PHYSICS 315

Physics of materials
Summary of lectures by

Birger Bergersen

Modified by Frank Linseisen

Instructor: Frank Linseisen

Department of Physics
University of British Columbia
Vancouver B.C., Canada V6T 1Z1
Office: Hennings 100
Tel.: (604)-822-3898
E-mail: [email protected]
WWW: http://www.physics.ubc.ca/ flinseis/Phys315/phys315.html
Revised: January 1999
Contents

1 Course organization 1

2 Solid structure 3
2.1 Bravais lattices and lattices with a basis . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Common crystal structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1.2 Packing fraction and coordination number . . . . . . . . . . . . . . . . . . . . 8
2.1.3 The 14 Bravais lattices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.4 Polycrystalline and noncrystalline solids . . . . . . . . . . . . . . . . . . . . . 13
2.1.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Structure determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.1 Reciprocal lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.2 Reciprocal lattice of a fcc lattice . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.3 Ewald construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.4 The Wigner-Seitz cell and the Brillouin zone . . . . . . . . . . . . . . . . . . 20
2.2.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3 Static properties 23
3.1 Atomic structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Lattice sums, binding energy and bulk modulus . . . . . . . . . . . . . . . . . . . . . 24
3.2.1 Molecular crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2.2 Ionic crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.3 Other common forms of bonding . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.4 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3 Stress and strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3.1 Elastic constants for a cubic crystal . . . . . . . . . . . . . . . . . . . . . . . 35
3.3.2 Elastic constants of an isotropic material . . . . . . . . . . . . . . . . . . . . 37
3.3.3 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.4 Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4.1 Point defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4.2 Dislocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.4.3 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5 Strength of materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.5.1 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

i
ii

4 Electrons and phonons 51

4.1 Lattice vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1.1 The harmonic approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1.2 Harmonic chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.1.3 Diatomic chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.1.4 Lattice vibrations in three dimensions . . . . . . . . . . . . . . . . . . . . . . 58
4.1.5 Density of states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.1.6 Phonons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.1.7 Debye model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.1.8 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2 Sommerfeld model for electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.1 Electron gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.2 Free electron gas at zero temperature . . . . . . . . . . . . . . . . . . . . . . 65
4.2.3 Low temperature electronic specific heat . . . . . . . . . . . . . . . . . . . . . 68
4.2.4 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3 Drude model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

5 Review 77
5.1 Midterm review problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.2 Old Midterms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.3 Old finals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.4 Values of physical constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

References 89
Chapter 1

Course organization

The preparation and use of materials plays an important role in all civilizations. In recognition of
this fact we label early cultures by the material which dominated technology at the time (stone,
bronze and iron age). Today thousands of new materials continue to revolutionize our everyday
existence. In the early part of this century the availability of inexpensive steel made it possible to
mass produce automobiles. During the last 50 years improvements in semiconductor technology
led to the computer revolution. More recently, improved understanding of ”soft matter”, liquid
crystals, polymers and biological materials is likely to bring profound changes in medicine and other
aspects of our lives.
In this course we study the physical properties of materials that have more structure than
simple liquids and gases. We start by considering crystalline solids, first from the point of view
of structure and then from the point of view of the cohesive energy. Next, we study the elastic
properties and move on to defects and their effect on the strength of materials. This is followed by
a discussion of thermal and transport properties due to electrons and phonons.
I will assume that the students in this course will have taken the usual complement of second
year mathematics courses (MATH 200, 215 and 221). A first course in thermodynamics such as
PHYS 213, or the corresponding course in chemistry or earth sciences will be assumed. Some
knowledge of modern physics (e.g. PHYS 200), and mechanics (e.g. PHYS 216) will be helpful,
but is not required.
There is no assigned text, but these notes will cover most of the course. A useful book which
covers most of the material in this course, but from a materials engineering rather than a physics
point of view, is that of Callister [1] . References to other appropriate books are given as we move
along.
There will be 10 problem sets and 1 midterm. The marking scheme is 25% assignments 25%
midterm and 50% final.
I have “borrowed” freely from old exams prepared by my colleagues David Balzarini , Birger
Bergersen , Myer Bloom , Roland Cobb , Dan Murray and Murray Neuman when selecting problems
and examples.
The course description in the UBC Calendar is
315 (3) Physics of Materials-Crystal structure, elasticity and phonons, elementary electronic
transport, defects, alloys, liquid crystals and polymers. Prerequisite: PHYS 203 or 213 or
CHEM 201 or 205. Corequisite MATH 215.[3-0-0].

1
2 Course organization
Chapter 2

Solid structure

The materials with the highest degree of order are the crystalline solids. Most solids exhibit such
order at least to some degree. This chapter is concerned with the geometrical arrangement of
atoms in a crystal. In particular we will discuss classification of crystals and how their crystalline
properties are measured. We will also briefly mention amorphous or glassy materials towards the
end of the chapter. For further reading on the material in this chapter I recommend the books by
Ashcroft and Mermin [2], Christman [5], Kittel [6] and Omar [8].

2.1 Bravais lattices and lattices with a basis

The most characteristic feature of a crystalline solid is its periodic structure. A crystal can be
thought to be made up of a space filling arrangement of basic building blocks called unit cells such
that the local arrangement of the atoms in any cell is reproduced exactly in all the other cells. We
define a lattice vector R as a vector which extends from a position in one cell to an equivalent
position in another cell. A translation by a lattice vector leaves the local environment unchanged,
and we say that a crystal exhibits translational symmetry, when the reference is shifted by a lattice
vector.
The concept of a perfect crystal is an idealization which cannot be completely realized for several
reasons

• The periodicity of a crystal will always be interrupted by its surface.

• In practice it is not possible to grow crystals without imperfections.

• At any non-zero temperature the atoms will vibrate about their equilibrium positions and
this will destroy the translational symmetry.

As we shall see later, lattice imperfections are responsible for many of the physical properties
of a solid, such as its strength, toughness, electronic properties and some of its optical properties.
However, before we can talk about the defects we must define the ideal which they disturb.
Crystallography is concerned with the geometrical properties of a crystal. A fundamental con-
cept is that of a Bravais lattice. Such a lattice is a set of points

R = n1 a + n2 b + n3 c (2.1)

3
4 2.1 Bravais lattices and lattices with a basis

where n1 , n2 , n3 are integers (negative, positive or zero). All the sites of a Bravais lattice are
equivalent and the local environment is identical at each lattice point. The three vectors a, b, c are
called primitive lattice vectors. The primitive lattice vectors must be non co-planar, but they need
not be orthogonal to each other. We define α to be the angle between b and c, β the angle between
c and a, γ the angle between a and b (see figure 2.1.a ). The three primitive lattice vectors span
a volume (figure 2.1.b) whose magnitude is

vc = |a · (b × c)| (2.2)

γ β b a

α c b
c

(a) (b)

Figure 2.1: (a) Primitive lattice vectors. (b) Primitive unit cell

We refer to this volume as a primitive unit cell. If a crystal structure is a Bravais lattice with
one atom at each lattice point, each primitive unit cell will have a volume vc = V /N where V is the
volume and N is the number of atoms of the crystal. Sometimes it convenient to work with a unit
cell which is larger than the primitive unit cell. To qualify as a unit cell it must still be possible
to use it as building block in a repeated pattern that fills the lattice. There are generally many
possible ways of choosing the primitive unit cell for a given crystal structure, but the volume of the
primitive unit cell will always be the same, and the volume of a non-primitive unit cell is always
an integer multiple of the volume of the primitive unit cell

2.1.1 Common crystal structures

We now apply the definitions which we have just made to the most common crystal structures.

Simple cubic lattice

The primitive unit cell is (as the name indicates) a cube. We take the side of the unit cube to be
a and choose the edges to be parallel to the three axes of a Cartesian coordinate system as shown
in figure (2.2-a)

a = ax̂, b = aŷ, c = aẑ

Solid structure 5

The volume of the primitive unit cell is a3 .

z
a
z b
y
x
a a
c
y
x

(a) (b)

Figure 2.2: (a) Simple cubic lattice. (b) Body centered cubic lattice

Body centered cubic lattice (bcc)

In the bcc structure identical atoms are located at the corners and at the center of a cube. The
unit cube is not a primitive unit cell - it is, however, a unit cell. We can construct the primitive
cell from the three lattice vectors a, b and c of figure 2.2-b.
a a a
a= (−x̂ + ŷ + ẑ); b = (x̂ − ŷ + ẑ); c = (x̂ + ŷ − ẑ) (2.3)
2 2 2
The volume of the primitive cell is from (2.2) 12 a3 . This result could be expected from the
following argument: assume that there are atoms at all the points of the bcc lattice. Each atom in
the center of a unit cube belongs to that unit exclusively, while each of the eight corner atoms are
shared by seven other cubes. The number of atoms per unit cube is thus 8/8 + 1 = 2. Since the
volume of the unit cube is a3 , the volume of the primitive unit cell must be vc = a3 /2.

Face centered cubic lattice (fcc)

The fcc lattice is another common Bravais lattice see figure (2.3). A possible choice of primitive
lattice vectors is
a a a
a=
(ŷ + ẑ); b = (x̂ + ẑ); c = (x̂ + ŷ) (2.4)
2 2 2
The volume of the primitive unit cell is

a3
a · (b × c) =
4
This result can also be found by noting that each of the six face centers of the unit cube is
shared by one other cube, while each corner is shared by seven other corners. Thus, there are 6/2
+ 8/8 = 4 atoms per unit cube. The volume of the primitive unit cell must thus be a3 /4.
6 2.1 Bravais lattices and lattices with a basis

a
b a

x c
y

Figure 2.3: Face centered cubic lattice

Triangular lattice
So far we assumed a three-dimensional Bravais lattice. It is easy to generalize to the case of two
dimensions. In this case the Bravais lattice is made up of the integer set {n1 , n2 } of points

R = n1 a + n2 b

where a and b are primitive lattice vectors. The area of the primitive unit cell is

Ac = |a × b|

The lattice is shown in figure 2.4(a). We choose the primitive lattice vectors to be
√
a 3
a = ax̂; b = x̂ + a ŷ
2 2
√
giving Ac = a2 3 /2 for the area of the unit cell.
Not all crystalline lattices are Bravais lattices. If we remove 1/3 of the lattice sites of the
triangular lattice we get the honey-comb lattice of figure 2.4(b) We notice from figure 2.4(b) that
the honey-comb lattice is made up of two distinct types of sites which we have respectively labeled
(i) and (ii) (or by filled and empty circles). The situation is unlike a Bravais lattice, where all
sites are equivalent. However, if we consider the sites of type (i) separately, we note they form
a triangular lattice which is rotated by 30o with respect to the original triangular lattice. The
primitive lattice vectors of this triangular lattice can be taken to be
√
3 3 √
a = a( x̂ + ŷ); b = a 3y
2 2
A translation by the vector d= ax̂ takes one from a site of type (i) to a site of type(ii). We call
the vector d a basis vector. The honey-comb lattice is an example of a lattice with a basis: all the
lattice sites are either on a Bravais lattice site
Solid structure 7

b x

(i) (ii)

x
b
a

Figure 2.4: (a) Triangular lattice. (b) Honey-comb lattice

R = n1 a + n2 b
or on a Bravais lattice site displaced by a basis vector
R = n1 a + n2 b + d
The concept of a lattice with a basis can be generalized in a straight forward manner to three
dimensions and the case where there are more than two non equivalent sites in the primitive unit
cell. A lattice with a n-point basis d1 · · · dn has its lattice points at one of the n types of sites
R = n1 a + n2 b + n3 c + di , i = 1, 2...n

Diamond structure
The lattice points of the diamond structure are localized either on a fcc lattice or on a fcc lattice
point displaced by a basis vector
8 2.1 Bravais lattices and lattices with a basis

a
d= (x̂ + ŷ + ẑ)
4

Figure 2.5: Tetrahedral bond in diamond lattice

A characteristic feature of the diamond structure is the tetrahedral bond (figure 2.5). The
orientation of the four tetrahedral directions are different for the fcc lattice points and on the
points displaced by the basis vector. The two types of sites are therefore not equivalent and the
diamond lattice is not a Bravais lattice.
In a number of compounds the two distinct sites of the diamond lattice are occupied by atoms of
different species. The structure is then called the zinc blende (ZnS) structure . A number of semi-
conductors such as GaAs and InSb crystallize in this structure, while the elemental semiconductors
silicon and germanium crystallize in the diamond structure.

Compound structures
All compounds (materials made up of more than one element) will necessarily crystallize in a
lattice with basis. An important crystal structure for compounds is the sodium chloride (NaCl)
structure(figure 2.6(a)) which is fcc with the basis d = a2 x̂.
A third fairly common compound structure is that of cesium chloride in figure 2.6(b) . This
structure is simple cubic with a basis vector d = a2 (x̂ + ŷ + ẑ). It differs from the bcc lattice only
in that the body center and corner sites are occupied by atoms of different species in the CsCl
structure.

2.1.2 Packing fraction and coordination number

There are a number of factors that dictate which crystal structure will be realized for a given
compound. As we indicated earlier, the tetrahedral covalent bond favors the diamond (or zinc
blende) structure. In the case where the forces that hold the crystal together are not directional, as
in the case of the covalent bond, packing considerations are often important. A convenient measure
Solid structure 9

z
z

a
y y

x x
(a) (b)

Figure 2.6: (a)NaCl lattice. (b) CsCl structure

of the packing efficiency of a given structure is the packing fraction f . We draw a sphere of radius
r around each atom (or ion). We first assume that we are dealing with a substance in which the
atoms are all of the same kind and which crystallizes in a Bravais lattice. We choose the radius r
so that the atoms touch, but do not penetrate each other. The packing fraction is then defined as

4πr 3
f=
3vc

where vc is the volume of the primitive unit cell.

Another important characteristic of a crystal structure is the coordination number . For a
Bravais lattice the coordination number is the number of nearest neighbors, or number of atoms
that touch each other. Generally, a high coordination number and a high packing fraction go
together. In the case of a compound the coordination number may be different for the different
constituents.

Simple cubic lattice

From figure 2.7(a) we see that r= a2 . The volume of the unit cell is a3 . The packing fraction is thus

4π π
fsc = = = 0.5236
3×8 6

The coordination number of the sc lattice is 6.

Fcc lattice
Figure√2.7(b) shows the packing arrangement on one of the cube faces. We see from the figure that
4r= a 2. Since there are four spheres per unit cube the packing fraction is
10 2.1 Bravais lattices and lattices with a basis

a a 2

(a) (b)

γ γ γ γ γ
β β β β
α α α α
γ γ γ γ γ
β β β β
α α α α
γ γ γ γ γ
β β β β
α α α α
γ γ γ γ γ

(c)

Figure 2.7: Packing fraction of (a) the simple cubic lattice. (b) The fcc lattice. (c) Close packing
of spheres.

√ !3 √
4 × 4π 2 π 2
ff cc = = = 0.7405
3 4 6

The coordination number of the fcc lattice is 12.

Hexagonal closed packed lattice

The most efficient packing of spheres in a single two-dimensional layer is the triangular lattice (figure
2.7(c)), with coordination number 6. We can construct a closed packed three dimensional lattice
by adding successive triangular layers on top of the original layer. The best way to do this is to put
the next layer either above the sites marked α or the sites marked β in figure 2.7(c). The reader
may convince her or himself that if the stacking sequence of successive layers is · · · γαβγαβγαβ · · ·
we recover the fcc lattice. Another possible sequence which gives the same packing fraction is
· · · γαγαγα · · · The resulting crystal structure (figure 2.8) is called hexagonal closed packed (hcp).
This lattice structure is not a Bravais lattice since the local environment of the atoms in γ and
α layers are different. The set of γ − points in the hcp structure does form a Bravais lattice, the
hexagonal lattice. The primitive lattice vectors of this lattice can be chosen as
√
1 3
a = ax̂; b = a( x̂ + ŷ); c = cẑ
2 2
The hcp lattice is then hexagonal with basis
Solid structure 11

b
a

Figure 2.8: The hexagonal closed packed lattice.

1 1 1 1 1
d = c + (a + b) = a( x̂ + √ ŷ) + cẑ
2 3 2 2 3 2
The coordination number of the hcp is 12, just as for the fcc lattice.
The concept of packing fraction can be extended to compounds. Consider an ideal crystal A+ B −
in which the A+ and B − separately form a Bravais lattice with volume vc of the primitive unit
cell. We associate an ionic radius r+ and r− to each species. Assuming that the closest approach
of two positive, two negative and ions of opposite kind are 2r+, 2r− and (r+ + r− ) , respectively,
we define the packing fraction as
4π
f= (r+3 + r− 3 )
3vc

2.1.3 The 14 Bravais lattices

There are 14 possible Bravais lattice for a three dimensional crystal (see figure 2.9). With the
angles α, β and γ defined as in (figure 2.1) these lattices are tabulated in 2.1. The reader may well
wonder why are there just 14 three dimensional Bravais lattices. Why e.g. do we not include a
base centered cubic lattice? We do not consider this lattice to be a distinct lattice since it is a
special case of the simple tetragonal lattice, as is the base centered tetragonal lattice. We refer the
interested reader to [2] for a more detailed discussion of the classification of the lattices.
The structure and cell dimensions of some elements and compounds are listed in table 5.1 below.
A standard notation has evolved to describe crystallographic directions and distances. We write
for a general lattice vector

R = n1 a + n2 b + n3 c = [n1 n2 n3 ]
(note the square bracket and the absence of commas). If some of the numbers are negative, it
is conventional to put the minus sign above the number rather than in front, e.g.
12 2.1 Bravais lattices and lattices with a basis

Trigonal
Triclinic Hexagonal

Simple monoclinic Base centered Simple tetragonal Body centered

monoclinic tetragonal

Simple Base centered Body centered Face centered

orthorombic orthorombic orthorombic orthorombic

Simple cubic Body centered Face centered

cubic cubic

Figure 2.9: The 14 Bravais lattices.

Solid structure 13

R = a − b + c = [11̄1]

Often several crystallographic directions are equivalent e.g. the [100] and [010] directions in a
cubic crystal. We indicate collectively all directions which are equivalent to [n1 n2 n3 ] as hn1 n2 n3 i
(angular brackets).
In the case of the cubic crystals it is conventional to let the indices n1 , n2 , n3 refer to the sides
of the unit cube rather than the primitive lattice vectors. For example, the three lattice vectors
(2.3) of the bcc lattice are written as
     
1̄ 1 1 1 1̄ 1 1 1 1̄
; ;
222 222 222

Triclinic a 6= b 6= c α 6= β 6= γ 6= 90o
Trigonal a=b=c α = β = γ 6= 90o
Hexagonal a = b 6= c α = β = 90o , γ = 120o
Simple monoclinic a 6= b 6= c α = β = 90o 6= γ
Base centered monoclinic
Simple orthorombic a 6= b 6= c α = β = γ = 90o
Base centered orthorombic
Body centered orthorombic
Face centered orthorombic
Simple tetragonal a = b 6= c α = β = γ = 90o
Body centered tetragonal
Simple cubic a=b=c α = β = γ = 90o
Body centered cubic
Face centered cubic

Table 2.1: 3-D Bravais lattices.

2.1.4 Polycrystalline and noncrystalline solids

It is rare to find a large single crystal in which the periodic arrangement extends throughout the
solid. When such crystals are cut one may see flat faces or facets such as in gemstones. The facet
planes will normally correspond to high symmetry directions of the structure which are the ones
for which the normals to the planes correspond to low index directions such as the h100i, h110ii
or h111i direction. A single crystal will be anisotropic, i.e. mechanical, optical and transport
properties will be different in different directions depending on the orientation of the crystalline
axes. Single crystals are difficult to grow, the process must be done very slowly and under carefully
controlled circumstances.
Most solids are polycrystalline, i.e. contain many small crystallites separated by grain bound-
aries. Polycrystalline materials will be isotropic (the same in all directions) on length scales which
are large compared to the size of the individual grains.
14 2.1 Bravais lattices and lattices with a basis

Some solids lack any regular arrangement of the constituent atoms or ions on length scales
more than a few lattice spacings. When this happens the solid is said to be amorphous or glassy.
Generally, in order to obtain an amorphous solid from a melt, the cooling has to be rapid. In the
case of metals or other materials with a simple crystal structure the cooling has to be extremely
rapid. Such a rapid quenching can typically only be achieved for thin strips and films. Certain
materials, which crystallize into very complicated structures, order only with difficulty and are
easy to produce in a glassy form. The freezing transition between a liquid and a crystalline solid
is a sharp phase transition. In the case of a liquid and a glass the transition is gradual and the
distinction is not that sharp. Typically, one defines the transition from a liquid to a glass to take
place when the viscosity (resistance to flow) exceeds a certain value. Of particular importance are
the silicate glasses used to make window glass and bottles. Solids made up of organic polymers can
be polycrystalline or amorphous.

2.1.5 Problems
Problem 2.1-1:

Compute the packing fraction f for the bcc lattice

Problem 2.1-2:
√
(a) Show that the packing fraction f for the diamond lattice is π 3/16.
(b) What is the packing fraction and coordination number of the honeycomb lattice?

Problem 2.1-3:

Consider the hexagonal closed packed lattice:

p
(a) Show that c = a 8/3 = 1.633a. Frequently a crystal structure is called hcp even if c is not
exactly equal to the ideal value. √
(b) Show that the packing fraction for the ideal hcp lattice is π 2/6 = 0.7405.

Problem 2.1-4:

The ionic compound A+ B − crystallizes in the NaCl structure (see figure 2.6). Plot the packing
fraction as a function of the ratio ξ = r+ /r− . Assume that ξ < 1.

Problem 2.1-5:

Repeat the calculation of problem 2.1-4 for the CsCl structure.

Problem 2.1-6:

Use the information in table 5.1 to calculate the densities (in kgm−3 ) of the following solids: (a)
Aluminum, (b) Iron, (c) Silicon and (d) Zinc. Atomic weights of some common elements are listed
in table 5.2.
Solid structure 15

Problem 2.1-7:
SrTiO3 crystallizes in the perovskite structure. The strontium atoms are at the corners of cubes
with side a, the titanium atoms are at the body centers, while the oxygen atoms occupy the cube
faces.
(a). What is the Bravais lattice type?
(b). Verify that the primitive unit cell contains one Sr, one Ti and three O atoms.
(c). Write down a set of primitive lattice vectors and basis vectors for the perovskite structure.

Problem 2.1-8:
The primitive lattice vectors of a certain bravais lattice can be written
1 1
R = (n1 + 2n2 )ax̂ + n1 bŷ + n3 cẑ
2 2
What is the lattice type?

Problem 2.1-9:
In each of the following cases indicate whether the structure is a Bravais lattice. If it is, give three
primitive lattice vectors, if it is not describe it as a Bravais lattice with as small as possible basis.
In all cases the length of the side of the unit cube is a.
(a) Base centered cubic (simple cubic with additional points in the centers of the horizontal
faces of the cubic cell).
(b) Side centered cubic (simple cubic with additional points in the centers of the vertical faces
of the cubic cell).
(c) Edge centered cubic (simple cubic with additional points at the midpoints of the lines joining
nearest neighbors).

2.2 Structure determination

2.2.1 Reciprocal lattice
The elementary theory of crystal structure determination by diffraction is based on Bragg’s law
(figure 2.10) We imagine a crystal as made up of lattice planes. An incoming plane wave is reflected
by these planes . We assume that the angle of incidence is the same as the angle of reflection
(specular reflection). We also assume that the wavelength of the reflected light is the same as the
wavelength of the reflected light (elastic scattering). If the wavelength of the incoming wave is
λ, the angle of incidence is θ, and the perpendicular distance between the planes is d, we have
constructive interference between waves reflected from a family of parallel planes if

2d sin θ = nλ (2.5)
where n is an integer. Consider a Bravais lattice. We define a lattice plane as any plane which
contains at least three points that do not lie on the same line (non co-linear points). A family of
planes is a set of lattice planes which together include all points of the Bravais lattice. The equation
for a plane is in vector notation
16 2.2 Structure determination

ki

G
θ
θ

kf
d
θ θ

(a) (b)

Figure 2.10: (a) Bragg’s law.(b) Elastic scattering.

R · n̂ = c

where R is a vector to a point in the plane, n̂ is a unit vector normal to the plane and c is the
perpendicular distance from the origin to the plane. Let us now return to the Bragg condition (2.5)
and let ki be the wave vector of the incident beam, kf the wave vector of the elastically scattered
beam and

G = kf − ki

the scattering vector. We have

2π
λ=
k
and find that with θ the angle of incidence

G = 2k sin θ

and (2.5) becomes

R · G = 2πn (2.6)
As n runs through the set of integers equation (2.6) defines a family of planes. The vector G is
called a reciprocal lattice vector.
Let a, b and c be a set of primitive lattice vectors of a Bravais lattice, we can construct a set
of primitive reciprocal lattice vectors from

A · a = 2π; A · b = 0; A · c = 0
B · a = 0; B · b = 2π; B · c = 0 (2.7)
C · a = 0; C · b = 0; C · c = 2π
with the solution
Solid structure 17

2πb × c 2πc × a 2πa × b

A= ; B= ; C= ; (2.8)
a · (b × c) a · (b × c) a · (b × c)
We can now construct a Bravais lattice of reciprocal lattice vectors

G = hA + kB + lC (2.9)
where h, k, l are integers. If R is a general vector (2.1) of the direct lattice we see that (2.7)
gives
G · R = 2π(n1 h + n2 k + n3 l) = 2πn
Suppose K is the shortest reciprocal lattice vector in a given direction, then K · R = 2πn defines
a family of planes, and the coefficients h, k.l of this family are called Miller indices. The distance
between the planes is 2π/K and it is customary to use the Miller indices to specify a reciprocal
lattice vector as
G = (hkl)
It can be seen from (2.8) that the volume of the reciprocal primitive unit cell is

(2π)3 (2π)3
Vc = |A · (B × C)| = = (2.10)
|a · (b × c)| vc
In the case of cubic crystals it is customary to let the Miller indices refer to unit cube not the
primitive unit cell. Figure (2.11) show some crystallographic planes and their Miller indices for a
cubic crystal. In a cubic crystal the (100),(01̄0)and (001) planes are equivalent. It is conventional
to refer collectively to {hkl} as the (hkl) family of planes as well as all the symmetry related ones.

{100} {110} {111}

Figure 2.11: Some crystallographic planes of a cubic crystal.

2.2.2 Reciprocal lattice of a fcc lattice

Applying (2.8) to (2.4) we find

2π 2π 2π
A= (−x̂ + ŷ + ẑ); B = (x̂ − ŷ + ẑ); C = (x̂ + ŷ − ẑ) (2.11)
a a a
18 2.2 Structure determination

and we see from (2.4) that the reciprocal of the fcc is a body centered cubic lattice with side of
unit cube 4π
a , if a is the side of the unit cube of the original fcc lattice. Similarly, the reciprocal to
the bcc lattice is fcc.

2.2.3 Ewald construction

We can illustrate some of the more common experimental methods of determining crystal structures
by a simple geometric construction called the Ewald construction (figure 2.12)

. . . . . . . .
. . . . . . . . .
. . . . . . . .
. . . . . . . . .
. . . . . . . .
. . k.f . . . . . .
ki
. . . . . . . .
G
. . . . . . . . .

Figure 2.12: Ewald construction.

We choose one of the reciprocal lattice points in figure 2.12 as the origin. We assume that there
is an incoming beam for which the magnitude and direction of the wave-vector ki is held fixed. If
the beam undergoes elastic scattering the magnitude of the scattered wave vector kf must be the
same as the magnitude of the incoming wave vector. The possible wave vectors of the diffracted
beam must therefore lie on the surface of a sphere which is called the Ewald sphere. This surface of
this sphere will in most cases only contain one reciprocal lattice vector - the one at the base of ki.
Whenever another reciprocal lattice vector touches the sphere the condition for Bragg reflection is
satisfied. There are two main methods to obtain x-ray diffraction from a single crystal. In the Laue
method one uses a beam which contains a continuous spectrum of wave vectors kmin < ki < kmax
but the orientation of the crystal studied and the direction of the beam is held fixed. From figure
2.13 we see that the reciprocal lattice points in the crescent shaped volume between the two spheres
will cause scattering and can be picked up on a film surrounding the crystal under study or by a
detector. The Laue method is frequently used to orient a single crystal with respect to a fixed axis
in an experimental set up.
Another important method, the rotating crystal method, uses a monochromatic beam. The
arrangement has the crystal mounted on a spindle located at the axis of a cylinder (see figure 2.14),
with an opening for the incoming beam and an x-ray film mounted on the inner surface. For some
orientations the Ewald condition will be satisfied and the diffracted beam will produce a black spot
Solid structure 19

. .
. . . . .
. . . . . .
. . . . . . .
max

. . . k. . .
. . . . .
k
min
. . . . . .
. . .

Figure 2.13: Ewald construction for the Laue method.

on the film.

AXIS

x FILM

INCOMING BEAM SAMPLE

Figure 2.14: Rotating crystal method.

Diffraction from a powder sample works in a similar fashion as the rotating crystal method.
The randomly oriented crystallites in the powder diffracts like a single grain which has been rotated
very many different ways.
X-rays are not the only available probes for diffraction experiments. Electron and neutron
beams are commonly used. The wavelength of a particle beam is given by the de Broglie relation
20 2.2 Structure determination

λ = h/p, where h is Planck’s constant and p the momentum. For diffraction to be a useful tool
in structure analysis, the wavelength should be of the same order of magnitude as typical lattice
spacings i.e. ∼Å. Using the formula for the energy E = p2 /2m, where m is the mass of a particle,
one can show that the wavelength of a neutron in Å can be written
0.28
λ= √
E
if the energy E is given in electron volts (eV ). For the wavelength to be 1 Å the energy would
have to be 0.08 eV . Since the thermal energy kB T at room temperature is 0.025 eV , the neutrons
used in diffraction experiments are referred to as thermal neutrons.
In the case of electrons the wavelength in Å can be written
12.25
λ= √
E

where again the energy is given in eV . At energies where λ is of the order a few Ångström
the penetration depth of electrons is typically only a few lattice spacings. Low energy electron
diffraction (LEED) is for this reason mostly used for studies of the surface properties of materials.

2.2.4 The Wigner-Seitz cell and the Brillouin zone

. .
Z

. . .
X Y

. .

(a) (b)

Figure 2.15: Wigner Seitz cell , (a) for the triangular lattice, (b) for the bcc lattice. In the latter
case the square faces are perpendicular to the [100] , [010] and [001] directions, while the hexagonal
faces are perpendicular to the [111] and symmetry equivalent directions.

It is often practical to divide a crystal into unit cells in a different way than what we did in
figure 2.1. Consider the cell defined by the following construction (for the two dimensional version
Solid structure 21

of this see figure 2.15). First draw lattice vectors connecting a lattice point at the origin with
neighboring lattice points. Then draw planes bisecting these vectors. The smallest volume which
is enclosed by such planes and includes the point at the origin is a Wigner Seitz cell. This cell is
frequently used as a primitive unit cell. In this way the crystal is divided into volumes which have
the property that all points inside each cell are closer to the lattice point at its origin than to any
other lattice point.
The Wigner Seitz construction can also be carried out for the reciprocal lattice. The resulting
unit cell is then called a Brillouin zone The Wigner Seitz cell can be generalized to non crystalline
structures. For a disordered set of points the resulting cell will be bounded by planes forming the
Voronoy polyhedra. The shape of each Voronoy polyhedron will generally depend on which point it
surrounds. The plane normals connecting the neighboring points form a network which is referred to
as the Delaunay tessellation. The Delaunay tessellation has many uses outside solid state physics.
For example, in geography one is often required to produce topographical maps where contours
correspond to point of equal altitude, mean rainfall, temperature, atmospheric pressure etc. These
maps must be generated from a grid of observation stations. The Delaunay tessellation of this grid
forms a useful basis for computer generated interpolation schemes.

2.2.5 Problems
Problem 2.2-1:
An x-ray source emits an x-ray line of wavelength λ = 1.54 Å. The lattice constant and crystal
structures of iron and aluminum are found in table 5.1.
(a) Find the Bragg angle(s) for reflections from the (111) plane of Al.
(b) Find the Bragg angle(s) for reflections from the (110) planes of Fe

Problem 2.2-2:
The construction (2.9) of the reciprocal lattice assumes that the direct vectors a, b, c of (2.8) are
primitive lattice vectors. If instead of using the primitive vectors one uses the lattice vectors of
the unit cube of e.g. the fcc or bcc lattices, what happens is that (2.9) predicts some reciprocal
lattice points that should not be there. Consider a simple cubic (sc) lattice for which the side of
the primitive unit cube is a. Construct the reciprocal lattice of this structure.
(a) Which of the reciprocal lattice vectors of the sc lattice are missing in the reciprocal lattice
constructed from the fcc lattice with the same unit cube side a.
(b) Which of the reciprocal lattice vectors of the sc lattice are missing in the reciprocal lattice
constructed from the bcc lattice with a unit cube of side a.

Problem 2.2-3:
It is observed that crystals generally cleave most easily along the crystal planes which are furthest
apart. Clearly, these are also the planes which have the highest density of atoms per unit area.
From (2.6) we see that the perpendicular distance between nearby lattice planes is 2π/G where
G is the shortest reciprocal reciprocal lattice vector in the direction normal to the planes. The
shortest reciprocal lattice vectors for the fcc lattice are normal to the (111) planes, while the (110)
planes are normal to the shortest reciprocal lattice vectors for the bcclattice. Assume that the side
of the unit cube is a.
22 2.2 Structure determination

(a) What is the density of atoms (number per unit area) on a (111) plane of a fcc lattice?
(b) What is the density of atoms on a (110) plane of a bcc lattice?

Problem 2.2-4:
The primitive lattice vectors of a 2-dimensional triangular lattice are
√
a a 3
a = ax̂; b = x̂ + ŷ
2 2
where a is the nearest neighbor distance.
(a) Find the reciprocal lattice
(b) Draw the Wigner Seitz cell and locate the coordinates of its corners.
(c) Draw the Brillouin zone and locate the coordinates of its corners.

Problem 2.2-5:
Find the coordinates of the corners of the Wigner Seitz cell for the bcc lattice (figure 2.15 (b)).

Problem 2.2-6:
An X-ray reflection from a certain crystal occurs at an angle of incidence of 45◦ when the crystal
is maintained at 0◦ C. When it is heated to 150◦ C the angle changes by 6.4 minutes of arc. What
is the linear thermal expansion coefficient of the material?
Chapter 3

Static properties

3.1 Atomic structure

Depending on the type of material, we will consider matter as made up of either atoms, molecules,
or ions. An atom is made up of a nucleus surrounded by electrons. The nucleus is in turn made
up of protons and neutrons. The mass of an electron is me = 9.109 × 10−31 kg, many times smaller
(many=1836) than the mass of a proton, mp = 1.6726×10−27kg, or neutrons mn = 1.6749×10−27kg.
The mass of an atom thus rests almost exclusively in the nucleus and is often given in terms of
the atomic mass unit or a.m.u. (1a.m.u. = 1.6605 × 10−27 kg). A chemical element is specified by
the number of protons in the nucleus. Atoms which differ only in the number of neutrons in their
nuclei are called isotopes. Different isotopes of the same element have except for the mass difference
essentially identical properties.
Molecules are tightly bound groups of atoms, covalently bonded by the sharing of electrons.
Ions are atoms that have a net electric charge because they have either given up one or more
electrons or accepted extra electrons.
The electrons in an atom are characterized by 4 quantum numbers. Electrons with principal
quantum number n = 1, 2, 3, 4... are referred to as K, L, M, N.. electrons, respectively. The electrons
with the lowest principal quantum are closest to the nucleus and most tightly bound. We refer
to these as core electrons. The electrons with the highest values of n are the valence electrons
and they are less tightly bound to the nuclei than the core electrons. The chemical properties
of a material depend almost exclusively on how the valence electrons are distributed among the
atoms, molecules and ions. Energies associated with chemical change are typically of the order
electron-volts per atom, which makes this a convenient unit of energy (1eV = 1.6022 × 10−19 J).
The second quantum number l specifies the orbital angular momentum of an electron. The
possible
p values of l are 0, 1, . . ., n − 1. According to quantum mechanics the angular momentum is
l(l + 1)h̄, where h̄ is Planck’s constant h divided by 2π. Electrons with l = 0, 1, 2, 3.. are referred
to as s, p, d, f.. electrons respectively. The number of electrons in the different n, l states specifies
the electronic configuration, e.g. the ground state of the element argon with chemical symbol Ar
is 1s2 2s2 2p6 3s2 3p6 indicates that there are 2 electrons in the K−shell with n = 1, these are both
s-electrons, i.e. have l = 0. There are 8 electrons in the L shell 2 of then s−electrons and 6
p−electrons with l = 1. There are also 2 s−electrons and 6 p−electrons in the M −shell.
The third quantum number m is associated with the component of the orbital angular mo-
mentum along a suitably chosen axis. This quantum number can take on any of the 2l + 1 values

23
24 3.1. Lattice sums, binding energy and bulk modulus

m = −l, −l + 1, . . ., l − 1, l. In an isolated atom electrons with different m quantum number have

the same energy. The energy levels will be split, however, in a magnetic field and may also be split
inside a crystal where different orientations need not be equivalent.
Finally, electrons carry an intrinsic angular momentum, called spin. The component of the
spin may take on one of two different values σ = ± 12 h̄ with respect to any axis. Electrons satisfy
the Pauli exclusion principle, only one electron may occupy a given quantum state. The main
importance of the spin degree of freedom for us will come when we need to count states - the effect
is to double the number of available states.

3.2 Lattice sums, binding energy and bulk modulus

What holds the crystal together? In all materials the forces are of electrostatic origin, but the
detailed mechanism varies widely. On the basis of the dominant mechanism for bonding we classify
crystals as either

• molecular (Van der Waals forces)

• ionic (Coulomb forces)

• covalent (“Quantum mechanical” forces)

• metallic (“Quantum mechanical” forces)

• hydrogen bonded (“Quantum mechanical” forces)

We will discuss these types of bonding in turn – unfortunately it is only in the two first cases
that we are able to go beyond very qualitative arguments.

3.2.1 Molecular crystals

The constituents of a molecular crystal are atoms and molecules with an inert electronic config-
uration. This configuration remains intact when the material is cooled down from the gas phase
through the liquid state into a crystal. To a reasonable approximation, the interaction between
the atoms or molecules can be described by a pair potential v(ri − rj ). The simplest cases are the
crystals formed from the inert gases neon (Ne), argon (Ar), krypton (Kr) and xenon (Xe). In this
case the atoms have spherical symmetry and the potential is central v = v(|ri − rj |) This potential
can be thought of as made up of two parts. At short range when the atomic cores overlap there
is a strong repulsion, whose origin lies in the Pauli exclusion principle. Since two electrons cannot
occupy the same state, the core states will have to be strongly distorted when two atoms get too
close together, and there is a large energy cost associated with such distortions. At larger distances
there is no significant overlap between the electron states of neighboring atoms. The dominant
force is then due to the van der Waals attraction. The corresponding potential falls off as the sixth
power of the distance– and since force is minus the gradient of the potential, the corresponding
force will be ∼ 1/R7 . This r dependence can be understood from the following argument:
If an atom is exposed to an electric field E the force on the negatively charged electrons and
the positively charged nucleus will be in opposite directions. This will induce a dipole moment
3. Static properties 25

1 V(r) / ε

r /σ
0.5 1.5

-1

Figure 3.1: Lennard-Jones potential (see equation 3.2).

pj = αE~ (3.1)
where α is the polarizability. If the electric field is absent the average dipole moment will be
zero, but there will still be a fluctuating dipole moment due to quantum effects (zero point motion).
The electric field due to a dipole falls off as the inverse third power of the distance, and it will
induce a dipole moment in neighboring atoms which is proportional to the field according to (3.1).
The interaction energy will then be of the form

h−pj · E~i i = α < Ei2 >

where pj is the induced dipole moment on atom j due to the fluctuating field E~i from atom i.
The angular brackets hi indicate average value. The average value of the fluctuating electric field
is hEi = 0, while the mean square value will be non-zero. Since the fluctuating field falls off as
the inverse third power of the distance, the mean square field will be proportional to the inverse
sixth power of the distance to the fluctuating atom. The van der Waals interaction energy is thus
∼ 1/r 6 .
An approximate way of combining the repulsive and attractive interactions between the atoms
in a molecular crystal is the Lennard-Jones potential

A B σ σ
v(r) = − = 4[( )12 − ( )6 ] (3.2)
r 12 r 6 r r
where A and B are constants which depend on which atom or molecule is involved. It is
conventional to parametrize the potential in terms of an energy parameter  and length parameter
σ, in terms of which A= 4σ 12 , B= 4σ 6 . We plot the Lennard -Jones potential in figure 3.1.
The potential changes sign at r = σ. By differentiating (3.2) we find that the potential has
minimum at r = 21/6 σ. The value of the potential at the minimum is −. In the case of the
inert gases one can estimate the potential parameters using data obtained from measurements of
corrections to the ideal gas law in the gas phase (virial coefficients). Some typical values are listed
in table 3.1.
If we assume that all the atoms are at their equilibrium positions in the perfect lattice, we can
estimate the cohesive energy U of the crystal as
26 3.1. Lattice sums, binding energy and bulk modulus

Element σ [Å]  [eV]

Ne 2.74 0.0031
Ar 3.40 0.0104
Kr 3.65 0.0140
Xe 3.98 0.0200

Table 3.1: Lennard-Jones parameters for the inert gases

NX
U= v(R) (3.3)
2 R
where the sum in (3.3) is over the lattice. Substituting (3.2) we find

( 6 ) ( 6 )
X  σ 12 
σ σ
12 X 
rnn
12 
σ
6 X 
rnn
U = 2N  − = 2N  − (3.4)
R
R R rnn R
R rnn R
R

where rnn is the nearest neighbor distance. We define the geometrical constants
X  rnn n
An = (3.5)
R
R
The constants An can easily be evaluated numerically for different lattice types. The inert gases
crystallize in the fcc structure at low temperatures. For the fcc lattice one finds A6 = 14.44862,
A12 = 12.13188 Using (3.5) in (3.4) we find for the cohesive energy
(  12  6 )
σ σ
U = 2N  A12 − A6 (3.6)
rnn rnn
if we differentiate (3.6) with respect to rnn we find that the cohesive energy has a minimum for
 1/6
2A12
rnn = σ = 1.09σ (3.7)
A6
for the fcc lattice. At this equilibrium separation we find U = −8.6N  for the cohesive energy.
Another quantity which can easily be evaluated in this model is the bulk modulus

∂P
B = −V (3.8)
∂V
where V is the volume per particle and P is the pressure which at low temperatures reduces to
∂U
P =−
∂V
It is convenient to introduce u = U/N and vc = V /N. We find

∂2u
B = vc (3.9)
∂vc 2
3. Static properties 27
√
For the
√ fcc lattice with a the side of the unit cube we have vc = a3 /4, rnn = a/ 2, hence
3 / 2 and
vc = rnn
√
∂ 2 ∂
=
∂vc 3rnn 2 ∂r

we find after some algebra

5
4(A6 ) 2
B= 3 (3.10)
σ 3 (A12 ) 2
In table 3.2 we compare the predictions of our simple theory with experiment

Element rnn [Å] rnn [Å] u[eV] u[eV] B[109 Pa] B[109 Pa]
theory exper. theory exper. theory exper.
Ne 2.99 3.13 -0.027 -0.02 1.81 1.1
Ar 3.71 3.75 -0.089 -0.08 3.18 2.7
Kr 3.98 3.99 -0.120 -0.11 3.46 3.5
Xe 4.34 4.33 -0.172 -0.17 3.81 3.6

Table 3.2: Some properties of inert gas crystals

We see that the agreement for rnn is quite good, the cohesive energy agrees well only for the
heavier elements and the discrepancy is somewhat larger for the bulk modulus. The systematic
dependence of the error on atomic mass suggest that the discrepancies have something to do with
the quantum zero point motion which decreases with mass. On the whole, table 3.2 indicates that
the rare gas solids are well understood, at least qualitatively. There is one somewhat embarrassing
fact. The rare gas atoms crystallize in the fcc rather than in the hcp lattice. The latter has
A12 = 12.13229 and A6 = 14.45489. If one calculates the cohesive energy for the hcp lattice
one finds u = −8.61106N  which is slightly lower than the value u = −8.61077N  one finds by
substituting (3.7) into (3.6). Since the difference is so minute, other factors such as the zero-point
energy and three-body corrections to the potential come into play.

3.2.2 Ionic crystals

As we shall see, the ionic crystals are another group of solids for which it is possible to construct
a simple theory of the cohesive energy that agrees reasonably well with experiment. To be specific
we consider a structure such as CsCl or N aCl, and assume that there are two singly charged ions
per primitive unit cell. There are N primitive unit cells in the crystal so that there are altogether
2N ions. The largest contribution to the cohesive energy will come from the Coulomb interaction

e1 e2 1
(3.11)
4π0 r
between two ions of charge e1 and e2 separated by a distance r.
Proceeding as in (3.4) we find for the Coulomb energy of the crystal
28 3.1. Lattice sums, binding energy and bulk modulus

 
N e2  X d

d 
Ucoul =− 1+ − (3.12)
4π0 d  R6=0
|R + d| R 

The sum over R is over the Bravais lattice. The term inside the curly brackets
X d d

α=1+ − (3.13)
R6=0
|R + d| R

is called the Madelung constant. It is a purely geometric factor which depends only on the
crystal structure and not on the lattice constant. When evaluating (3.13) we run up against the
problem that the series is not convergent, and depending on the order in which the terms are
summed one can add up the series to different results. To see this let us attempt to evaluate the
sum by splitting it up into two contributions
X d d
 X 
d d
 X  d d

− = − + −
R6=0
|R + d| R R6=0,R>R0
|R + d| R R>R0
|R + d| R

X   Z Z Z
d d Nd 2π θ |R + d| − R
≈ − − R2 dR dφ dθ
R6=0,R<R0
|R + d| R V R>R0 0 −θ |R + d|R

For fixed angles φ and θ the integrand will approach a value which is approximately independent
of R for large R. The value of the integral will thus depend on the shape of the crystal volume V .
The physical origin of this problem is that the long range of the Coulomb interaction small surface
charges can cause a significant change in the electrostatic potential, which in turn can effect the
sum in (3.13). At equilibrium the net charge and dipole moment of the crystal will be zero. The
series will be summed correctly if the terms are summed up in blocks that have no net charge, no
dipole moment nor quadrupole moment. The sum over blocks will then be convergent, and the
value of this sum is the correct Madelung constant. The value of α for some crystal structures is
given in table 3.1

Structure α
NaCl 1.7476
CsCl 1.7627
ZnS 1.6381

Table 3.3: Madelung constant for some crystal structures

In addition to the Coulomb interaction we need to consider the core-core repulsion in estimating
the lattice constant, bulk modulus and cohesive energy. One way to proceed is to write the cohesive
energy on the form

N αe2 NC
U = Ucoul + Ucore = − − (3.14)
4πo rnn (rnn )m
where C is a constant. In the case of molecular crystals we chose m = 12 for convenience. Since
there is no free parameter associated with the Coulomb term we can now afford to use m as an
3. Static properties 29

adjustable parameter. If we e.g. use experimental data to fit C and m to the observed nearest
neighbor distance rnn and bulk modulus B ( 3.9) one can test the theory by comparing theoretical
values of the cohesive energy with experimental data. In the case of simple ionic crystals with the
NaCl structure, the theoretical and experimental values for the cohesive energy differ by at most
a few percent and it would appear that he simple theory we have outlined is qualitatively correct.
The form (see equation 3.14) for the core-core interaction is somewhat arbitrary. Another common
two parameter form for Ucore is

Ucore = A exp(−rnn /λ)

with A and λ adjustable parameters. These parameters can be fitted in a similar spirit to what
we used for C and m of (see equation 3.14).

3.2.3 Other common forms of bonding

The covalent crystal can be thought of as giant molecules and the cohesive energy comes about
largely from the chemical energy of contributing bonds. This type of bond comes about by the
sharing of electrons between adjacent atoms, and it is quantum mechanical in origin. Reasonable
semi-quantitative results for the binding energy can be obtained using chemical rules of thumb, the
subject is beyond the scope of this course.
Similarly, there is no simple theory for cohesion in metals . The valence (or conduction) electrons
occupy delocalized orbits percolating through the crystal and can be thought of as constituting an
electron fluid which is confined within the crystal by the requirement of charge neutrality. There
are important contributions to the energy from the attractive interaction between electrons and the
positive ions and the repulsive electron-electron interaction. The Pauli exclusion principle raises the
energy by forcing the electrons to occupy electron states with a nonzero kinetic energy. Quantum
mechanical considerations play a dominant role and the electrons participate in a complicated
quantum dance in which they avoid each other and take advantage of the attraction to the ions.
This energy contribution is called the exchange and correlation energy .
The hydrogen bond index Hydrogen bond is even more subtle, but is responsible for the com-
plicated structure of a number of important materials such as ice and many organic materials e.g.
the spatial configuration of proteins owes much to the hydrogen bond.

3.2.4 Problems
Problem 3.2-1
Derive (3.10) from (3.7) and (3.9) and use the data in table (5.1) to verify the numbers in the 6’th
column of table (3.2).

Problem 3.2-2
The lattice parameters of KCl are given in table 5.1
(a) Calculate the Coulomb energy between a K+ and a Cl− ion at the nearest neighbor distance
in units of eV.
(b) Assume that the parameters σ and  of the van der Waals attraction between the ions
(the term proportional to 1/r 6 in the Lennard -Jones potential) are the same as for Ar (table 3.1).
30 3.2 Stress and strain

Calculate the van der Waals energy between a K+ and a Cl− ion at the nearest neighbor distance
of KCl. Compare with the result found under (a).

Problem 3.2-3
The cohesive energy of an ionic crystal can be approximated as
( )
αe2 C
U =N − + m
4πo r r

where N is the number of unit cells, each assumed to contain one positive and one negative
charge ±e. In the formula α is the Madelung constant (see table 3.3), r is the nearest neighbor
distance, and C and m are adjustable parameters. The cohesive energy of N aCl has be measured
to be U/N = −1.27 × 10−18 J, (For details of the NaCl structure see table 5.1 and figure 2.6).
(a). Use the above data to determine the constants C and m.
(b). Show that the bulk modulus can be determined from
∂P ∂ 2U (m − 1) αe2
B = −V =V =
∂V ∂V 2 18 4πo r 4
where V is the volume of a crystal containing N unit cells.
(c). Calculate the bulk modulus of NaCl.

3.3 Stress and strain

In this section we will consider solids which differs from perfect crystals in that the atoms are
displaced from their equilibrium positions by a small amount1 . We consider the stresses and
strains associated with static external forces, and postpone to section 4.1 a discussion of elastic
waves. One purpose in discussing the theory of elasticity is to lay the foundation for understanding
the strength of materials.
Consider a solid which is deformed by external forces. The position of a given atom in the
crystal will only differ from that of a perfect crystal by a small amount if the forces causing the
deformation are weak. We write

r=R+u

Here r is the actual position of an atom which was at R before the crystal was deformed. The
vector u(R) represents the displacement of an atom which used to be at R. If the displacement
vector u is constant the only thing which has happened is that the crystal has moved a distance
u, and there is no elastic energy associated with this. In elastic theory one is interested in how the
displacements change in space. We define the deformation tensor (or matrix)
   
∂ux ∂uy ∂uz
xx xy xz
  
∂x ∂x ∂x

 yx yy yz  =  
∂ux ∂uy ∂uz
 ∂y ∂y ∂y  (3.15)
zx zy zz ∂ux ∂uy ∂uz
∂z ∂z ∂z
1
An elementary, but more detailed discussion than is presented here can be found in Barber and Loudon [3]. For
a more advanced treatment of elasticity theory see Landau and Lifschitz [7].
Static properties 31

Equation (3.15) can be written more compactly as

∂uβ
αβ = (3.16)
∂xα
where xα = x, y or z and uβ = ux , uy or uz . Let x̂, ŷ, ẑ be unit vectors in three cartesian
directions inside a solid. After the deformation the new axes are given by the vectors x0 , y 0 and
z0 (figure 3.2(a)). The three new axes need no longer be perpendicular to each other, but we do
require the deformation to be small. After the deformation we have for the new x-axis

∂ux ∂uy ∂uz

x0 = (1 + )x̂ + ŷ + ẑ
∂x ∂x ∂x
and in general
    
x0 1 + xx xy xz x̂
 0    
 y  =  yx 1 + yy yz   ŷ  (3.17)
z0 zx zy 1 + zz ẑ

^z z’ u
y dφ

R
^
y φ
x
y’
^
x
x’
(a) (b)

Figure 3.2: (a) Distortion of a Cartesian coordinate system. (b) Rotation about the z-axis

It is conventional to define the strain tensor eαβ in terms of the symmetric part of the defor-
mation tensor
   
exx exy exz xx xy + yx xz + zx
   
 eyx eyy eyz  =  xy + yx yy yz + zy  (3.18)
ezx ezy ezz xz + zx yz + zy zz
The reason for this is that the anti-symmetric part of the deformation tensor is associated with
a rotation of the whole crystal and just as in the case of uniform translation there is no elastic
energy associated with a pure rotation. To see why, consider as an example a rotation of a vector
R in the x − y plane by a small angle dφ as shown in figure 3.2(b). Before the rotation the polar
coordinates of the vector were r,φ.
32 3.2 Stress and strain

R = x + y = r cos φx̂ + r sin φŷ

u = dR = −r sin φdφx̂ + r cos φdφŷ = −ydφx̂ + xdφŷ

The associated deformation matrix is
   
xx xy xz 0 dφ 0
   
 yx yy yz  =  −dφ 0 0 
zx zy zz 0 0 0

We see that the small rotation is associated with a deformation tensor whose only nonzero
components are yx = −dφ, xy = dφ. The result of the small rotation is thus an anti-symmetric
deformation tensor.
The notation in (3.18) is the one commonly used. It is somewhat awkward in that it might
have been more logical to define the strain tensor as
!
1 ∂uα ∂uβ
ηαβ = +
2 ∂xβ ∂xα

as is done in [7].
Consider next the shear deformation of figure 3.3 in which the x- and y-axes are deformed in
the opposite directions by a rotation by a small angle ∆. Using sin ∆ ' ∆, cos ∆ ' 1 we have

x0 = x̂ cos(∆) + ŷ sin(∆) ' x̂ + ∆ŷ

y 0 = ŷ cos(∆) + x̂ sin(∆) ' ŷ + ∆x̂

y’ y’
y^ ^y

∆
2∆

X’
∆
^
X ^ =X’
X
(a) (b)

Figure 3.3: Two distortions with the same shear strain.

We find xy = yx = ∆, exy = eyx = 2∆, i.e. the deformation is symmetric.
Static properties 33

If instead the deformation had been on the form depicted in figure 3.3(b), in which the x-axis
is held fixed but the y-axis is rotated clockwise by 2∆ we find

x0 = x̂

y 0 = ŷ cos(2∆) + x̂ sin(2∆) ' ŷ + 2∆x̂

We now have xy = 0, yx = 2∆, but the strain tensor is unchanged exy = eyx = 2∆. The two
configurations have the same strain, since, for small deformations, they only differ by a rotation by
the angle ∆.
Figure 3.4(a) depicts a situation in which there is a change of volume as a consequence of the
stretching of a crystal in the z-direction. Suppose a volume V is deformed into a volume V 0 . We
define the dilation as

V0−V
δ= (3.19)
V
The volume of the parallelepiped formed by three vectors a, b and c can be written (see figure
2.1(b))

ax ay az
a · (b × c) = bx by bz
cx cy cz

y σ
xy

σ σ
yx yx
^z

x
σ
^y xy

(a) (b)

Figure 3.4: (a) Dilation strain ezz = δ. (b) For equilibrium the shear stress σxy = σyx

and we find

1 + xx xy xz

V0 = 1+δ = yx 1 + yy yz
zx zy 1 + zz
34 3.2 Stress and strain

or

δ ' xx + yy + zz + terms of order 2

Only the diagonal components of the strain tensor gives rise to a volume change for small
distortions. A positive volume change is a dilation while a negative change is a compression.
Since the strain tensor is symmetric only six of its nine component are independent and it is
conventional to number the components:

e1 = exx = xx
e2 = eyy = yy
e3 = ezz = zz
(3.20)
e4 = eyz = yz + zy
e5 = exz = xz + zx
e6 = exy = xy + xy

In general one needs to apply a force in order to produce strain. Force per unit area is called
stress. We let σαβ represent the force in the α direction acting on a surface whose normal is in
the β direction. If the force is parallel to the surface (perpendicular to the normal) we have shear
stress (figure 3.4(b)), while if the force is parallel to the normal (perpendicular to the surface) we
are dealing with tension or compression, (figure 3.5). The diagonal components of the stress tensor
represents tensile forces when positive, compaction forces when negative.

σ σ
xx xx

Figure 3.5: Tensile stress

For a material in equilibrium the forces acting on a volume element must be balanced and there
can be no net torques. In the case of shear stresses this will give rise to a double couple situation
as depicted in figure 3.4(b). The consequence is that for equilibrated stresses the stress tensor will
be symmetric, i.e. σαβ = σβα . Of the 9 components of the strain tensor only 6 are thus distinct. In
analogy with (4-8) we write for these components
Static properties 35

σ1 = σxx
σ2 = σyy
σ3 = σzz
(3.21)
σ4 = σyz = σzy
σ5 = σxz = σzx
σ6 = σxy = σyx
In the theory of elasticity one assumes that stresses and strains are proportional to each other
according to Hooke’s law. Complications arise from the fact that a given component of stress may
affect more than one strain component and vice versa. For example, if a beam is elongated due to
tensile stress the cross sectional area will also change. The most general form of Hooke’s law is

X
6
σi = Ca ej (3.22)
j=1

The coefficients Cij are called elastic constants.

3.3.1 Elastic constants for a cubic crystal

There are 36 possible elastic constants, but they will not be independent because of symmetry
requirements. The number of independent elastic constants will depend on the crystal structure
and we refer to [7] for details on this point. We will here limit our attention to crystals of cubic
symmetry and argue that there are then 3 independent elastic constants.
To be specific the [100],[010] and [001] directions are equivalent in a cubic crystal. We must
thus have C11 = C22 = C33 , C12 = C21 = C13 = C31 = C23 = C32 and C44 = C55 = C66 . All
the other off-diagonal elastic constants will be zero. To see that C36 = 0, assume that a crystal is
strained so that e6 = exy 6= 0. A nonzero C36 would mean that there is a resulting tensile stress in
the z-direction. If we relabel the axes so that x̂ → ŷ, ŷ → −x̂ we see that in the new coordinate
system the e4 strain will change sign (figure 3.6). If C36 6= 0, the σzz stress would also change
sign, say, from tension to compression. However, this is clearly impossible since the crystal hasn’t
changed. Similar arguments can be used to show that the remaining off-diagonal elastic constants
are zero. To summarize: for a cubic crystal Hooke’s law takes the form
    
σ1 C11 C12 C12 0 0 0 e1
 σ2   C12 C11 C12 0 0 0  e2 
    
    
 σ3   C12 C12 C11 0 0 0  e3 
 =    (3.23)
 σ4   0 0 0 C44 0 0  e4 
    
 σ5   0 0 0 0 C44 0  e5 
σ6 0 0 0 0 0 C44 e6

Cubic crystal under uniaxial compression strain

As an example consider a crystal with cubic symmetry which is put in a vise so that there is a
uniform compression in the [11̄0] direction (figure 3.7). The dilation of the crystal is from the figure
given by -δ. Let us find the components of the strain and stress tensor. We find xx = yy = − δ2 ,
xy = yx = 2δ , so that the strain tensor becomes
36 3.2 Stress and strain

y’
^
y ^x
x’

x’
-x^ y^

x^ ^
-x’ y’
(a) (b)

Figure 3.6: Relabeling the axes

 
−δ δ 0
 2 
e= δ − δ2 0 
0 0 0

y^

δ /2
x^

Figure 3.7: Compression perpendicular to the (11̄0) plane

giving e1 = − δ2 , e2 = − δ2 , e3 = 0, e4 = 0, e5 = 0, e6 = δ. Substitution into (3.23) gives

σ1 = σxx = − δ2 (C11 + C12 )

σ2 = σyy = − δ2 (C12 + C11 )
σ3 = σzz = −δC12
(3.24)
σ4 = σyz =0
σ5 = σxz =0
σ6 = σxy = δC44
Static properties 37

Note that while there is no strain component in the z-direction there will be a compressional
stress.

3.3.2 Elastic constants of an isotropic material

If we have the stress components in one coordinate system we can find the stresses acting on any
lattice plane by the rules of vector addition, taking into account the fact that stress is force per
unit area. E.g. when the Cartesian components of the stress is given by (3.24), the stress normal
to the (11̄0) plane is,

1 1 δ
σn = √ √ (σxx − σxy + σyy − σyx ) = − (C11 + C12 + 2C44 ) (3.25)
2 2 2
where the first factor of √1
comes from taking the component of a force in the x−direction (or
2 √
y− direction). The second factor comes that the (11̄0) plane of a unit cube has area 2. There
will be no shear stress on this plane in our example.
Consider next a pure compression δ in the [100] direction, e1 = −δ, e2 = e3 = e4 = e5 = e6 = 0.
We find for a cubic material

σ1 = −δC11 , σ2 = −δC12 , σ3 = −δC12 , σ4 = σ5 = σ6 = 0. (3.26)

A polycrystalline or amorphous material will be isotropic. For an isotropic system we must
require that if the dilation is the same, the normal stress is the same, irrespectively of the plane
chosen for the application of the strain. By comparing (3.25) with (3.26) we then find for an
isotropic system

C11 = C12 + 2C44 (3.27)

It is conventional to describe an isotropic system in terms of the the two Lamé constants

λ = C12 , µ = C44 (3.28)

From (3.27) an isotropic system must have C11 = λ + 2µ. We conclude that Hooke’s law for an
isotropic system takes the form
    
σ1 λ + 2µ λ λ 0 0 0 e1
 σ2   λ λ + 2µ λ 0 0 0  e2 
    
    
 σ3   λ λ λ + 2µ 0 0 0  e3 
 =    (3.29)
 σ4   0 0 0 µ 0 0  e4 
    
 σ5   0 0 0 0 µ 0  e5 
σ6 0 0 0 0 0 µ e6

Uniaxial stress of an isotropic solid.

An isotropic solid is subject to uniaxial tension T in the x−direction. We wish to find the strain
components.
If σ1 = T, σ2 = σ3 = σ4 = σ5 = σ6 = 0 we find the following equations for the strains
38 3.2 Stress and strain

T = (λ + 2µ)e1 + λe2 + λe3

0 = λe1 + (λ + 2µ)e2 + λe3 (3.30)
0 = λe1 + λe2 + (λ + 2µ)e3
in addition to e4 = e5 = e6 = 0. The solution to (3.30) is

T (λ + µ)
e1 =
µ(3λ + 2µ)

−T λ
e2 = e3 = (3.31)
2µ(3λ + 2µ)
The Poisson ratio of an isotropic material is defined as

e2 λ
− = (3.32)
e1 2(λ + µ)
The Young’s modulus is defined as the ratio of an uniaxial stress T and the tensile strain
component along T . If the stress is acting in the x−direction we have Y = T /e1 . We find that

µ(3λ + 2µ)
Y = (3.33)
λ+µ
for an isotropic material.

3.3.3 Problems
Problem 3.3-1:
An isotropic material with elastic constants λ and µ is put under uniform hydrostatic pressure
(σ1 = σ2 = σ3 = −P, σ4 = σ5 = σ6 = 0). Show that the bulk modulus is

∂P 2
B = −V = λ+ µ (3.34)
∂V 3

Problem 3.3-2:
A cubic crystal with elastic constants C11 , C12 and C44 is put under uniform hydrostatic pressure
(σ1 = σ2 = σ3 = −P, σ4 = σ5 = σ6 = 0). Assume that the stresses act normal to the (100), (010)
and (001) planes. Find the bulk modulus.

Problem 3.3-3:
Find the Young’s modulus of a cubic crystal with elastic constants C11 , C12 and C44
(a) if uniaxial tension is applied in the [100] direction?
(b) if uniaxial tension is applied in the [110] direction?
(c) if uniaxial tension is applied in the [111] direction?
Check that your formulas reduce to (3.33) if the elastic constants satisfy (3.27)(3.28).
Static properties 39

Problem 3.3-4:
A polycrystalline material has Young’s modulus Y = 2.5 × 1010 N m−2 and Poisson ratio 0.46.
(a) What are the elastic constants λ and µ?
(b) What is the bulk modulus?
(c) A person weighing 75kg stands on a 1cm3 sample shaped like a cube, what is the change
in volume? what is the change in height?
(d) A sample of the material is dumped into the ocean and lands at the depth of 1000m. What
is the fractional change in volume?

3.4 Defects
So far, we have been concerned with perfect crystals in which the unit cell is repeated without
error in all directions throughout space. Any real crystal will contain imperfections. Many bulk
properties, such as the elastic behavior, discussed in the previous section, and the thermal properties
to be discussed later are relatively insensitive to a small amount of imperfections. Other properties,
such as the strength of materials, and electric conduction, are strongly affected by defects.

3.4.1 Point defects

Figure 3.8 illustrates four different types of point defects. Vacancies are simply missing atoms or
ions. If the removed atom or ion is allowed to migrate to the surface we have a Schottky defect.
On the other hand, if the missing atom is stuck in an interstitial position we have a Frenkel defect.
Another common form of defect, is an impurity. A substitutional impurity simply consists of a
foreign atom replacing a host atom on one of the lattice sites. This is the common form of impurity
if the guest and host atoms are of comparable size. If the impurity atom is much smaller than
the host it can sometimes find room in the empty space between the atoms, we then say that the
impurity is interstitial.
The point defects of figure 3.8 are always present in small amounts in a crystal at equilibrium.
We will here concentrate on vacancies. Let Ev be the vacancy formation energy, which for a
Schottky defect is the energy of removing an atom or ion and placing it on the surface of a crystal.
The equilibrium concentration nV of Schottky vacancies will then be approximately given by a
Boltzmann factor
NV
nV = = e−βEV
N
where NV is the number of vacancies, N the total number of sites in the crystal and β =
1
kB T . Typical vacancy formation energies for metals and ionic crystals are in the range 0.5-1.5
eV . At room temperatures and below the equilibrium vacancy concentration will be very low (at
room temperature β ' 40 eV −1 and e−40 ' 4. × 10−18 ). At elevated temperatures the vacancy
concentration will be much higher, e.g. at 1000K the Boltzmann factor is ' 10−5 .
In an ionic crystal the positive ion is often smaller than the negative counter ion. The vacancy
formation energy for the positive ion may then be significantly smaller than the corresponding
energy for the negative ion. If this leads to a charge imbalance a huge negative electrostatic
potential will develop in the interior of the crystal and prevent the escape of any more positive
ions. We can thus write
40 3.3 Defects

(a) (b)

(c) (d)

Figure 3.8: Point defects: (a) Schottky defect, (b) Frenkel defect, (c) Substitutional impurity, (d)
Interstitial impurity

E+ = + + eφ; E− = − − eφ

where E+ (E− ) is the true vacancy formation energy of a positive (negative) ion and + , and −
are the corresponding quantities at constant electrostatic potential, e is the ion net charge and φ is
the electrostatic potential difference between surface and bulk. This potential will be established
by a layer of surface dipoles called the Debye layer and as a consequence we have for the positive
(negative) ion vacancy concentration n+ (n− )
β
n+ ' n− = e− 2 (+ +− ) (3.35)

n+ n− = e−β(+ +− ) (3.36)

Equation (3.37) is a special case of the law of mass action . One way of introducing a larger
equilibrium vacancy concentration is by doping. Consider e.g. a K + Cl − crystal that contains a
concentration nCa of calcium impurities. The calcium ions are divalent and as a consequence we
must replace (3.35) by

n+ − n− = nCa (3.37)
Static properties 41

while (3.36) remains valid. If nCa is much larger than the equilibrium vacancy concentration
in the absence of doping (3.36) and (3.37) will give n+ ' nCa . One consequence of this is that
although a calcium ion is heavier than a potassium ion the density of a crystal will be reduced by
doping.
In practice, a crystal at room temperature and lower will often have a vacancy concentration
which is much higher than predicted by equilibrium theory. To understand how this comes about
we must consider vacancy diffusion. A vacancy will move to a neighbor site in a crystal if an atom
from that site jumps into the empty site. In order to do so it will, however, have to overcome an
energy barrier EB figure (3.9). We write for the jump rate p (probability per unit time that a
neighbor will jump into the empty site)

p = νe−βEB (3.38)
Here ν is the attempt frequency, the rate at which the atom or ion attempts to overcome the
barrier. This rate is of the order of a typical lattice vibration frequency (see section 4.1). The
Boltzmann factor exp(-βEB ) represent the probability that an attempt to jump over the barrier
will be successful.
Energy

ν
Vacancy
E
B

Position

Figure 3.9: Potential energy barrier for atom next to vacancy

We let n(x) be the number of vacancies per unit volume, and assume that the concentration
depends on a coordinate x in a direction normal to a family of lattice planes. The concentration of
vacancies per unit area in these planes will be a n(x) where a is the spacing between planes. The
particle current density, i.e the number of vacancies per unit area moving in the positive x direction
minus the number moving in the negative x-direction will then be

dn −βEB
j = −a2 νe (3.39)
dx
Conventionally one writes

dn
j = −D (3.40)
dx
where D is the diffusion constant. We thus have

D = νa2 e−βEB (3.41)

42 3.3 Defects

A typical value for the factor νa2 for a metal or ionic crystal is 10−6 m2 /s. The energy barrier
EB will normally be a few tenths of an electron volt. At room temperature the Boltzmann factor
exp(−βEB ) may then be of the order 10−16 , yielding a diffusion constant of the order 10−22 m2 /s.
You will have learned in other courses that the typical time t it takes to travel a distance x will be
√
x' Dt

If x is 1mm = 10−3 m, we find t ' 1016 s, which is a significant fraction of the age of the universe.
On the other hand if we increase the temperature from room temperature to 1000K the diffusion
time will be reduced to times of the order days. Typically a crystal will be grown from a melt of
the order 1000K, at these temperatures the vacancy concentration will be quite high. The extent
to which the vacancy concentration will have time to reach its equilibrium value, as a crystal is
cooled down, will then depend critically on the cooling rate.

3.4.2 Dislocations
The type of defect that influences the mechanical properties of a material the most is dislocations.
Imagine that we remove a half-plane of atoms in a family of lattice planes as shown in figure 3.10.
The edge of this half-plane is perpendicular to the plane of the figure. The associated line defect
is called an edge dislocation . In the immediate neighborhood of the edge there is considerable
distortion of the crystal, while further away the crystal appears almost regular. The presence of
the dislocation will still be felt far away. To see this, imagine the following “walk” on a lattice
plane perpendicular to the edge: go 8 lattice spacings W , then 8 lattice spacings S, then 8 spacings
E and finally 8 spacings N . If this walk takes place in an undistorted part of the crystal it would
take us back to the starting point, but if we go around an edge dislocation we would be one lattice
spacing short. The vector b by which the walk is short of returning is called Burger’s vector .

Figure 3.10: Edge dislocation

Static properties 43

A screw dislocation can be obtained by making a cut along a half-plane and shifting the parts on
either side by one lattice spacing (figure 3.11). After this distortion the lattice planes perpendicular
to the edge of the cut forms a helicoidal surface (somewhat analogous to the parking floors in a
parkade). Both in the case of edge and of screw dislocations, we refer to the region near the edge
of the cut, where the crystal is most distorted, as the core. If we make the same walk as we did
before, around the edge, we would be short one lattice spacing parallel to the edge of the cut. The
Burger’s vector of a screw dislocation is parallel to the dislocation core for a screw dislocation, and
perpendicular to it for an edge dislocation. In both cases the associated strain away from the axis
will be approximately 1 lattice spacing divided by the distance to the axis, i.e the strain field will
be inversely proportional to the distance to the dislocation core. Since the strain falls off so slowly
with distance the energy associated with dislocations will be large compared to vacancies. Thermal
energies are therefore too small to create dislocations. Once formed they can migrate, but not be
got rid of, without a major rearrangement of the atoms in the crystal. The pattern of dislocations
depends on the history of the crystal. It is almost impossible to grow a dislocation free crystal, and
further dislocations are formed in a crystal subject to bending and other forms of severe stain.

Figure 3.11: Screw dislocation.

3.4.3 Problems
Problem 3.4-1:

A N a+ Cl − crystal contains small amount of the radioactive isotope N a22 . The sodium ions in
N a+ Cl − occupy a fcc lattice. Assume that the N a22 ions can jump to one of the 12 nearest
neighbor positive ion site, only if it contains a vacancy, otherwise it is stuck. Assume that the
effective value for νa2 in the formula (3.40) for the diffusion constant is 10−6 m2 /s × the positive
ion vacancy concentration. Also assume that the jump activation energy EB is 0.5eV . Estimate
the diffusion constant for Na22 at 600K if
44 3.4 Strength of materials

(a). The vacancy concentration is e−β with  = 1eV .

(b). The crystal contains 1% Ca++ .
(c). How long would it typically take for a N a22 ion to diffuse 1mm in the two cases.

Problem 3.4-2:Dislocation pairs

Consider a crystal which contains two edge dislocations with cores parallel to the z-axis. One
cut-out half-plane extends from the origin in the positive x-direction. The second dislocation core
intersects the x − y plane at L(−x̂ + ŷ), where the distance L corresponds to many lattice spacings.
Find the Burger’s vector for a counter clockwise (seen from the positive z-axis) path around (i) the
first dislocation, (ii) the second dislocation, (iii) both dislocations if the second half-plane that has
been cut out extends
(a). in the negative x-direction
(b). in the positive y-direction.

3.5 Strength of materials

If a material is put under sufficient stress it will undergo permanent damage and the material will
not return to its original shape if the stress is removed. The theoretical shear strength of a perfect
crystal can be estimated by the following simple argument. Consider two nearby lattice planes in a
sheared crystal as seen in cross-section in figure 3.12. We assume that in the undistorted crystal the
lattice planes are directly above each other, and let x be the shear displacement of an atom in one
lattice plane relative to the next plane a distance d below. The lattice spacing in the x-direction is
a. The strain is exy = e6 = x/d. For small displacements the shear stress will be

Gx
σxy = Gexy = (3.42)
d
where in our case the shear modulus is G = C66 . If we increase the strain, the shear stress
cannot continue to increase linearly with the strain without limit. When the relative slip between
the two planes reaches a full lattice spacing we recover the periodic crystal and the stress is zero.
The stress will also be zero, corresponding to unstable equilibrium, when the successive planes are
slipped by one half lattice spacing as shown in figure (3.12). We conclude that the stress must
be a periodic function of the strain with period a. The simplest way to model this is to assume a
sinusoidal variation of the stress with the strain. We require that we recover (3.42) for small strains
and obtain

Ga 2πx
σxy ≈ sin( ) (3.43)
2πd a
Equation (3.43) predicts that there is a maximum shear stress σc beyond which the crystal will
yield. In our simple model this stress is given by

Ga
σc = (3.44)
2πd
Since a and d will be about the same size, (3.44) predicts that the shear modulus should at
most be one order of magnitude larger the the yield stress. This prediction is very far from the
Static properties 45

Gx/d

Shear stess σ
x d
a 2a

Displacement x

Figure 3.12: The shear stress is a periodic function of the strain.

truth. In table 3.4 we list some typical values of G and σc for aluminum samples of varying quality.
Similar values would obtain for other materials.
Material G N m−2 σc N m−2 G/σc
Single crystal Al 2.8×1010 4.0×105 60 000
Pure polycrystalline Al 2.5 × 1010 2.6 × 107 900
Commercial drawn Al wire 2.5 × 1010 3.8 × 107 250
Duralumin 2.5 × 1010 3.6 × 108 70

Table 3.4: Yield stress and shear modulus for aluminum

We note that not only is the disagreement between theory and experiment quite bad. More
embarrassing, considering that the theory was developed for a perfect crystal, is the fact that the
disagreement gets worse the better the quality of the crystal. It is now generally accepted that
shear slip is mediated by dislocations. Consider the edge dislocation of figure 3.10. If the edge is
slipped, one lattice spacing at the time, to the right, this is equivalent to a shear deformation of the
top part of the crystal, relative to the bottom part by one lattice spacing. Since it is only the edge
which is moving, and not the whole lattice plane, the required stress is much less than it would be
in a perfect crystal without dislocations.
An analogous situation occurs if one tries to pull a heavy carpet across the floor in an empty
room. Brute force pulling requires more strength than most of us possess. A much easier method is
to produce a wrinkle at one end of the carpet and push the wrinkle across. We can now understand
why σc increases in a crystal with an increasing number of imperfections. Dislocations can not
easily cross grain boundaries of a polycrystalline material. The strain produced by drawing an
aluminum wire will produce a tangle of dislocations that pin other dislocations. The process of
increasing the strength of a material through dislocation producing deformations is called work
hardening. The impurities of an alloy such as duralumin also act as pinning centers. Returning to
our carpet analogy, the wrinkle method of pulling a carpet does not work too well if the room is
full of furniture.
A ductile material can be subjected to plastic deformation, if the strain exceeds an elastic limit.
Figure 3.13 describes the outcome of a typical tensile test of a ductile material. In the test a
cylindrical sample is subject to uniaxial tension σ, and the strain component  = e11 is measured.
46 3.4 Strength of materials

We start with an unstressed sample (a). When the stress is increased to (b) the behavior is elastic,
the stress is proportional to the strain as predicted by Hooke’s law, and the sample will return to its
original shape if the stress is released. The point (b) is called the yield point. Beyond this point the
relationship between stress and strain is nonlinear. If the stress is released at (c) the sample will
not return to its original shape but there will be a plastic deformation. We now must distinguish
between two ways of registering the stress. The true stress is defined as

tensile f orce
σt =
actual area
and the engineering stress is given by

tensile f orce
σe =
original area

The dashed line in figure 3.13 represents the true stress, while the full line corresponds to the
engineering stress. The marked difference between the two curves arises because the reduction in
cross sectional area tends not to be uniform, but concentrated in a “neck”. If the stress is reapplied
at (d) the new yield point (e) will typically be somewhat higher than the old one, due to work
hardening. Finally there is a maximum stress which the sample can endure before it breaks.
Stress σ

True stress σt

Engineering stress σ
e
c

a d

Strain ε

Figure 3.13: Stress vs. strain for a ductile material.

Some materials are brittle and break suddenly rather than undergo plastic deformation. Im-
portant examples are glass and rocks. In brittle materials small voids called cracks develop when
a sample is stressed. In order to explain the importance of cracks we need to comment on how
nonuniform stress is distributed. Consider a medium which is subject to a spatially varying stress
consisting possibly both of a shear component and tensile stress. The requirement that forces
and torques are balanced, leads to a symmetric stress tensor which satisfies the partial differential
equation
Static properties 47

X ∂σij
=0 (3.45)
j=x,y,z
∂rj

where i = x, y or z. One can solve (3.45) for a spherical cavity. The boundary condition is
uniaxial tensile stress σ∞ far away, and zero normal component of the stress on the cavity surface.
At the top and bottom there will be a compressive stress equal to σ∞ . At the equator the stress
is tensile and enhanced to 3σ∞ . If the void is an oblate ellipsoid of revolution with the tensile
stress σ∞ applied parallel to the minor axis b, the stress enhancement at the equator is even more
pronounced and given by

2c
σeq = σ∞ (1 + )
b
where c is the major axis.

Mode I Mode II Mode III

Figure 3.14: The three crack propagation modes.

In the case of a long thin crack the stress enhancement can thus be enormous and we see that
once a crack opens up, the stress enhancement at the tip makes it tend to grow, drastically weak-
ening the strength of a sample. As shown in figure (3.14) there are three main crack propagation
modes. We note that the deformation associated with a mode II crack is the same as for an edge
dislocation, while a mode III crack is analogous to a screw dislocation. A complication is that the
material near the tip of a crack may be subject to chemical corrosion. As anyone who has had to
cut a glass pane to fit a frame would know, it is easier to crack glass when it is wet. Corrosion
processes can be very slow and this can give rise to the phenomenon of static fatigue. A window
pane subject to a steady stress may suddenly, for no apparent reason, decide to shatter. Similarly
a ductile material may become brittle if the dislocation density gets to high due to excessive work
hardening. This phenomenon is called dynamic fatigue. . For further reading on strength of ma-
terials see the book by Callister [1]. For a discussion of ductile materials see also the article by
Sprackling [10], a readable discussion of the properties of brittle materials can be found in Scholz
[9].
48 3.4 Strength of materials

We conclude this section by defining a few more different properties associated with the strength
of materials:

• Hardness is a measure of the resistance of a material to plastic deformations such as dents

and scratches. A somewhat arbitrary but often used index of hardness is the Mohs scale.
This scale is based on the idea that a harder material can scratch a softer material. The
hardest known material is diamond which is given an index of 10 on the Mohs scale. Next
are corundum and sapphire with 9, topaz with 8, quartz 7, orthoclase (a feldspar mineral)
6, apatite (calcium fluorophoshate) 5, fluorite 4, calcite 3, gypsum 2 and talc 1. There are
a number of more quantitative hardness measures such as the Brinell hardness number. In
a Brinell test a spherical indenter of tungsten carbide or hardened steel with a diameter of
typically 1cm is applied for a specified length of time and with a specified force. The diameter
of the indentation is then measured and converted to a hardness number using a chart. The
Rockwell hardness test is a similar method.

• Resilience is the ability of a meterial to return to its original shape after being bent, com-
pressed or stretched. The resilience modulus Ur is the energy per unit volume which is stored
and can be retrieved before the yield point (point b in figure 3.13). This energy is givent by
Z y 1
Ur = σd = σy dy
0 2

where σ and  are engineering stress and strain. A related property which is more difficult to
quantify is Toughness. It is defined as the abilty to absorb energy up to fracture or other
failure.

3.5.1 Problems
Problem 3.5-1: Weibull statistics
There are a number of situations, e.g. in problems involving fracture or failure where what matters
is not the average property of the system, but rather the behavior of the weakest link. An empirical
approach to such problems was taken by the Swedish engineer Waloddi Weibull.
Consider a fiber of length n in some unit. Assume a segment i of unit length has a yield
threshold stress σi which may depend on microscopic flaws and will vary from segment to segment.
We assume the probability that a segment breaks with a stress σ is given by

P (σ) = 1 − e−φ(σ)

The probability that the link will not break is then e−φ(σ) , where φ(σ) is some monotonically
increasing function of σ. If any one of the links fails the whole chain fails. Therefore, the probability
that the fiber survives is e−nφ(σ) and the probability that it will break under stress σ is

Pchain = 1 − e−nφ(σ) (3.46)

Weibull suggested the form

φ(σ) = a(σ − σ0 )ρ ; f or σ > σ0 (3.47)

Static properties 49

where σ0 is some minimum failure stress (which may be zero). The exponent ρ must be expected
to be a property of the fiber material (Weibull found ρ ≈ 3 for steel, and ρ ≈ 1.46 for cotton.).
The proportionality constant a in (3.47) will from (3.46) not be a material property, but will also
depend on the unit of length chosen and σ0 will depend on the manufacturing process.

(a). If length is measured in feet rather than meters, and stresses in lb per square inches
rather than N m−2 by what factor will the constant a change for steel? for cotton?
(b). A 10m long fishing line will withstand a certain force 99 times out of a hundred.
What is the probability that a 100m long fishing line of the same material can support
the same force?
(c). Plot the probability that (i) a steel and (ii) a cotton fibber will support the
stress σ in units where a = σ0 = 1. The length of the fiber is 1 in these units.
50 3.4 Strength of materials
Chapter 4

Electrons and phonons

We will in this chapter treat the properties of solids related to the motion of the atoms or ions and
of the electrons. The vibrations of atoms and ions in the lattice will be discussed mainly for their
importance in understanding the thermal properties of solids. In the case of the electrons we will
also be concerned with the electric and optical conductivity.

4.1 Lattice vibrations

In our discussion of the structure of solids we have assumed that the atoms (or ions) sit still at their
equilibrium positions. We now wish to turn to lattice vibrations. Our formal development will be
based on classical mechanics and the basic assumption of a harmonic lattice. At low temperatures
the theory must be modified to take into accounts quantum mechanical effects and we introduce the
concept of phonons. We then show that most thermal properties of solids can be explained by an
idealized model due to Debye. Two important effects which cannot be explained in the harmonic
approximation are thermal expansion and the lattice contribution to thermal conduction. We come
back to these effects towards the end of this section.

4.1.1 The harmonic approximation

In order to establish a manageable formalism we will make two important simplifying assumptions
which allow us to picture the vibrating lattice as a system of masses and springs. Firstly, since the
electrons in a solid are very light compared with the atoms, the electrons will respond very fast to
a disturbance of the ion positions. It is thus natural to make the approximation that the electrons
will adjust instantaneously to the ion position when the ions move. This implies that the potential
energy of the ions can be expressed as a function of the ion coordinates alone.
Let Ri be the equilibrium position of the i’th ion, ui the displacement of this ion from the rest
position. The actual position of the ion is then

ri = Ri + ui

We write uiα (α = x, y, z) for the components of the displacement. The components of the force
on the i’th ion is
∂V (r1 , r2 , ..ri, ..rN )
fiα = −
∂riα

51
52 4.1 Lattice vibrations

or in the more compact vector notation

∂V
fi = − (4.1)
∂ri
For small amplitude vibrations we can make a Taylor series expansion of the potential energy
function
X ∂V (r1 , r2, ..rN )
V (r1 , r2 , ..rN ) = V (R1 , R2 , ..RN ) + uiα +
iα
∂riα riα =Riα f or all i

1 X ∂ 2 V (r1 , r2, ..rN )

+ uiαujβ +··· (4.2)
2 ijαβ ∂riα∂rjβ riα =Riα f or all i

When all the ions are at their equilibrium positions, the net force (4.1) on them is zero. The
second term on the right hand side of (4.2) must therefore vanish. The second assumption necessary
to establish the harmonic approximation consists of stopping at the last term on the right hand
side of (4.2). This allows us to think of the lattice can be thought of as made up a set of masses
connected by springs. Each spring is represented by a spring constant

∂ 2 V (r1 , r2 , · · · rN )
Φijαβ = for all i (4.3)
∂riα∂rjβ riα =Riα
The equation of motion for the α’component of the displacement of the i’th atom is then
according to Newton’s second law.
X
müiα = fiα = − Φijαβ ujβ (4.4)
jβ

Because the motion is three dimensional, the notation is a bit cumbersome, and it is instructive
to consider some simple special cases before we return to the general case (4.4).

4.1.2 Harmonic chain

Consider a long chain consisting of N masses m connected by springs with spring constant K (figure
4.1). Let a be the equilibrium spacing between the masses, so that x = na is the equilibrium position
of the n’th mass. The actual position of the n’th mass is

rn = na + u(n)

2 [u(n + 1) −
and the potential energy of the spring connecting the n’th and the (n + 1)’st mass is K
u(n)]2 . We assume periodic boundary conditions, i.e. we let the N ’th mass be connected with the
first mass. Formally, we do this by requiring that u(N + 1) = u(1) and in general

u(N + n) = u(n) (4.5)

The force on the n’th mass is K[u(n + 1) + u(n − 1) − 2u(n)] and Newton’s second law gives
rise to a coupled set of differential equations

mü(n) = K[u(n + 1) + u(n − 1) − 2u(n)] (4.6)

Electrons and phonons 53

We can find solutions to (4.6) on the form

u(n) = Aei(nka−ωt) (4.7)

where k is the wave vector. Substitution of (4.7) into (4.6) yields
n o
−mω 2 Aei(nka−ωt) = AK ei[(n+1)ka−ωt] + ei[(n−1)ka−ωt] − 2ei[nka−ωt]

or

u(n)
K m

a +u(n) - u(n-1) a +u(n+1) - u(n)

Figure 4.1: The harmonic chain

2K
ω2 = [1 − cos(ka)]
m
We use the trigonometric identity

ka
2 sin2 ( ) = [1 − cos(ka)] (4.8)
2
By convention the frequency ω is positive and we find
s
K ka
ω=2 sin (4.9)
m 2
The periodic boundary condition (4.5) determines which values of k are allowed. Substitution
of (4.7) into (4.5) gives
eikN a = 1
or k = 2πl 0
N a with l an integer. From (4.7) we see that l = l + N and l give rise to identical solutions,
and we can without loss of generality restrict l to the values

N N N
l=− , − 1, · · · 1, 2, .. − 1 (4.10)
2 2 2
From (4.10) we note that there are N distinct values of l. This result is compatible with the
fact that the system has N degrees of freedom. The restriction on the wave vector is
π π
− ≤k< (4.11)
a a
54 4.1 Lattice vibrations

Κ
2 m

k
π π
-a a

Figure 4.2: The harmonic chain

We note that G = 2π a is a reciprocal lattice vector of the one dimensional lattice consisting to the
chain atoms being located in their equilibrium positions. The restriction (4.11) is then equivalent
to saying that k is within the Brillouin zone. The relationship (4.9) between frequency and wave
vector is plotted in figure (4.2)
The physical displacement is, of course, real. Since the differential equation (4.6) is linear
and real, both the real and imaginary part of a complex solution are also solutions. With these
considerations the general solution to (4.6) is then with A, B, Cl and Sl arbitrary constants of
integration
N
X
2
−1  
2πln 2πln
u(n) = Cl cos( − ωl t) + Sl sin( − ωl t) + A + V t (4.12)
N N
l=− N
2
,l6=0

The case l = 0 in (see equation 4.12) requires special attention. If l = 0 the wave vector k = 0.
This means that all displacements u are the same. The differential equation (4.6) now becomes

mü(n) = 0

with solution u = A + V t where A and V are constants. Usually the the situation where the
chain moves uniformly (as a whole) is not of much interest in lattice vibrational problems, but it is
important when considering the recoil in the decay of a radioactive nucleus in the lattice (as in the
Mössbauer effect), or in particle detectors analyzing or particles from cosmic rays or high energy
accelerators.
It is instructive to solve the harmonic chain problem with various boundary conditions. First,
consider the situation where we ”pluck” the chain by displacing the atom at the origin by a fixed
amount δ.
The initial conditions are then

u(0, t = 0) = δ; u(n, t = 0) = 0 f or n 6= 0 (4.13)

Electrons and phonons 55

u̇(n, t = 0) = 0 f or all n (4.14)

It is convenient to use the complex form for the normal mode expansion
N/2−1
X
u(n) = Al ei(2πnl/N −ωt) + A + V t
l=−N/2,l6=0

giving
N/2−1
X
u(n, t = 0) = Al ei(2πnl/N ) + A (4.15)
l=−N/2,l6=0

N/2−1
X
0 = u̇(n, t = 0) = −iωAl ei(2πnl/N ) + V (4.16)
l=−N/2,l6=0

The fundamental formula for inverting discrete Fourier series is

−1
( 0
X
N 1−e2πi(l−l )
= 0; f or l 6= l 0
2πi(l−l0 )n/N 1−e2πi(l−l0 )/N
e = (4.17)
n=0 N; f or l = l 0
0
Equating (4.13) and (4.15) and multiplying by eiπl n/N and summing over n gives

δ
A = Al =
N
i.e. al modes are excited with equal amplitude. The actual motion is the real part of the solution
giving
 
N/2−1
δ  X
u(n) = cos(2πnl/N − ωt) + 1 (4.18)
N l=−N/2,l6=0

Next consider the case where the atom at the origin is given a ”kick” so that it has velocity v
(in units of lattice spacings per unit time) starting at the origin with all the other atoms at rest.
The initial condition is now

u̇(0, t = 0) = v; u(n, t = 0) = 0 f orn 6= 0 (4.19)

u(n, t = 0) = 0 f or all n (4.20)

We find equating (4.16) and (4.19) using the fourier transform formula (4.17)
v
V =
N
and
iv
Al =
ωN
56 4.1 Lattice vibrations

The actual motion of the masses is obtained by taking the real part or
 
N/2−1
v  X 1
u(n) = − sin(2πnl/N − ωt) + t (4.21)
N ω
l=−N/2,l6=0

Returning to the general case: we refer to the individual terms in the [ ] of 4.12 as normal modes
and to (4.11) as a normal mode expansion.
From (4.9) we note that there is a maximum frequency
s
K
ωmax = 2
m

If an atom is subject to a periodic force with frequency ω < ωmax this will set up traveling
waves with wave vectors satisfying (4.9). The phase velocity of this wave is
s
ω 2 K ka
vp = = sin( )
k k m 2

while the group velocity is

s
∂ω K
vg = =a cos(ka)
∂k m

Note that the group velocity → 0 as ω → ωmax . A local disturbance with frequency > ωmax
cannot propagate through the lattice, but stays trapped. In the long wave length limit k → 0, λ =
2π/k → ∞ the phase and group velocities approach a common limit
s
K
s=a
m

We refer to long wave length lattice vibrations as sound waves with s the speed of sound.

4.1.3 Diatomic chain

We next consider a slightly more complicated case in which there are two atoms per unit cell with
mass m and M , respectively. The equilibrium size of the unit cell is a and nearest neighbors are
connected by springs with spring constant K (see figure 4.3). The equations of motion for the two
masses are

mü1 = K[u2 (n) + u2 (n − 1) − 2u1 (n)]

M ü2 = K[u1 (n + 1) + u1 (n) − 2u2 (n)] (4.22)

The amplitude of vibration for the two masses will now be different and we look for solutions
of (4.22) on the form

u1 = 1 ei(kna−ωt)
Electrons and phonons 57

M u (n) u (n)
1 2
K m

a/2+u2(n-1)-u1(n-1) a/2+u1(n)-u2(n-1)

Figure 4.3: The diatomic harmonic chain

u2 = 2 ei(kna−ωt) (4.23)
Substitution of (4.23) into (4.22) gives

(mω 2 − 2K)1 + K(1 + e−ika )2 = 0

K(1 + eika )1 + (M ω 2 − 2K)2 = 0 (4.24)

For (4.24) to have a solution we must have

mω 2 − 2K K(1 + e−ika )
= mM ω 4 − 2K(m + M )ω 2 + 2K 2 [1 − cos(ka)] = 0
K(1 + eika ) M ω 2 − 2K

Using (4.8) we find the solutions

s
K 4M m ka
ω = (1 ±
2
1− sin2 ( )
µ (M + m)2 2

mM
where µ = m+M is the reduced mass. The allowed values of the wave vector k are the same as
for the case of the monatomic chain (4.10),(4.11). For each allowed wave vector there are now two
frequencies.
The top branch of figureq 4.4 is called optical and the bottom branch acoustic. The highest
allowed frequency ωmax = 2K π
µ occurs for k= 0. At the zone boundary k= a the frequencies are
given by

K M −m
ω2 = (1 ± | |)
µ M +m
q q
2K 2K
and the two roots are ω = M and ω = m . Suppose M > m. We see that there is an
intermediate band of frequencies
s s
2K 2K
<ω<
M m
58 4.1 Lattice vibrations

ω
M
=2
Optical branch m

Acoustical branch

k
π π
- a a

Figure 4.4: The frequency vs. wave-vector for the diatomic chain.

for which no lattice wave can propagate. Finally let us consider the acoustical branch near
k = 0. We find
s
Ka2
ω'k
2(M + m)

and the sound velocity is given by

s
Ka2
s=
2(M + m)

Once the frequencies are found one can find the ratio 1 /2 and thus get a qualitative idea of
the nature of the vibrations. Simple calculations show that for the acoustic branch near k = 0
1 ' 2 i.e. the light and heavy atom vibrate together. As one moves up the acoustic branch the
amplitude 1 of the lighter atom is reduced and at the Brillouin zone boundary only the heavier
atom moves. Along the optical branch near k = πa only the lighter atom moves while near k = 0
neighboring atoms move in opposite directions in such a way that the center of mass is stationary.

4.1.4 Lattice vibrations in three dimensions

We now return to the general problem (4.3) (4.4). If the crystal potential is known, e.g. if the atoms
interact pairwise via a Lennard–Jones potential (3.2) , the force constants (4.3) can be computed by
straightforward if tedious algebra. Suppose we are dealing with a crystal in which the equilibrium
positions of the atoms form a Bravais lattice. In analogy with (4.7) we assume a solution to (4.4)
on the form

uj = ~ei(k·Ri −ωt) (4.25)

From the translational symmetry of the crystal it is clear that the force constant (4.3) is a
function of the distance s = Rj − Ri .
Electrons and phonons 59

Φijαβ = Φαβ (s)

Substitution of (4.25) into (4.4) gives

X
−mω 2 α = − Φαβ (s)eik·sβ
s,β

The Fourier transform of the force constant matrix Φ is the dynamical matrix
X
Dαβ (k) = Φαβ (s)eik·s
s

The determination of the normal mode frequencies involves solving the determinantal equation

Dxx(k) − mω 2 Dxy (k) Dxz (k)

Dyx (k) Dyy (k) − mω 2
Dyz (k) =0 (4.26)
Dzx (k) Dzy (k) Dzz (k) − mω 2
The dynamical matrix will be real since if s is a Bravais lattice vector so is −s. From (4.3)
we see that the dynamical matrix will be symmetric. It is a theorem of linear algebra that a real
symmetric matrix has positive eigenvalues and that the eigenvectors ~ will be orthogonal to each
other. The cubic equation which results when one multiplies out the determinant in (4.26) will thus
have three positive roots for ω 2 . If we take the square root of ω 2 we thus get real frequencies. One
can show that for small values of k one of the eigenvectors ~ will be parallel to k, i.e. the vibrations
will be longitudinal. The two other modes will have ~ perpendicular to k and the vibrations will be
transverse. In the general case the amplitude vector will not be exactly parallel or perpendicular
to k, but the three vectors ~ corresponding to a given k will be orthogonal to each other.
The restrictions on the allowed values of k are analogous to what we had in the case of the
linear chain. It is clear from (4.25) that adding a reciprocal lattice vector to k will not lead to any
change in the displacement vector u. We thus restrict k to the Brillouin zone . The allowed values
of k inside the Brillouin zone depends on the boundary conditions. Suppose a crystal is made up
of L unit cells in the directions of each of the three primitive lattice vectors a, b and c, and we
apply periodic boundary conditions . A little reflection should convince the reader that the allowed
k-values can be written
n1 n2 n3
k= A+ B+ C (4.27)
L L L
where n1 , n2 and n3 are integers and A, B and C are the reciprocal lattice vectors. Just as we
found in connection with our discussion of (4.10) the total number of allowed k-vectors inside the
Brillouin zone will be L × L × L = N, where N is the total number of unit cells. There will thus be
3N normal modes, which is expected from the fact that each atom can vibrate in three different
directions.
If we are dealing with a lattice which is not a Bravais lattice, but a Bravais lattice with a basis,
the situation becomes more complex and we will skip the details. In analogy with what we did
for the diatomic chain we will have an expression on the form (4.25) for each atom in the basis.
For a lattice with a p-point basis the diagonalization of the resulting dynamical matrix will give
60 4.1 Lattice vibrations

rise to a 3p × 3p determinant which must be put equal to zero. In analogy with what we found
for the diatomic chain 3 of the normal modes will be acoustic (i.e. the frequency approaches zero
in the long wave length limit) and the remainder optical (with a non-zero frequency in the long
wavelength limit).

4.1.5 Density of states

In what follows we will need to calculate thermal averages which involves summing over contri-
butions from all the individual modes. For this purpose we need to find a way to convert sums
over allowed k-values into a volume integral over d3 k. The volume of the primitive unit cell of the
reciprocal lattice is from (2.10)

(2π)3
|A · (B × C)| = (4.28)
vc
where vc is the volume of the primitive unit cell of the direct lattice. We have for the total
volume

V = L3 vc

Consider the modes for which the integers n1 , n2 and n3 in (7-22) are between
n1 and n1 + dn1
n2 and n2 + dn2
n3 and n3 + dn3
The total number of such modes is

dn1 dn2 dn3 = d3 n

From (4.27) and (4.28) we see that these modes occupy a ‘volume’ in k-space

(2π)3 3
d3 k = d n
V
If the volume V is macroscopically large the k-values for successive normal modes are very close
together and we can replace a sum of allowed k-states by an integral according to
X Z
V
= d3 k (4.29)
k
(2π)3
Equation (4.29) will prove quite useful later on.

4.1.6 Phonons
We treat the normal modes of the lattice as a set of independent harmonic oscillators. According
to quantum mechanics the energies of the individual oscillators are quantized with energy levels
1
En = (n + )h̄ω (4.30)
2
where n can take on integer values 0,1,2... , h̄ is Planck’s constant divided by 2π and ω is the
normal mode frequency. The energy of the lowest state, the ground state energy, is 12 h̄ω and an
Electrons and phonons 61

excitation with energy h̄ω above the ground state energy is called a phonon. The probability p(n)
that an oscillator will be in the n’th state will be proportional to the Boltzmann factor

e−βEn = e−β(n+ 2 )h̄ω

1

1
where β = kB T . The probabilities have to add up to 1 so that

1 −β(n+ 1 )h̄ω
p(n) = e 2
Z
where
∞
X e− 2 βh̄ω
1
−β(n+ 12 )h̄ω
Z= e =
n=0
1 − e−βh̄ω

The average energy stored in each mode is then

∞
X ∞
1 X 1 1 ∂Z
(n + )h̄ωe−β(n+ 2 )h̄ω = −
1
< E >= En p(n) =
n=0
Z n=0 2 Z ∂β

Straightforward algebra gives

h̄ω h̄ω
< E >= + βh̄ω
2 e −1

Each mode will contribute a term

∂<E> ∂ < E > ∂β 1 (h̄ω)2 eβh̄ω

= =
∂T ∂β ∂T kB T 2 (eβh̄ω − 1)2

to the specific heat. At high temperatures β → 0, and

h̄ω
eβh̄ω → 1 + βh̄ω = 1 +
kB T

and the contribution to the specific from each mode approaches kB . If the crystal has N atoms
there are 3N normal modes. We thus find that the heat capacity in the high temperature limit will
be close to 3NkB . This result is known as the law of Dulong and Petit.
In general, the lattice heat capacity will be the sum of the contributions from the individual
modes. We find

Z
1 X [h̄ω(k, α)]2eβh̄ω(k,α) V 1 X 3 [h̄ω(k, α)]2eβh̄ω(k,α)
CV = = d k (4.31)
kB T 2 k,α (eβh̄ω(k,α) − 1)2 (2π)3 kB T 2 α (exp[βh̄ω(k, α)] − 1)2
BZ

Here, ω(k, α) is the frequency of one of the three modes with wave vector k. The integral in
(4.31) is over the Brillouin zone.
62 4.1 Lattice vibrations

4.1.7 Debye model

We next describe a simplified model which allows us to evaluate (4.31), to a good approximation.
In the Debye model one makes the following assumptions:

(1). The frequency of a normal mode with wave vector k is approximated as

ω(k, α) = sk (4.32)
where s is an average sound velocity.
(2). The Brillouin zone is approximated by a sphere. The radius qD of the sphere
is given by the requirement that the number of allowed wave vectors inside the sphere
should be equal to the number of atoms N
4π 3 V
qD =N
3 (2π)3
or

N 1/3
qD = (6π 2 ) (4.33)
V
The maximum phonon frequency in the Debye model is called the Debye frequency

ωD = sqD (4.34)
We also define the Debye temperature as

h̄ωD
θD = (4.35)
kB
We substitute (4.32) into (4.31)
ZqD
3V 4π h̄2 s2 k2 eβh̄sk
CV = k2 dk
(2π)3 kB T 2 (eβh̄sk − 1)2
0

(The factor of 3 comes from the sum over α.). We introduce the dimension less variable x = βh̄sk
and find
Z D
βh̄sq
3V kB 2 T 3 x4 ex dx
CV =
2π 2 h̄3 s3 (ex − 1)2
0

which simplifies, using (4.33) (4.34) and (4.35), to

θZ
D /T
T x4 ex dx
CV = 9N kB ( )3 (4.36)
θD (ex − 1)2
0

Equation (4.36) can be evaluated analytically in some limits. At low temperatures θD /T → ∞.

The integral in (4.36) can the be evaluated analytically to
Electrons and phonons 63

Element θD [K] Compound θD [K]

Li 335 N aCl 280
Na 156 KCl 230
K 91 CaF2 470
Cu 343 LiF 680
Ag 226 Quartz 255
Au 162
Al 428
Ge 378
Si 647
Diamond 1860

Table 4.1: Debye temperatures of some elements and compounds

Z∞
x4 ex dx 4π 4
=
(ex − 1)2 15
0

and we get
 3
12π 4 T
lim CV = N kB (4.37)
T →0 5 θD

CV
3
Nk B
2.5

1.5

0.5

0.2 0.4 0.6 0.8 1 1.2 1.4

T/TD

Figure 4.5: Heat capacity in the Debye model.

i. e. the low temperature specific heat will be proportional to T 3 . For high temperatures we
note that for small values of x the integrand in (4.36) is ' x2 . The integral is then ' 13 (θD /T )3
and we recover the Dulong and Petit heat capacity. Finally, we not that if we plot the specific heat
per mol vs. the temperature in units of the Debye temperature, (4.36) predicts a universal curve
64 4.2 Sommerfeld model for electrons

(see figure 4.5). This prediction is remarkably accurate. To get the good agreement the Debye
temperature should be considered as a fitting parameter, i.e. the value of θD was chosen which
gave the best overall fit. Some fitted Debye temperatures are tabulated below.

4.1.8 Problems
Problem 4.1-1:
Consider a linear chain of atoms. Each atom interacts with its nearest neighbor on either side via
a Lennard-Jones potential (3.2). Assume parameter values appropriate to krypton (tables 3.1and
5.2).
(a). Find the equilibrium spacing between the atoms.
(b). Find the sound velocity.
(c). What is the maximum frequency?

Problem 4.1-2:
Plot the state of a monatomic chain of 100 atoms with periodic boundary conditions that has been
”kicked” or ”plucked” by evaluating (4.18) and (4.21) for a sequence of times. Give a physical
interpretation of the evolving pattern, by making use of the fact that the sound velocity increases
with increasing wavelength.

4.2 Sommerfeld model for electrons

In order to find the electronic states in a crystal one must solve the Schrödinger equation for the
electrons in the presence of a potential with the periodicity of the lattice. The resulting electronic
band structure calculation represents a problem which is beyond this course. Instead we will address
an idealized problem in which the electrons experience a constant potential. This is the so called
Sommerfeld model, and it turns out to be sufficient to explain a number of qualitative properties
of metals.

4.2.1 Electron gas

The volume of the box is V = L3 . First consider the simplest case of periodic boundary conditions.
The wave function of the particle satisfies

ψ(x = 0) = ψ(x = L).

The quantum eigenstates of the system are then given by

1 1
ψ(r) = √ eik·r = √ exp(i[kxx + ky y + kz z])
V V

where
2πnx 2πny 2πnz
kx = , ky = , kz =
L L L
Electrons and phonons 65

and nx , ny , nz can take on the values 0, ±1, ±2... i.e. are integers. The x-component of the
momentum is px = h̄kx = hnx /L and we have similar expressions for the other components.
Consider now a volume in phase space W = L3 ∆px ∆py ∆pz where the components are in the
intervals

p1 < px < p1 + ∆px

p2 < px < p2 + ∆py (4.38)

p3 < pz < p3 + ∆pz

We find that the number of states in W is given by

L3 ∆px∆py ∆pz W
NW = = 3 (4.39)
h3 h
The counting rule (4.39) must be applied with some discretion. Suppose we change the boundary
conditions and consider a box with hard walls. The boundary condition is ψ = 0 at surface. The
eigenstates are now
r
8
ψ= sin(kxx) sin(ky y)sin(kz z)
π
with
πnx πny πnz
kx = , ky = , kz = (4.40)
L L L
Note the absence of the factor of 2 in (4.40). The number of states in the interval (4.38) is then

23 ∆px ∆py ∆pz L3 8W

3
= 3
h h
The reason for this apparent paradox is that the ‘standing wave’ states associated with the hard
wall boundary condition must be considered to be mixtures of states propagating back and forth.
The states kx and -kx are not distinct. So while
# of states
= 8 times larger
V olume of phase space

only 1/8 th of the momentum states are distinct.

4.2.2 Free electron gas at zero temperature

At zero temperature the system will be in its lowest energy state, the ground state. Electrons are
Fermions and the Pauli exclusion principle for Fermions states that no two particles can occupy
the same state. The single particle states can be characterized by a wave vector k and spin σ. The
number of allowed k-values inside a “volume” d3 k in k-space is from (4.39), noting the relation
p = h̄k between momentum and wave-vector
66 4.2 Sommerfeld model for electrons

V d3 k
(2π)3

When summing over allowed k-states we can replace a sum by an integral in the large V , or
continuum limit
X Z
V
→ d3 k (4.41)
k
(2π)3
We will neglect the interaction between the electrons. The energy of a state with wave vector
2 2
k is Ek = h̄2m
k
, i.e. it increases monotonically with increasing k. The lowest energy state then has
all momentum states occupied up to a maximum wave vector kF , the Fermi wave vector . The
number of states with wave vector less than kF is then (including a factor 2 for spin)

2 × 4πkF 3 V kF 3
N =V =
3(2π)3 3π 2

We refer to the k-space volume with k < kF as the Fermi sphere and its surface as the Fermi
surface . The momentum of a particle with wave vector kF is called the Fermi momentum pF = h̄kF .
The energy of a particle with wave vector kF is the Fermi energy

pF 2
F = (4.42)
2m
The free electron gas is often used as a starting point to describe the electronic properties of
metals. Let us make some order of magnitude estimates from this application.
Length: The Wigner-Seitz radius rs is defined as the ratio

radius of sphere containing one electron

rs =
Bohr radius
so that
V 4πa0 3 rs3
=
N 3
Here a0 = 0.529 × 10−10 m is the Bohr radius. If Z is the valence of a metal then the number
of conduction electrons is N = ZNatoms . From tables of the density, atomic weight and Avogadro’s
number one can then work out the value of rs . For the metallic elements one typically has 2 < rs < 6.
Aluminum is a high density metal with rs = 2.07, cesium is a low density metal with rs = 5.62.
Wave vectors: We have

kF 3 N 3
= = (4.43)
3π 2 V 4πa0 3 rs 3
giving
 1/3
9π 1 1.92
a0 kF = = (4.44)
4 rs rs
Electrons and phonons 67

The Fermi wave vector kF is of the order of the inverse Bohr radius.
Velocity: The Fermi velocity is given by

h̄kF 4.2
vF = = × 106 m/s
m rs

This velocity is of the order 1% of the velocity of light, or 1000 times a typical sound velocity.
We conclude that electrons in metals are supersonic, but non-relativistic.
Energy: It is instructive to compare the Fermi energy with the Rydberg, the ground state energy
of the hydrogen atom. We have

e2
1Ry = = 13.6eV
4π0 2a0

From (4.44)

 2/3
h̄2 kF 2 h̄2 9π 1
F = = (4.45)
2m 2ma0 2 4 rs 2

Substituting numbers gives

50.1eV
F =
rs 2

For aluminum, F = 11.7eV , for cesium F = 1.59eV . The model which we have described above
is commonly called the Sommerfeld model. We have neglected the electron-electron interaction.
Real electrons are charged particles. The energy associated with the Coulomb repulsion between
two electrons a distance r apart is

e2
4π0 r

A typical nearest neighbor distance between electrons is ∼ a0 rs and we see that the Coulomb
energy and the kinetic energy is comparable. Nevertheless, the Sommerfeld model explains many
of the qualitative properties of the physics of metals. Since the kinetic energy is ∼ 1/rs2 and the
model will work best when rs is small- and for many purposes aluminum is the most free electron
like metal.
Temperature: Thermal energies available to electrons are of the order kB T. It is instructive to
compare this quantity to the Fermi energy. For this purpose one can define the Fermi temperature
as

TF = F /kB

The Fermi temperature is 13.6×104 K for Al and 1.84×104 for cesium. These are high temper-
atures compared to room temperature.
68 4.2 Sommerfeld model for electrons

4.2.3 Low temperature electronic specific heat

At nonzero temperatures not all the electrons will occupy states inside the Fermi sphere. Some
electrons will be excited from states inside the Fermi sphere to an electronic state outside the sphere
leaving a hole behind, see figure 4.7(b). Let f () be the the probability that a state is occupied
and 1 − f () the probability that it is not. We also introduce the chemical potential µ. If  < µ
the normal state of the affairs is that the state is occupied, while for  > µ it is normally empty.
We require that the ratiosatisfies

f ()
= e−β(−µ) (4.46)
1 − f ()
i.e. the ratio is given by a Boltzmann factor.

1 1

f(ε) f(ε)

ε ε
µ µ

Figure 4.6: (a) The Fermi-Dirac distribution at non-zero temperature. (b) Fermi-Dirac distribution
at T = 0

From (4.46) the probability that a state with energy  is occupied is given by the Fermi-Dirac
distribution (figure 4.6)

1
f () = (4.47)
eβ(−µ) +1
The mean number of particles is then given by
Z
V
hN i = 2 d3 kf (k ) (4.48)
(2π)3
The condition (4.48) determines the chemical potential µ. For T = 0 we have µ = F , while for
T 6= 0 µ will be a slowly varying function of T . When discussing thermodynamic properties of the
Fermi gas it is convenient to change the integration variable from wave vector to energy. We have

h̄2 k2 3 h̄2 k
= , d k = 4πk2 dk, d = dk
2m m
Z∞
hN i = d D()f ()
−∞
Electrons and phonons 69

where the density of states is given by (see figure 4.7)

 3/2
V 2m √
D() = 
2π 2 h̄2
The density of states is a convenient quantity when taking thermal averages. The internal
energy of the ideal Fermi gas is e.g.
Z ∞
U= D()f ()d (4.49)
−∞

particle
X
D( ε )

kF

hole
D( ε ) f( ε)

ε
(a) (b)

Figure 4.7: (a)Density of states of free electrons. (b) Particles and holes.

Let us use this result to find a formula for the specific specific heat
 
∂U
CV =
∂T VN

The only quantity in (4.49) which depends on temperature is f (). The trouble is that the other
independent variable is µ not N . When differentiating f with respect to T we must therefore con-
sider µ to be an implicit function of T, N, V . We can obtain the leading term in a low temperature
expansion for the specific heat by noting that for T = 0, µ = eF . The temperature dependence of
U at constant N comes about because of thermal excitations in which particles are excited into
previously unoccupied states outside the Fermi surface leaving holes behind (figure 4.7 (b)).
We use F as our reference energy: Then  − F = is the excitation energy of a particle outside
Fermi surface. Similarly the excitation energy of a hole, i.e. a particle missing from a single particle
state of energy  is F − . The change in internal energy due to thermal excitations is then
Z∞ ZF
∆U = d( − F )f ()D() + d(F − )(1 − f ())D()
F −∞
70 4.2 Sommerfeld model for electrons

We differentiate inside the integrand to get

Z ∞ ∂f
CV = d( − F ) D() (4.50)
−∞ ∂T
We now approximate µ ' F and neglect the temperature dependence of µ

∂f  − F eβ(−F )
=
∂T kB T 2 (eβ(−F ) + 1)2

Because of the factor ∂f /∂T the integrand in (4.50) will be sharply peaked near  = F . If we
assume that near  = F , D() ' D(F ) and put x = β( − F ) we get

Z∞
x2 ex
CV ' kB T D(F )
2
dx (4.51)
(ex + 1)2
−∞

π2
The integral in (4.51) can be evaluated to yield 3 and we get the formula for the electronic
specific heat

T π 2 kB 2
CV = D(F ) (4.52)
3
2 2
Note that we have not made use of the formula  = h̄2m k
for the electron energy. In a metal
electrons occupy energy bands where k may be a complicated function of k. Equation (4.52) still
remains valid in this situation.
If we substitute the free electron value for the energy we get

π 2 N kB T
CV = (4.53)
2 TF
An important aspect of (4.52) and (4.53) is the linear temperature dependence. It can be
shown that for low temperatures the contribution to the specific heat from lattice vibrations will
be proportional to T 3 . Therefore, the specific heat of metals at low temperatures (T≤ 10K) will
be dominated by the electronic contribution. This contribution is still small. The factor T /TF
in (4.53) indicates that only a small fraction of the conduction electrons will be thermally excited
even at room temperature. At ordinary temperatures the lattice specific heat dominates.

4.2.4 Problems
Problem 4.2-1:

In sodium metal each ions contributes one conduction electron. Using the data in table 5.1 calculate
for sodium
(a). The Fermi energy
(b). The Fermi velocity
(c). The Fermi temperature.
Transport properties 71

Problem 4.2-2:
3 Heatoms can be considered as Fermi particles. At low temperatures 3 He forms a liquid with a
volume of 4.62 × 10−29 m3 per helium atom. The mass of a 3 He atom is 5 × 10−27 kg. Estimate the
Fermi temperature of 3 He.

Problem 4.2-3:
Show that the kinetic energy of a three dimensional gas of N free electrons is 35 N F .

Problem 4.2-4:
Estimate the the electronic and lattice specific heats for Al at temperatures 1K, 10K, 100K in
units of J mol −1 K −1 . Use the Somerfeld model for the electrons and the Debye model for the
phonons. Aluminum is trivalent.

4.3 Drude model

In this section we will discuss transport properties associated with the conduction electrons, mainly
the conductivity of a metal subject to a DC electric field or an alternating electromagnetic field.
The latter extension of the theory will also will allow us to discuss the optical properties of a metal.
We will employ a simple classical model, the Drude model. For simplicity we will assume that the
metal is isotropic and neglect all band structure effects. Our discussion will of necessity be rather
brief and we refer the reader to Ashcroft and Mermin [2], Callaway [4], Wooten [12] and Ziman [11]
for a more extensive discussion.
Let us first consider the DC conductivity of a metal containing n conduction electrons per unit
volume. Each electron will have a velocity v which will be some fraction of the Fermi velocity, but
in the absence of an electric field the electrons are equally likely to move in any direction, so the
average velocity hvi = 0. However , if we apply an electric field E~ the electrons will be accelerated
in the field, and there will be an average nonzero drift velocity vdrif t. We let the charge of the
elctrons be e = −|e| and the mass be me . From time to time the electrons will undergo collisions
with impurities and phonons and we assume that after each collision the electron will be equally
likely to travel in any direction, so that immediately after each collision the average velocity is zero.
Let v(0) be the velocity immeditately after a collision. Some time t after that collision, but before
the next collision the electron velocity will be

~
eEt
v(t) = v(0) + (4.54)
me
The average time between collisions, τ , is commonly called the relaxation time. We see from
(4.54) that the drift velocity in our simple model will be

~
eEτ
vdrif t = hv(τ )i = (4.55)
me
The electric current density (current per unit area) is then
72 4.3 Drude model

ne2 τ ~
j = nevdrif t = E (4.56)
me
According to Ohm’s law the electric field needed to produce a given current is given by

E~ = ρj

where ρ is the resistivity. Conversely we put

j = σ0 E~

with σ0 the DC conductivity. We thus find in the Drude model

ne2 τ
σ0 = (4.57)
me
It is instructive to construct a differential equation for the average velocity hv(t)i. If the average
time between collisions is τ , the probability of a collision taking place in the time interval dt will
be dt/τ . The mean speed will then increase by

~
eEdt
dhv(t)i = (4.58)
me
with probability (1 − dt/τ ) or drop to zero with probability dt/τ . Neglecting terms of order dt2
we find

dhv(t)i eE~ hv(t)i

= −
dt me τ

In the steady state hv(t)i → vdrif t and

eEτ
~
vdrif t =
me
and we get the same result as before.
We can also consider the response to a AC electric field

−iωt
E~ = Re[E(ω)e
~ ]

Assuming a steady state response

hv(t)i = v(ω)e−iωt (4.59)

we find

eE(ω)
~ v(ω)
−iωv(ω) = −
me τ

Putting j(t) = Re[j(ω)e−iωt ] we write

Transport properties 73

j(ω) = σ(ω)E(ω)
~

and we find after some algebra

σ0
σ(ω) = (4.60)
1 − iωτ
We see that (4.60) exhibits two distinct régimes: a low frequency régime where ωτ << 1 in
which the current is approximately in phase with the field, and a high frequency régime ωτ >> 1
in which the current will be approximately 90o out of phase .
Another way of looking at the system is in terms of a complex dielectric constant.

(ω) = r (ω)0

where r is the relative dielectric constant and 0 is the permittivity of vacuum. We have

D
~ = 0 E~ + P
~

where D ~ is the electric displacement and P

~ the electric polarization per unit volume, P
~ = neu
and u is the dispalcement of an electron, caused by the applied field
du
=v
dt
In anlalogy with (4.59) we put u = u(ω)e−iωt and find −iωu(ω) = v(ω) and after a little algebra
we find

iσ(ω)
r = 1 + (4.61)
0 ω
We define the Plasma frequency as
s
ne2
Ωpl = (4.62)
0 m
The formula (4.61) for the relative dielectric constant can be rewritten as

Ωpl
r = 1 − (4.63)
0 ω(ω + i/τ )
You will learn from other courses that an electromagnetic wave with frequency ω propagates
√
in a medium with dielectric constant r with wave vector q = q0 r where q0 = ω/c is the wave
vector in vacuum. If the dielectric constant is complex the wave will be damped and it is customary
to write
√
r = n + ik

where n and k are optical constants. The normal incidence reflectivity from vacuum can be
shown to be given in terms of the optical constants as
74 4.3 Drude model

5 1
R
4 0.8

3 0.6
k

2 0.4

1
n n 0.2

0 0
0 0.5 1 1.5 2 0.25 0.5 0.75 1 1.25 1.5 1.75 2
ω/Ωpl ω/Ωpl

(a) (b)

Figure 4.8: (a) Optical constants n snd k and (b) normal incidence reflectivity in the Drude model.
Frequencies are shown in units of the plasma frequency Ωpl and the curves are drawn for the case
Ωpl τ = 20

(n − 1)2 + k2
R= (4.64)
(n + 1)2 + k2

Inf figure 4.8 we plot the optical constants and the reflectivity in the Drude model.
For frequencies which are small compared to the plasma frequency the imaginary part of the
index of refraction will be large and electromagnetic waves can only penetrate a short distance in
a metal (skin effect). The intensity of the wave will decay according to

2ωk
I ∝ E 2 ∝ exp (− z)
c

where z is the penetration depth.

For frequencies higher than the plasma frequencies k will typically be very small and the metal
will be transparent to electromagnetic radiation. Similarily, the reflectivity of a metal will be high
for frequencies below the plasma frequency, and low for higher frequencies.

4.4 Problems
Problem 4.3-1
(a). The relaxation time τ for an aluminum sample is 0.8 × 10−14 s. Calculate the DC resistivity.
(b). Calculate the plasma frequency for Aluminum (note that Al is trivalent, values of required
physical constants can be found near the end of Part I of the lecture notes).
(c). To what electromagnetic wavelength in vacuum will the plasma frequency correspond in
Aluminum.
Transport properties 75

Problem 4.3-2
Consider a metal for which ωpl τ >> 1. The dielectric constant for frequencies larger than the
plasma frequencies will then be approximately real, but less than one. The phase velocity for
electromagnetic waves is ω/q, with the wavevector q = 2π/λ. The phase velocity for ω > Ωpl will
then be faster than the speed of light. This will not violate special relativity since signals propagate
with the group velocity given by vgroup = dω/dq.
(a). Show that the group velocity of light is less that the speed of light for ω > Ωpl in the
Drude model.
(b). One consequence of the fact that for ω > Ωpl the index of refraction n < 1 in the Drude
model is that there will be total reflection of X-rays impinging with an angle of incidence larger
than a critical angle given by sin(θc ) = n. Calculate the critical angle of incidence for light with
frequency 2Ωpl .

Problem 4.3-3
Calculate the penetration depth at which the intensity of an electromagnetic wave in Aluminum
relaxation time τ = 0.8 × 10−14 s will be reduced by a factor of 2 at the frequency of
(a). ω = 2π107Hz.
(b).ω = 2π1014Hz.
76 4.3 Drude model
Chapter 5

Review

5.1 Midterm review problems

Problem 5.1-1:
(a). Which crystal planes of the copper metal structure have ithe highest density
of atoms?
(b). What is the density of atoms per cm2 in these planes?
(c). Find the Bragg angle(s) for reflection of x-rays of wave-length 1.50 Å against
these planes?
(d). Calculate the density of copper in g/cm3 from the data in 5.1. The atomic
weight of Copper is 63.54g/mol

Problem 5.1-2:
(a). Describe the diamond lattice. Is it a Bravais lattice? If not, describe it as a
Bravais lattice with the smallest possible number of basis vectors.
(b). Find the angle between any two of the lines (bonds) joining a site of the
diamond lattice to its four nearest neighbors.

Problem 5.1-3:
The primitive lattice vectors of a lattice are

a a a
a= (−x̂ + ŷ + ẑ); b = (x̂ − ŷ + ẑ); c = (x̂ + ŷ − ẑ)
2 2 2

where x̂, ŷ, ẑ are three unit vectors along Cartesian axes.
(a). What is the Bravais lattice?
(b). Find three primitive reciprocal lattice vectors.
(c). Which lattice plane has the highest density of atoms? Calculate the density of
atoms per unit area for this plane.

77
78 Old Midterms

Problem 5.1-4:
(a). Which of the following crystal structures are Bravais lattices: the hexagonal
closed packed, the body centered cubic, the face centered cubic, the diamond structure?
(b). Find a set of primitive reciprocal lattice vectors for the structures listed above.

Problem 5.1-5:
Zinc crystallizes in the hexagonal
p closed packed structure, with a c/a ratio slightly
different from the ideal value 8/3). The atomic weight of Zn is 65.37. Calculate the
mass density of Zn.

Problem 5.1-6:
(a). Explain the effect of dislocations on the mechanical strength of materials.
(b). Discuss the strengths of very well prepared crystals, perfect crystals, and
relatively poorly prepared crystals. Which are strongest or weakest and why?
(c) In an ionic crystal such as NaCl the vacancy formation energy will be different for
positive and negative ion vacancies. Nevertheless, if the crystal is pure the concentration
of positively and negatively charged vacancies will be roughly the same. Describe the
mechanism through which this comes about.
(d) Define Burger’s vector.

5.2 Old Midterms

Midterm examination February 1992

“Closed book exam”

1:
Silicon crystallizes in the diamond structure (fcc with a basis vector 14 [111])anda= 5.43 Å the side
of the unit cube. The atomic number of Si is 28.09.

(a). Calculate the density of Si in kg/m3 .

(b). What type of structure is the reciprocal lattice of Si?
(c). Write down a set of primitive reciprocal lattice vectors for Si.

2:
The element Xe crystallizes in the fcc structure at low temperatures. Assume an interatomic
Lennard-Jones potential with σ = 3.98 Å and  = 0.020eV . Calculate the nearest neighbor distance
in Å of Xe and the cohesive energy in eV /atom at low temperatures.
Review 79

3:
Describe briefly the Ewald construction and its use in the rotating crystal method of crystal struc-
ture determination.
Some formulas:
Avogadro’s number is 6.022 × 1023, an atomic mass unit is 1.67×10−27kg. 1Å= 10−10 m. The
Lennard Jones potential is

σ σ
v(r) = 4[( )12 − ( )6 ]
r r

X rnn
An = ( )n
R
R

A6 = 14.45, A12 = 12.13 for the fcc lattice.

Midterm examination February 1993

Allowed aids: 2 page handwritten “Cheat sheet”, calculator. Answer all questions.

1:
A certain lattice has a primitive unit cell with the dimensions a = 3Å, b= 4Å, c= 5Å, α = β = 90◦ ,
γ = 60◦ .

(a). Write down a set of primitive lattice vectors a, b and c on Cartesian form.
(b). Construct a set of primitive reciprocal lattice vectors.

2:
Iron is a body centered cubic metal. The edge of the unit cube is a = 2.87Å.

(a). Which lattice planes of the iron structure have the highest density of atoms?
(b). What is the density of atoms/m2 in these planes? (1 Å= 10−10 m).
(c). Find the Bragg angle(s) for reflection against these planes if x-rays of wave-
length 1.54 Å are employed. (The formula for Bragg scattering is 2d sin θ = nλ).

3:
(a). Sketch the Wigner Seitz cell of the triangular lattice.
(b). Define edge and screw dislocations.
(c). What is the difference between the bulk modulus B and Young’s modulus Y ?
Which one will be largest of the two? (Describe the difference in words, no formulas
are required).
80 Old Midterms

Midterm examination February 1994

Allowed aid: Calculator, 2 page “cheat sheet”
Answer all 3 questions.
1:
In each of the following cases indicate if the structure is a Bravais lattice. If it is, give three
primitive lattice vectors, if it is not describe it as a Bravais lattice with as small as possible basis.
In all cases the side of the unit cube is a.
(a). Base centered cubic (simple cubic with additional points in the centers of the horizontal
faces of the cubic cell).
(b). Side centered cubic (simple cubic with additional points in the centers of the vertical faces
of the cubic cell).
(c). Edge centered cubic (simple cubic with additional points at the midpoints of the lines
joining nearest neighbors).
2:
(a). Write down a set of Bravais primitive lattice vectors a, b and c and basis vector d for the
hexagonal close packed lattice.
p
(b). Show that the ideal c/a ratio is 8/3.
(c). Zinc has atomic weight 65.37 and crystallizes in a hexagonal closed packed structure with
a = b = 2.66Å, c = 4.95Å, i.e. the c/a ratio is slightly different from the ideal. Calculate the mass
density of Zinc in g/cm3.
3:
An isotropic elastic material is shaped like a cube. The faces of the cube are subject to a
uniform pressure P , and no other stresses.
(a). Express the components of the strain in terms of P and the Lamé coefficients λ and µ
defined by C11 = λ + 2µ, C1 2 = λ, C44 = µ.
(b). Find the dilation (fractional volume change).
(c). Define the Young’s modulus and the bulk modulus. Why are they different?
Values of physical constants
Atomic mass unit amu = 1.66042 × 10−27 kg
Avogadro’s number NA = 6.0225 × 1023 mol −1

Midterm examination February 1995

Allowed aid: Calculator, 2 page “cheat sheet”
Answer all 3 questions.
1: Zinc crystallizes in the hexagonal closed packed structure,
p with a = 2.66 Å, c = 4.95 Å
(The c/a ratio is slightly different from the ideal value 8/3). The atomic weight of Zn is 65.37,
1a.m.u. = 1.6604 × 10−27 kg.
(a). Calculate the mass density of Zn.
(b). Write down a set of reciprocal lattice vectors.
(c). Calculate the distance between the (011) planes.
Review 81

2: The element Kr crystallizes in the fcc structure at low temperatures. Assume an interatomic
Lennard-Jones potential

σ σ
v(r) = 4[( )12 − ( )6 ]
r r

We have A6 = 14.45, A12 = 12.13 for the fcc lattice with

X rnn
An = ( )n
R
R

The nearest neighbor distance is rnn = 3.99 Å for Kr and the cohesive energy is -0.120 eV /atom
at low temperatures.
(a). Derive formulas which express the cohesive energy and the nearest neighbor distance in
terms of σ and .
(b). Calculate the Lennard-Jones parameters σ and .

3: An isotropic elastic material is subject to a uniform tension T in the z−direction and no other
stresses.
(a). Express the components of the strain in terms of T and the Lamé coefficients λ and µ
defined by C11 = λ + 2µ, C12 = λ, C44 = µ.
(b). Find the dilation (fractional volume change).

Midterm exam February 1996

Allowed aids: Calculator, 2 page ”cheat sheet”.
Answer all questions
1: The CsCl structure is simple cubic with a basis vector d = a2 (x̂ + ŷ + ẑ). It is a common lattice
structure for ionic materials and differs from the bcc lattice only in that the body center and corner
sites are occupied by atoms of different species.
(a) If the two types of ions in the CsCl are of the same size what is the packing fraction f ?
(b) What is the packing fraction if the radius of one type of ions is twice that of the other type?
2: Iron is a bcc metal with a = 2.87 Å.
(a) Find a set of primitive reciprocal lattice vectors for iron.
(b) Find the scattering angles (angles between incident and scattered wave vectors) for x-rays of
wave vectorλ = 1.54 Å impinging upon a powder sample of iron. (Bragg’s law of scattering is
2d sin θ = nλ, where θ is the angle of incidence and d is the distance between lattice planes).
3: In an isotropic material it is conventional to describe the elastic constants in terms of the Lamé
constants λ = C12 and µ = C44 . The elastic constant C11 is then given by λ + 2µ. In a certain
material λ = 1. × 1010 N m−2 , µ = 0.5 × 1010 N m−2 . A wire made of this material of cross section
1mm2 is pulled (subject to uniaxial tension) so that its length increases by 1%.
(a)What is the force required to pull the wire?
(b)What is the change in the cross sectional area of the wire?
82 Old finals final

5.3 Old finals

April 1993 Final

Time 3 hours. Allowed aids: One double sided ‘cheat sheet’, calculator. Answer 5 out of the 7
questions. All questions have equal value. If you answer more than 5 questions you will be given
credit for the 5 best answers.
1:
An A+ B − ionic solid crystallizes in the N aCl structure (i.e. the lattice is fcc with a basis
vector d = a2 x̂, where a is the side of the unit cube).
(a). Assuming that the positive and negative ions have equal size what is the packing fraction
of the material?
(b). If the positive ion has a radius which is 12 that of the negative ion what would the packing
fraction be?
2:
(a). What is the Ewald construction?
(b). Describe the Laue method of X-ray diffraction.
(c). Describe the rotating crystal method of crystal structure determination.
3:
An isotropic elastic material is shaped like a cube. The faces which are parallel to the x − z
and y − z planes are subject to a uniform pressure P , while the faces parallel to the x − y plane are
not subject to any external force. The elastic constants of the material are C11 = λ + 2µ, C12 = λ
and C44 = µ.
(a). Find the components of the strain.
(b). What are the stresses acting on a (110) surface?
4:
Aluminum is face centered cubic with the side of the unit cube a = 4.05 × 10−10 m. The Debye
temperature is 428K.
(a). Find the Debye wave vector qD .
(b). What is the sound velocity?
5:
Beryllium has the h.c.p. structure with lattice constants a = 2.27 × 10−10m, c = 3.59 × 10−10m.
Each ion contributes 2 conduction electrons. Assuming Beryllium to be a free electron metal
(a). Calculate the Fermi energy of Beryllium.
(b). Calculate the Fermi velocity.
6:
(a). What is chirality?
(b). Describe the features of chiral nematic liquid crystals which makes them useful as ther-
mometer materials.
7:
(a). Describe the micellar and lamellar phases of soaps.
(b). Describe how a polymer dispersed liquid crystal display works.
Review 83

April 1994 final

Time 2 1/2 hours. Allowed aids: One double sided hand written ‘cheat sheet’, calculator. Answer
4 out of the 5 questions. All questions have equal value. If you answer more than 4 questions you
will be given credit for the 4 best answers.
—————————————
1:
Given the table of physical properties of sodium and the values of the physical constants listed
at the end of the paper find
a: the nearest neighbor distance of two sodium ions.
b: a set of primitive reciprocal lattice vectors.
c: the scattering angles (angle between incident and scattered wave vector) φ for x-rays of
wavelength λ = 1.54 Åimpinging upon a powder sample of sodium. (Bragg’s law of scattering is
2d sin θ = nλ, where θ is the angle of incidence, and d is the distance between planes.)
—————————————
2:
Assume that the conductivity σ of sodium is given by σ = ne2 τ /m, where n is the number of
valence electrons per unit volume, e the electron charge, and m the electron mass. Use the table
of physical properties of sodium and the values of the physical constants listed at the end of the
paper to find
a: the relaxation time τ
b: the electron drift velocity if a current of 105 Amp m−2 passes through a sodium sample.
c: Assume that the sodium specific heat is given by the Dulong-Petit value CV = 3N kB . How
fast will the temperature rise if a current of 105 Amp m−2 passes through a sodium sample and no
heat is transported away?
—————————————
3:
Use the table of physical properties of sodium and the values of the physical constants listed at
the end of the paper to estimate
a: the sound velocity
b: the Fermi velocity
c: the Fermi energy of sodium in electron volts.
4:
a: What is a plastic deformation?
b: Why is the maximum shear stress σc of a material usually much less than the theoretical
value predicted for a perfect crystal?
c: Define Burger’s vector and draw pictures showing the Burger’s vector for an edge and a
screw dislocation.
—————————————
5:
a: What is the the micellar phase of a soap solution?
b: What is a chiral nematc liquid crystal?
c: Why are chiral nematics useful as thermometer materials?
—————————————
Some properties of Sodium Metal
84 5.7 April 1994 final

Valence +1 Structure body centered cubic Debye temperature 156 K

Density 0.97 g cm−3 Conductivity 0.21×106 Ω−1 cm−1 Weight of 1 mol 22.99 g

April 1995 final

Time 2 1/2 Hours. Allowed aids: One double sided hand written ‘cheat sheet’, calculator. Answer
4 of the 5 questions. If you answer all the questions you will be given credit for the 4 best answers.
All questions have equal value. You will find values of some physical constants on the last page.
——————————
1:
Potassium is a bcc metal with one valence electron per potassium ion. The side of the unit cube
is a = 5.23 Å. The Debye temperature is 91K and the atomic weight of potassium is 39.10gmol −1.
What is
a: The plasma frequency Ωpl ?
b: The velocity of sound?
c: The heat capacity at room temperature of 1kg of potasium (in JK −1 )?
d: The Fermi energy (in electron volts)?
——————————
2:
Consider a linear chain of N atoms. Each atom interacts with its nearest neighbor on either side
with a potential
K
V (x) = V0 + (x − a)2
2
where x is the distance between two nearest neighbor atoms and V0 , K and a are constants.
a: Find the velocity of sound.
π
b: Find the phase and group velocity of a lattice wave with wave vector k = 4a .
c: Sketch the displacements u along the chain for the lattice wave in (b) at an instant in time
when the amplitude of an atom at the origin is at its maximum value.
——————————
3:
In a good conductor the relaxation time τ will be sufficiently long that ωτ >> 1, when the frequency
ω is larger than the plasma frequency Ωpl . The relative dielectric function r can then be taken as
√
real, as can the index of refraction n = r . An electromagnetic wave propagates in such a metal
with wave vector q = nq0 , where q0 = ω/c and c is the velocity of liqht in vacuum. Consider a
metal for which
Ω2pl
r = 1 −
ω2
a: Find the phase velocity vp = ω/q for the wave when ω > Ωpl .
b: Find the group velocity vg = dω/dq for the wave when ω > Ωpl .
c: Show that vg vp = c2 .
d: Give a numerical answer to (a) and (b) when ω = 2Ωpl .
——————————
4:
Review 85

a: What type of crystal exhibits optical phonons?

b: Outline the main features of the Debye model of lattice vibrations.
c: Show that in the Debye model a solid will satisfy the Dulong-Petit law for the heat capacity
for temperatures much higher than the Debye temperature.
——————————
5:
a: Describe the principles behind the twist nematic liquid crystal cell in a liquid crystal display.
b: What is chirality? Give an example of an object which is chiral and one which is not.
c: What is a chiral nematic?
——————————

April 1996 final

Time 2 12 Hours. Allowed aids: One double sided hand written ‘cheat sheet’, calculator. Answer 4
of the 5 questions. If you answer all the questions you will be given credit for the 4 best answers.
All questions have equal value. You will find values of some physical constants on the last page.
——————————
1:
(a) Describe the main types of point defects in crystals.
(b) Consider an ionic crystal of the type A+ B − . Assume that the energy associated with moving a
positive or negative ion from the bulk of the crystal to its surface is respectively + and − . Derive
formulas for the equilibrium vacancy concentration (% vacant sites) of positive and negative ions
for (i)pure crystals and where c % of the positive ions have been replaced by divalent C ++ ions.
(c) Often the actual concentrations of vacancies in a crystal is much larger than the equilibrium
concentration. Discuss how this can happen.
(d) Discuss why a small amount of edge dislocations reduces the shear strength of a crystal, while
adding more dislocations and/or impurities can increase the strength of a crystal.
——————————
2:
Consider a chain of atoms, all of mass m. The equilibrium distance between nearest neighbor atoms
are alternatively a1 and a2 so that the length of the unit cell is

a = a1 + a2

and there are two atoms per unit cell. The spring constant of the bond of length a1 is K1 , while
the spring constant of the bond of length a2 is K2 .
(a)Write down the equations of motion for the displacements u1 and u2 for the two atoms in a unit
cell.
(b)Find solutions to the equation of motion on the form

u1 = e1 ei(kna−ωt)

u2 = e2 ei(kna−ωt)
86 5.9 Values of physical constants

and find an equation for the frequency ω as a function of the wave-vector k.

(c)Consider the special case k = π/a and find the possible values of the frequency ω.
(d)Describe the vibrational modes when k = π/a.
——————————
3:
(a) Outline the main features of Debye model of lattice vibrations.
(b) Gold is face centered cubic with the side of unit cube 4.08 × 10−10 m. The atomic weight of
gold is 197 a.m.u. and the Debye temperature is θD = 162K. Find the Debye wave-vector qD .
(c) Estimate the sound velocity in gold.
(d) What is the room temperature heat capacity of a 1kg gold bar?
——————————
4:
(a). Outline the main features of the Sommerfeld (free electron) model of the conduction electrons
in a metal.
(b). Present an argument why the electronic specific heat of a metal CV is proportional to the
temperature T and inversely proportional to the Fermi temperature TF .
(c). K (potassium) is a monovalent, body centered cubic metal with a = 5.23 × 10−10 m. Find the
Fermi wave vector kF .
(d). Find the plasma frequency Ωpl .
——————————
5:
(a) Show that the kinetic energy of N free electrons at temperature T = 0 is 35 N F where F is
the Fermi energy.
(b)The relaxation time τ for an aluminium sample is 0.8 × 10−14 s. Calculate the DC conductivity
of Al. (Aluminium is face centered cubic with a = 4.05 × 10−10 m, and trivalent).
(c) Calculate the real and imaginary part of the relative dielectric constant of Aluminium at the
frequency ω = 2π × 1014 rad s−1 .
——————————

5.4 Values of physical constants

1 Å=10−10 m = 10nm
Atomic mass unit amu = 1.66042 × 10−27 kg
Avogadro’s number NA = 6.0225 × 1023 mol −1
Boltzmann constant kB = 1.381 × 10−23 J/K
Electron charge e = −1.602 × 10−19 C
Electron mass me = 9.109 × 10−31 kg
Permittivity of vacuum 0 = 8.8542 × 10−12 F/m
Plancks constant h̄ = 1.054 × 10−34 kg m2 /s
Review 87

Substance Structure a,Å cÅ Substance Structure a,Å cÅ

Ag fcc 4.09 Ge diamond 5.66
Al fcc 4.05 InSb ZnS 6.46
AgBr NaCl 5.77 K bcc 5.23
Ar[at 4K] fcc 5.31 KCl NaCl 6.29
Au fcc 4.08 Kr[at 4K] fcc 5.64
Be hcp 2.27 3.59 Na bcc 4.23
C(diamond) diamond 3.57 NaCl NaCl 5.63
CdS ZnS 5.82 Ne[at 4K] fcc 4.46
Co hcp 2.51 4.07 Pb fcc 4.95
Cr bcc 2.88 Si diamond 5.43
CsCl CsCl 4.11 Xe[at 4K] fcc 6.13
Cu fcc 3.61 Zn hcp 2.66 4.95
CuCl ZnS 5.41 ZnS ZnS 5.41
Fe bcc 2.87

Table 5.1: Crystal structure of some common substances. Unless specified the temperature is 300K

Element name Symbol Atomic weight

Aluminum Al 26.982
Argon Ar 39.948
Copper Cu 63.55
Iron Fe 55.847
Krypton Kr 83.80
Silicon Si 28.086
Zinc Zn 65.38

Table 5.2: Atomic weights of selected elements

88 5.9 Values of physical constants
Bibliography

[1] W.D. Callister Jr. (1994). Materials science and engineering: an introduction Wiley.

[2] N. W. Ashcroft and N.D. Mermin (1976). Solid state physics, Holt Rinehart and Winston.

[3] D.J. Barber and R. Loudon An introduction to the properties of condensed matter Cambridge
University Press (1989).

[4] J. Callaway, J. (1991). Quantum Theory of the Solid State, 2nd edition, New York: Academic
Press.

[5] J. R. Christman (1988) . Fundamentals of solid state physics, John Wiley.

[6] C. Kittel (1986). Introduction to solid state physics, 6th edition, John Wiley.

[7] L. D. Landau and E.M. Lifshitz (1986), Theory of Elasticity, 3rd Edition, Pergamon Press.

[8] M.A. Omar (1975). Elementary solid state physics.

[9] C.H. Scholz (1990). The mechanics of earthquakes and faulting, Cambridge University Press.

[10] M.T. Sprackling (1990). Mechanical properties of solids, p. 727 in The Encyclopedia of Physics,
ed R.M. Besançon, Van Nostrand.

[11] Ziman, J. M. (1964). Principles of the Theory of Solids, Oxford: Clarendon Press.

[12] F. Wooten (1972) Optical properties of solids Academic Press

89
Index

Acoustic modes, 57 Current density

Ashcroft N.W., 3 of particles, 41
Atomic quantum numbers, 23
Atoms, 23 DC conductivity, 71
Attempt frequency, 41 DC Conductivity, 72
de Broglie relation, 20
Balzarini D., 1 Debye frequency, 62
Barber, D.J., 30 Debye layer, 40
Basis vector, 6 Debye model, 62–64
Bergersen B., 1 Debye temperature, 62
Bloom M., 1 Defects, 39–44
Body centered cubic lattice, 5 Deformation tensor, 30
Boltzmann factor, 39, 61, 68 Delaunay tessellation, 21
Bragg’s law, 15 Density of states, 60, 69
Bravais lattice, 3–15 Diamond structure, 7
Brillouin zone, 21, 54, 59 Diatomic chain, 56–58
Brittle materials, 46 Dielectric constant, 73
Burger’s vector, 42 Diffraction, 15
Diffusion constant, 41
Callister W.D., 1 Dilation, 33
Central Potential, 24 Dipole moment, 24
Cesium chloride structure, 8 Direct lattice, 17
Christman J.R., 3 Dislocation, 45
Close packing, 10 Dislocation core, 43
Cobb R., 1 Dislocations, 42–43
Cohesion in metals, 29 Displacement from equilibrium, 30
Cohesive enrgy, 25 Doping, 40
Collisions, 71 Double couple, 34
Compression, 34 Drift velocity, 71
Contour maps, 21 Drude model, 71–75
Coordination number, 9 Ductile materials, 45
Core electrons, 23 Dulong and Petit heat capacity, 61, 63
Coulomb energy, 67 Dynamic fatigue, 47
Covalent crystal, 29 Dynamical matrix, 59
Cracks, 46
Crystalline solid, 3 Edge dislocation, 42
Crystallography, 3 Elastic constant, 35
CsCl structure, 14 Elastic constants for cubic crystals, 35

90
References 91

Elastic scattering, 18 Isotropic materials, 37

Electric displacement vector, 73
Electric polarization, 73 Jump rate, 41
Electron gas, 64
Kittel C., 3
Electron gas at T=0, 65
Electronic band structure, 64, 70 Lamé constants, 37
Electronic specific heat, 68, 70 Landau, L.D.., 30
Engineering stress, 46 Lattice planes, 15
Ewald Lattice vibrations, 51–64
construction, 18, 19, 79 in three dimensions, 58
sphere, 18 Lattice with a basis, 6
Exchange and correlation energy, 29 Laue method, 18
Law of mass action, 40
Face centered cubic lattice, 5
LEED, 20
packing fraction, 9
Lennard-Jones potential, 25, 58
reciprocal lattice, 17
Lifshitz E.M., 30
Family of planes, 15
Line defect, 42
Fermi
Loudon R., 30
energy, 66, 67
Low energy electron diffraction, 20
sphere, 66
surface, 66 Madelung constant, 28
temperature, 67 Maximum shear stress, 44
velocity, 67 Mermin N.D., 3
wave vector, 66 Metals, 29
Fermi-Dirac distribution, 68 Miller indices, 17
Fermions, 65 for cubic lattices, 17
Frenkel defect, 39 Mohs hardness scale, 48
Molecular crystals, 24–27
Geography, 21 Molecules, 23
Group velocity, 56, 75 Murray D., 1
Hardness, 48 NaCl structure, 14
Harmonic approximation, 52 nearest neighbor, 9
Harmonic chain, 52–56 Neuman M., 1
Hexagonal closed packed lattice, 10, 14 Neutron scattering, 20
Hexagonal lattice, 10 Normal modes, 61, 62
Holes, 68, 69
Honey-comb lattice, 6 Ohm’s law, 72
Hooke’s law, 35 Omar M.A., 3
Optical constants, 73
Interaction energy, 25 Optical modes, 57
Interstitials, 39 Orbital angular momentum, 23
Ionic crystals, 27–29, 39
Ions, 23 Packing fraction, 8–11, 14
Isotopes, 23 Pair potential, 24
Isotropic, 13 Pauli Exclusion principle, 24
92 References

Periodic boundary conditions, 52, 59, 64 Structure determination, 15–22

Perovskite structure, 15 Substitutional impurity, 39
Phase transition, 14 Symmetric stress tensor, 34
Phase velocity, 56, 75
Phonons, 60 Tensile forces, 34
Plasma frequency, 73 Tension, 34
Plastic deformation, 45, 46 Thermal neutrons, 20
Point defects, 39–42 Triangular lattice, 6
Poisson ratio, 38 True stress, 46
Polarizability, 25
Unit cell, 3
Primitive
lattice vectors, 4, 21 Vacancy, 39
reciprocal lattice vector, 16 diffusion, 41
unit cell, 4 formation energy, 39
Principal quantum number, 23 Valence electrons, 23
Van der Waals interaction, 24
Quantized energy levels, 60
Virial coefficients, 25
Reciprocal lattice, 15–17, 21 Viscosity, 14
vector, 16 Voronoy polyhedra, 21
Reflectivity, 73 Weakest link, 48
Relative dielectric constant, 73 Weibull W., 48
Repulsive interaction, 24 Wigner Seitz cell, 21
Resistivity, 72 Wigner-Seitz radius, 66
Rotating crystal method, 18 Work hardening, 45
Rotation of crystal, 31
X-ray diffraction, 18
Scattering vector, 16
Schottky defect, 39 Yield point, 46
Schrödinger equation, 64 Young’s modulus, 38
Screw dislocation, 43
Shear Zinc blende (ZnS) structure, 8
deformation, 32
stress, 34
Shear strength, 44
Simple cubic lattice, 4
packing fraction, 9
Sodium chloride (NaCl) structure, 8
Sommerfeld model for electrons, 64
Sommerfeld model for electrons, 71
Sound veloicty, 56
Static fatigue, 47
Strain, 30–39
Strain tensor, 31
Strength of materials, 44–49
Stress, 30–39