Solid State Physics2 PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/326145592
SOLID STATE PHYSICS Course by E. Kogan
Presentation · August 2018
CITATIONS READS
0 6,230
1 author:
Eugene Kogan
Bar Ilan University
184 PUBLICATIONS 825 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Magnetic Resonances View project
Critical Phenomena View project
All content following this page was uploaded by Eugene Kogan on 23 September 2018.
The user has requested enhancement of the downloaded file.

SOLID STATE PHYSICS
Course by E. Kogan
CONTENTS 1. Molecular Orbitals and Valence Bonds 28

2. Hybridization 30
I. About Condensed Matter Physics 2 3. Fractionally Ionic Covalent Bonds 31
A. About These Lecture Notes and About B. Metallic Solids 32
Solving the Problems 2 C. Crystals of Inert Gases 32
D. Hydrogen Bonding 33
II. SPECIFIC HEAT OF SOLIDS 2
A. Dulong - Petit Law 2 X. WAVE SCATTERING BY CRYSTALS. I 34
B. Boltzmann’s Calculation 3 A. The Reciprocal Lattice 34
C. Einstein’s Calculation 4 B. Lattice Planes and Miller Indices 34
D. Debye’s Calculation 4 C. Fermi’s Golden Rule 35
D. The Laue Condition 36
III. ONE-DIMENSIONAL CHAIN 7 E. The Bragg Condition 36
A. Compressibility, Sound, and Thermal F. Equivalence of Laue and Bragg
Expansion 7 Conditions 37
B. Vibrations of A One-Dimensional Chain 8
C. Canonical Quantization 8 XI. WAVE SCATTERING BY CRYSTALS. II 37
D. Phonons and Improvement of Debye’s A. Scattering Amplitudes 37
Calculation 9 B. Experimental Techniques 39
E. Diatomic Basis 10 1. The Rotating-Crystal Method 39
2. The Laue Method 39
IV. ELECTRONS IN METALS: DRUDE 3. Powder Diffraction 39
THEORY 11 C. Examples 40
A. Conductivity 11 D. Scattering in Liquids and Amorphous
B. Thermal Transport 12 Solids 42
E. Scattering and Nobel Prizes 43
V. SOMMERFELD (FREE ELECTRONS)
THEORY 13 XII. BLOCH (ELECTRONS IN A PERIODIC
A. Basic Fermi-Dirac Statistics 13 POTENTIAL) THEORY 43
B. Electronic Heat Capacity 14 A. Brillouin Zones and Bloch Theorem 43
C. Magnetic Susceptibility 15 B. Nearly Free Electron Model 44
D. Pauli Paramagnetism 16 C. Tight Binding Model 45
1. s-state Bands 46
VI. CRYSTALS 16 2. p-state Bands 47
A. Bravais Lattices 16 D. Fermi Surface 48
B. Crystal Structures 18
1. Diamond Structure 19 XIII. ELECTRON DYNAMICS 49
2. Hexagonal Close Packed Structure 19 A. Quasi-Classical Dynamics of Electrons 49
C. Allotropy 19 B. Boltzmann Equation 50
VII. CHEMICAL ELEMENTS 20 XIV. SEMICONDUCTORS 51

A. Periodic Table 20 A. Optical Properties of Semiconductors 51
B. Electrons in Atom 22 B. Electrons and Holes 52
C. Effective Mass 52
VIII. CHEMICAL BOND. I 23 1. Silicon and Germanium 52
A. Ionic Solids 23 D. Doping 54
B. Why are the Caesium Chloride and 1. Impurity States 54
Sodium Chloride Structures Different? 26 E. Statistical Mechanics of Semiconductors 55
C. Zinc Blende and Fluorite Structures 26
D. Perovskite Structures 26 XV. MAGNETIC PROPERTIES OF SOLIDS 56
E. Spinel Structures 28 A. Atomic Energy Levels 56
B. Atom in Magnetic Field 57
IX. CHEMICAL BOND. II 28 C. Lande g-factor 58
A. Covalent Solids 28 D. Rare Earth Ions 58
2
E. Iron Group Ions 59 importantly, however, the physics of solids provides a

F. Paramagnetism 59 paradigm for learning other topics in physics. The things
1. Langevin Paramagnetism 59 we learn in our study of solids will form a foundation for
2. Van Vleck Temperature Independent study of other topics both inside the field of condensed
Paramagnetism 60 matter, and outside of it.
G. Diamagnetism 61
XVI. DIELECTRICS AND FERROELECTRICS 62

A. Electromagnetic Fields in Material
Medium 62
B. Dielectric Constant and Polarizability 62
C. Ferroelectric Crystals 63
D. Polaritons 63
XVII. LANDAU THEORY OF PHASE

TRANSITIONS 64
A. Second-Order Transition 65
B. First-Order Transition 65
XVIII. METALLIC ALLOYS 66

A. Rules for Primary Solubility 66
B. The Phase Diagram 66
C. Thermodynamics: Free Energy and FIG. 1. The discovery of the elements mapped to significant
periodic table development dates (pre-, per- and post-).
Entropy 67
D. Polymorphic Transformation 67
E. The Mixing Entropy of a Substitutional
Alloy 67
F. Melting and Structure 68 A. About These Lecture Notes and About Solving
the Problems
G. Free Energy of Alloy Phases 68
1. Microscopic Model 68
H. The Phase Diagram and Free Energy 69 The Lecture Notes are based on the first 17 Sec-
I. Intermediate Phases 70 tions (plus the Section 19) of the book by Steven
J. Electron Concentration and the Zone H. Simon The Oxford Solid State Basics, on his
Theory of Alloy Phases 71 Lectures (https://podcasts.ox.ac.uk/series/oxford-solid-
state-basics), on the book by C. Kittel Introduction to
Solid State Physics, and on the book by M. A. Omar
I. ABOUT CONDENSED MATTER PHYSICS Elementary Solid State Physics: Principles and Applica-
tions. So the problems which will be suggested at the
exams will be taken either from those books, or from
There are several reasons to study the solid state these Lecture Notes. In general, solving the problems
physics. First of all, solid state physics is by far the is absolutely necessary while learning any subject. One
most technologically useful subfield of physics. Even in who thinks that he/she will be able to learn just by listen-
ancient times human mastery of metals such as copper ing to the Lectures and reading the books, but without
(around 8000 BC), bronze, the first alloy used on a large solving the problems, is like a person who believes that
scale and made of tin and copper, (around 3000 BC), he/she will be able to learn to swim without entering into
and iron (around 1200 BC), completely changed agricul- the water.
ture, weaponry, and pretty much every other aspect of
life. Metallurgy of gold and silver also dates back to
3000 BC. (How much better to get wisdom than gold, II. SPECIFIC HEAT OF SOLIDS
to get insight rather than silver!) Almost all materials
that have found their way to industrial application are A. Dulong - Petit Law
in their solid state. Nowadays, paramount among these
materials are the solids known as semiconductors which
are the basis of the entire electronics industry. Indeed, In 1819 it was found by Dulong and Petit that for many
frequently the electronics industry is even called the solid solids the specific heat is
state industry. Cv = 3R per mole, (1)
Secondly, solid state physics is the biggest and the most
successful subfield of physics. We know far more about where R = 8.3 J/mol·K = 1.99 cal/mol·K is gas constant.
solids than we know about other types of matter. More For diatomic solids the result should be doubled. The
3
specific heat of sodium chloride at room temperature is

12.0 cal/K/mole. Material C/R
Aluminum (Al) 2.91
Antimony (Sb) 3.03
Copper (Cu) 2.94
Gold (Au) 3.05
Silver (Ag) 2.99
Diamond (C) 0.735
TABLE I. Heat capacities Cv of some solids at room temper-

ature and pressure.
FIG. 2. Temperature variation of Cp and Cv in sodium chlo-

ride and copper.
The specific heat at constant volume, Cv , is very dif- FIG. 3. Ludwig Eduard Boltzmann (1844 - 1906) was
ficult to measure and is usually obtained from the ther- an Austrian physicist famous for his founding contributions
modynamic relation in the fields of statistical mechanics and statistical thermo-
dynamics. He was one of the most important advocates for
Cp − Cv = V0 T α2 K, (2)
atomic theory when that scientific model was still highly con-
where troversial.

1 ∂V
α= (3)
V ∂T p Taking into account, that
is the coefficient of thermal expansion, (∂S/∂P )T = −(∂V /∂T )P , (6)

∂P we obtain Eq. (2).
K = −V , (4)
∂V T Problem A.2 (a) Rewrite the Table I presenting heat
is the isothermal bulk modulus, and V0 is the molar vol- capacities in units of J/kg. (b) Find the data and expand
ume if Cp and Cv are molar specific heats. Inserting the Table by including into it 3 other elements (which are
typical figures, (Cp − Cv ) for most liquids and solids at solid at room temperature) of your choice.
room temperature are of the order of R/3 and R/10 re-
spectively. Though the volume expansion is very small,
the pressures which would be needed to counteract it are B. Boltzmann’s Calculation
very large so that their product is by no means negligi-
ble. For gold, for example, α = 14 × 10−6 K−1 , K = 180 In 1896 Boltzmann constructed a model that ac-
GPa, V0 = 10−6 m3 /mol. For very hard substances, such counted for the Dulong and Petit law fairly well. He
as diamond, (Cp −Cv ) however is very small, of the order suggested that (much of the) heat capacity (specific heat)
of R/1000. of materials is due to vibrations of the atom, and con-
Problem A.1 (a) Calculate Cp − Cv for gold. (b) Find sidered each atom in a solid as being in a harmonic well
the necessary data and calculate Cp − Cv for copper and formed by the interaction with its neighbors.
NaCl. Compare your result with Fig. 2. Consider a three-dimensional simple harmonic oscilla-
tor with mass m and spring constant k (i.e., the mass is
attracted to the origin with the same spring constant in

∂S

∂(S, V ) ∂(S, V )/∂(T, P ) all three directions). The Hamiltonian is
Cv = T =T =T
∂T v ∂(T, V ) ∂(T, V )/∂(T, P ) p2 kx2
H= + . (7)
(∂S/∂T )P (∂V /∂P )T − (∂V /∂T )P (∂S/∂P )T 2m 2
=T
(∂V /∂P )T The classical partition function is
Z Z
(∂V /∂T )P (∂S/∂P )T dp
= CP − T . (5) Z= dxe−βH(p,x) , (8)
(∂V /∂P )T (2π~)3
4
where ~ = 1.05×10−34 J·s is the reduced Planck constant, Consider the Hamiltonian (7). The quantum partition
β = 1/kB T , kB = 1.4 × 10−23 J/K = 8.6 × 10−5 eV/K is function is
the Boltzmann constant. Note that X
Z= e−βEj , (17)
R = NA k B , (9) j
where NA = 6.02 × 1023 is the Avogadro number. Since where the sum over j is a sum over all eigenstates.
Z ∞ For a 1d harmonic oscillator
2 p
e−ay dy = π/a, (10)
−∞
En = ~ω(n + 1/2). (18)
in three dimensions we get Hence

3 1
~ω Z1d = . (19)
Z = (~ωβ)−3 = , (11) 2 sinh(β~ω/2)
kB T
p The energy is then
where ω = k/m. From the partition function we get
free energy 1 ∂Z
U =− = ~ω[nB (ω) + 1/2], (20)
Z ∂β
F = −kB T ln Z, (12)
where nB is (a particular case of) the Bose distribution
entropy function
1

∂F T ∂Z
S=− = kB ln Z + , (13) nB (ω) = . (21)
∂T Z ∂T eβ~ω −1
energy Hence the heat capacity is
T 2 ∂Z 1 ∂Z eβ~ω
U = F + T S = kB =− , (14) C = kB (β~ω)2 2. (22)
Z ∂T Z ∂β (eβ~ω − 1)
and heat capacity In 3d
∂U eβ~ω
C= . (15) C = 3kB (β~ω)2 2. (23)
∂T (eβ~ω − 1)
From Eq. (11) follows that
The high-temperature limit agrees with the law of Du-
C = 3kB . (16) long and Petit. Einstein quantum analysis shows that
at temperatures below the oscillator frequency, degrees
If you can consider a solid to consist of N atoms all of freedom freeze out, and heat capacity drops exponen-
in harmonic wells, then the heat capacity should be tially. Einstein frequency is a fitting parameter.
3NA kB = 3R, in agreement with the law of Dulong and Einstein’s theory of specific heat was extremely suc-
Petit. cessful, but still there were clear deviations from the pre-
dicted equation. At low temperatures most materials
Problem B.1 a) Using Eq. (2) and equation of state have a heat capacity that is proportional to T 3 , but Ein-
for an ideal gas calculate Cp − Cv for such gas. stein’s formula at low temperature is exponentially small
b) Using Boltzmann’s method calculate Cv of a in T .
monoatomic ideal gas. Notice that the only chemical el-
ements which are stable single atom molecules at stan-
dard temperature and pressure (STP) are the noble gases. D. Debye’s Calculation
These are helium, neon, argon, krypton, xenon and
radon. Find the data of these gases heat capacity and
In 1912 Peter Debye discovered how to better treat the
compare with the result of your calculation.
quantum mechanics of oscillations of solids, and managed
to explain the T 3 dependance of the specific heat at low
temperatures. He treated the oscillations as sound waves
C. Einstein’s Calculation
with the dispersion law ω(k) = v|k|, where v is the ve-
locity of sound, and quantized these oscillations the same
In 1907 Einstein started wondering about why this law way as Planck had quantized light waves in 1900.
does not hold at low temperatures (for diamond, low A wave (in 1d) is described by the space dependence
temperature appears to be room temperature!). He con- eikx . We can put periodic boundary conditions
sidered the same model Boltzmann did, but treated the
oscillations quantum-mechanically. eikx = 1. (24)
5
FIG. 6. Albert Einstein (1879 - 1955) was an ethni-

cally Jewish, German-born, probably the greatest theoretical
physicist of the 20th century who is best known for his theo-
ries of special relativity and general relativity. He also made
FIG. 4. Plot of molar heat capacity of diamond from Ein- important contributions to statistical mechanics: theories of
stein’s original paper. The fit is to the Einstein theory. The Brownian motion, of specific heat, and of fluctuations and
y axis is C in units of cal/(K·mol). In these units, 3R ≈ 5.96. dissipation, and to quantum mechanics and, indirectly, quan-
The fitting parameter TEinstein = ~ω/kB is roughly 1320K. tum field theory, primarily through his theoretical studies of
Figure from A. Einstein, Ann. Phys., 22, 180, (1907). the photon.
FIG. 7. Max Karl Ernst Ludwig Planck (1858-1947)

was a German theoretical physicist whose discovery of energy
quanta won him the Nobel Prize in Physics in 1918. Planck
made many contributions to theoretical physics, but his fame
as a physicist rests primarily on his role as the originator of
quantum theory, which revolutionized human understanding
of atomic and subatomic processes.
In three dimensions the space dependence is eik·r .

Hence
FIG. 5. Heat capacity of diamond is proportional to T 3 at
low temperature. Note that the temperatures shown in this 2π
k= (n1 , n2 , n3 ). (27)
plot are far far below the Einstein temperature and therefore L
correspond to the very bottom left corner of Fig. 4.
As a result
Z ∞
X L3
→ dk. (28)
Hence (2π)3 −∞
k
2πn
k= , (25) Following Planck, Debye has written (there are three po-
L larizations not two)
and the summation with respect to k can be changed to X
integration U =3 ~ω(k)[nB (ω(k)) + 1/2]
k
L ∞
Z
L3
X Z
→ dk. (26) =3 dkω(k)[nB (ω(k)) + 1/2]. (29)
2π −∞ (2π)3
k
6
is called Riemann ζ function. In general

Z ∞
dxxs−1
= Γ(s)ζ(s). (36)
0 ex − 1
The zeta function of even argument can be calculated
analytically. We’ll present Euler’s approach to such cal-
culation, based on the identity
∞
z2

sin z Y
= 1− 2 2 . (37)
z n=1
π n
Expanding both sides we get

FIG. 8. Peter Joseph William Debye (1884 – 1966) was ∞
a Dutch-American physicist and physical chemist. He made z2 z4 X 1
1− + − · · · = 1 − z2
important contributions to several fields of physics, including 6 120 n=1
π n2
2
solid state physics, and received Nobel Prize in Chemistry for ∞
his studies of molecular structure through the diffraction of z4 X 1 1
X-rays and electrons in gases.
+ − .... (38)
2 π 2 n2 π 2 m 2
n,m=1;n6=m
Problem D.1 Show that

Material TDebye (K)
Diamond (C) 1850 π2 π4
Beryllium (Be) 1000 ζ(2) = , and ζ(4) = . (39)
6 90
Silicon (Si) 625
Copper (Cu) 315 Thus we get
Silver (Ag) 215
∞
dxx3 π4
Z
Lead (Pb) 88
x
= . (40)
0 e −1 15
TABLE II. Some Debye temperatures.
Hence the heat capacity is
Eq. (29) can be presented as 12π 4 T 3

C= R 3. (41)
5 TD
9N ~ ∞
Z
U= 3 dωω 3 [nB (ω) + 1/2], (30)
ωD 0 To recover Dulong – Petit law Debye introduced max-
imum frequency cutoff ωc necessary to obtain a total of
where we defined the Debye frequency by the equation only 3N degrees of freedom
3 6π 2 N v 3 Z ωc
ωD = ; (31) 3N = dωg(ω). (42)
V 0
the latter is sometimes replaced by the Debye tempera- It turns out that ωc = ωD . Hence
ture
9N ~ ωD dωω 3
Z
kB TD = ~ωD . (32) U= 3 . (43)
ωD 0 eβ~ω − 1
Ignoring the T independent terms we obtain
Problem D.2 Show that for T TD from Eq. (43)
9N ~ ∞ dωω 3 9N (kB T )4 ∞ dxx3 follows
Z Z
U= 3 = .
ωD 0 eβ~ω − 1 (~ωD )3 0 ex − 1
C = 3R, (44)
(33)
and for T TD specific heat calculated from Eq. (43)
Z ∞ 3 ∞ Z ∞ ∞ coincides with that calculated from Eq. (33).
dxx X X 1
= dxx3 e−nx =6 . (34)
0 ex − 1 n=1 0 n=1
n4 For intermediate temperatures integral in Eq. (43)
should be calculated numerically. It should be empha-
The function sized that the Debye theory makes predictions without
∞ any free parameters, as compared to the Einstein theory
X 1 which had the unknown Einstein frequency ω as a free
ζ(z) = z
(35)
n=1
n fitting parameter.
7
FIG. 10. Potential between neighboring atoms (thin black).

The thick light gray curve is a quadratic approximation to
the minimum.
FIG. 9. Heat capacity of silver compared to the Debye and

Einstein models. The high-temperature asymptote is given by The usual description of compressibility is
C = 3R = 24.945 J/(mol- K). Over the entire experimental
range, the fit to the Debye theory is excellent. At low T it cor- 1 ∂V
rectly recovers the T 3 dependence, and at high T it converges β=− (47)
V ∂P
to the law of Dulong–Petit. The Einstein theory clearly is
incorrect at very low temperatures. The Debye temperature (one should ideally specify if this is measured at fixed
is roughly 215 K, whereas the Einstein temperature roughly T or at fixed S. Here, we are working at T = S = 0
151 K. for simplicity). In the onedimensional case, we write the
compressibility as
Problem D.3 Einstein versus Debye. In both the 1 ∂L 1

β=− = (48)
Einstein model and the Debye model the high-temperature L ∂F κa
heat capacity is of the form You may recall from your study of fluids that in an
isotropic compressible fluid, one predicts sound waves
C = N kB (1 − κ/T 2 + . . . ). (45) with velocity
s
a) For the Einstein model calculate κ in terms of the B
r
1
Einstein temperature. v= = , (49)
ρ ρβS
b) For the Debye model calculate κ in terms of the Debye
temperature. From your results give an approximate ratio where ρ is the mass density of the fluid,B is the bulk
TEinstein /TDebye . Compare your result to the values for modulus, which is B = 1/βS with βS the adiabatic com-
silver given in Fig. 9. (The ratio you calculate should be pressibility. Thus using our previous result, we predict a
close to the ratio stated in the caption of the figure. It is sound wave with velocity
not exactly the same though. Why might it not be?) r
κa2
v= . (50)
m
III. ONE-DIMENSIONAL CHAIN
Problem A.2 Classical Model of Thermal Expan-
sion.
A. Compressibility, Sound, and Thermal Expansion
(i) In classical statistical mechanics, we write the expec-
tation of x as
Let us imagine a one-dimensional system of atoms
dxxe−βV (x)
R
(atoms in a single line). The potential V (x) between
< x >β = R (51)
two neighboring atoms is drawn in Fig. 10. dxe−βV (x)
Let us now Taylor expand the potential around its min-
imum. Substituting x = xeq + u we obtain Although one cannot generally do such integrals for ar-
bitrary potential V (x), one can expand the exponentials
κ κ3 as
V (x) = V (xeq ) + (x − xeq )2 − (x − xeq )3 + . . (46)
..
2 3! βκ 2
h κ3 i
e−βV (x) = e− 2 (x−xeq ) 1 + (x − xeq )3 + . . . (52)
Problem A.1 Why there is no linear term in the expan- 3!
sion (46)? and let limits of integration go to ±∞.
8
a) Why is this expansion of the exponent and the exten-

sion of the limits of integration allowed?
b) Use this expansion to derive < x >β to lowest or-
der in κ3 , and hence show that the coefficient of thermal
expansion is
1 dL 1 d < x >β 1 kB κ3
α= ≈ = (53)
L dT xeq dT xeq 2κ2
c) In what temperature range is the above expansion

valid?
d) While this model of thermal expansion in a solid is
valid if there are only two atoms, why is it invalid for the
case of a many-atom chain? (Although actually it is not
so bad as an approximation!) FIG. 12. Dispersion relation for vibrations of the one-
dimensional harmonic chain.
B. Vibrations of A One-Dimensional Chain
For small oscillations we can ignore the higher order

terms in the expansion of the potential, and substituting
x = xeq + u we obtain a simple harmonic chain with the
Lagrangian
" #
X m dun 2 κ
2
L= + (un+1 − un ) . (54)
n
2 dt 2
The equation of motion is
FIG. 13. Paul Adrien Maurice Dirac (1902 - 1984) was a

British theoretical physicist. Dirac made fundamental contri-
butions to the early development of both quantum mechanics
and quantum electrodynamics. Dirac shared the Nobel Prize
in physics for 1933 with Erwin Schrodinger
FIG. 11. The one-dimensional harmonic chain. Each ball has C. Canonical Quantization
mass m and each spring has spring constantκ. The lattice
constant, or spacing between successive masses at rest is a.
Canonical quantization is one of several procedures
for quantizing a classical theory. Paul Dirac introduced
d2 un it in his 1926 doctoral thesis, the ”method of classical
m = κ(un+1 + un−1 − 2un ). (55) analogy” for quantization. The word canonical arises
dt2 from the Hamiltonian approach to classical mechanics,
The solution is in which a system’s dynamics is generated via canonical
Poisson brackets. In quantum mechanics Poisson brack-
un = Ae−iω(k)t+ikna , (56) ets turn into commutators.
Consider harmonic oscillator Hamiltonian
where
p̂2 mω02 q̂ 2
2
ω (k) = 2ω02 (1 − cos ka), (57) H(p, q) = + , (60)
2m 2
where ω02 = κ/m. where the operators q and p satisfy commutation relation
For ka 1 we obtain
[q̂, p̂] = i~. (61)
ω(k) = aω0 k. (58)
Let us introduce new operator
Thus we predict a sound wave with velocity r r
mω0 1
v = aω0 . (59) â = q̂ + i p̂. (62)
2~ 2m~ω0
9
Operator â is not Hermitian; Solving Eqs. (62) and (63) we obtain

r r r
† mω0 1 q̂ =
~
[â + â† ] (75)
â = q̂ − i p̂. (63) 2mω0
2~ 2m~ω0 r
m~ω0
The Hamiltonian in the new operators is p̂ = −i [â − â† ]. (76)
2

1 Thus we can calculate matrix elements of any function
H = ~ω0 â† â + . (64)
2 of operators q̂ and p̂.
From Eq. (61) we obtain commutation relations for the Problem C.1 Use Eq. (74) in Eq. (64) to confirm the
new operators eigenvalues and eigenvectors of the Harmonic oscillator.
The eigenfunctions of the Hamiltonian can be written
[â, â† ] = 1. (65)
as
n
To obtain spectrum and eigenfunctions of the Hamil- â†
tonian (64) consider operator |n >= √ |0 > . (77)
n
N̂ = â† â. (66) We can also easily find the eigenfunctions in q represen-
tation. Wave function of the ground state is found Eq.
It is a Hermitian operator. Consider some eigenvector (71), which in q representation is
of these operator with the eigenvalue n (at the moment r r !
considering it some real number) mω0 ~ ∂
q+ ψ0 (q) = 0. (78)
2~ 2mω0 ∂q
N̂ |n >= n|n > . (67)
The solution is
From the commutation relations we obtain
2
ψ0 (q) = e−mω0 q /2~
. (79)
N̂ â|n >= (n − 1)â|n > . (68)
Acting consecutively by operator â† we can obtain higher
Hence â|n > is also an eigenvector of N̂ , corresponding eigenstates.
to eigenvalue n − 1.
Problem C.2 Write explicitly ψ1 (q) and ψ2 (q).
â|n >= |n − 1 > . (69)
Problem C.3 Using the approach due to Dirac, derive
Thus from a single eigenvector with the eigenvalue n we expressions for < q >, < q 2 >, < p >, and < p2 >
obtain a succession of eigenvectors with the eigenvalues for the harmonic oscillator in eigenstate n. Use this to
n − 1, n − 2, n − 3 . . . . However, any eigenvalue of the obtain the uncertainty product ∆q∆p. Comment on your
operator N̂ must be non-negative. In fact, result.
n =< n|N̂ |n >=< n|â† â|n >=< n − 1|n − 1 >≥ 0. (70)
D. Phonons and Improvement of Debye’s
Hence, the succession mentioned above should terminate.
Calculation
It can happen only in the case if the initial eigenvalue we
started from (and hence all the other eigenvalues) are
integers, and From the Lagrangian (54) we can obtain the Hamilto-
nian
â|0 >= 0. (71) X p2 κ

n 2
H= + (un+1 − un ) . (80)
Also from the commutation relations we obtain n
2m 2
N̂ â† |n >= (n + 1)â† |n >; (72) Fourier transforming the variables

r
thus the spectrum of operator N̂ is semi-infinite, and the 1 X ~
un = √ âk + â†−k eikn
spectrum of the Hamiltonian (64) is N k 2mωk
r
1 1 X ~mωk
En = ~ω0 n + . (73) pn = −i √ âk − â†−k eikn (81)
2 N 2
k
Eq. (68) and (72) can be rewritten as we obtain the Hamiltonian in the form
√
< n0 |â|n > = δn0 n−1 n

X † 1
√ H= ~ω(k) âk âk + . (82)
< n0 |â† |n > = δn0 n+1 n + 1 (74) k
2
10
we obtain the dispersion relation
M1 M2 ω 4 − 2κ(M1 + M2 )ω 2 + 2κ2 (1 − cos ka) = 0.

(87)
FIG. 15. The one-dimensional harmonic chain with diatomic

basis.
FIG. 14. Satyendra Nath Bose (1894 - 1974) was an Indian
physicist, specializing in mathematical physics. He is best
We can solve Eq. (87) exactly, but it is simpler to
known for his work on quantum mechanics in the early 1920s,
providing the foundation for Bose-Einstein statistics and the examine the limiting cases ka 1 and ka = ±π at the
theory of the Bose–Einstein condensate. He is honoured as zone boundary. For small ka we have
the namesake of the boson. κ
ω2 = k 2 a2 (acoustical branch) (88)
2(M1 + M2 )

Problem D.1 Check up that introduced above operator 1 1
ω 2 = 2κ + (optical branch). (89)
â†k and âk are creation and annihilation operators, i.e. M1 M2
satisfy the appropriate commutation relations.
At kmax = ±π/a the roots of Eq. (87) are
The energy of each harmonic oscillator is quantized

1
ω 2 = 2κ/M1 ω 2 = 2κ/M2 . (90)
En (k) = ~ω(k) n + . (83)
2 The dependence of ω on k is shown in Fig. 16 for M1 >
We can think of the quanta of the chain oscillations as M2 .
quasiparticles (bosons) having energy ~ω(k) and crystal
momentum ~k. Problem E.1 Find dispersion law for the oscillations of
Thus for the chain we obtain the diatomic linear chain of identical atoms, connected by
springs of alternating strengths.
N a π/a
Z
U= dk~ω(k)[nB (ω(k)) + 1/2]. (84)
2π −π/a
For d = 3, if there are p atoms in the primitive cell,
Thus we improved Eq. (29) (in 1d). Debye only knew there are 3p branches to the dispersion relation: 3 acous-
about sound, where ω(k) = vk is linear in the wavevec- tical branches and 3p 3 optical branches. Thus germa-
tor. We, on the other hand, have calculated ω(k) for our nium (Fig. 17) with two atoms in a primitive cell, has
microscopic mass and spring model (and have found that six branches: one LA, one LO, two TA, and two TO.
the dispersion is linear in k only for small k).
Problem D.2 Using Eq. (84) analyse behavior of the

heat capacity of the chain for high and low temperatures.
E. Diatomic Basis
Consider a chain with two atoms in the primitive cell.

The dynamic equations are
d2 un
M1 = κ(vn+1 + vn−1 − 2un )
dt2
d2 vn
M2 2 = κ(un+1 + un−1 − 2vn ). (85)
dt
On substitution
FIG. 16. Optical and acoustical branches of the dispersion
un = ue−iωt+ikna relation for a diatomic linear lattice, showing the limiting fre-
vn = ve−iωt+ikna (86) quencies at k = 0 and k = π/a. The lattice constant is a.
11
tron returns to momentum p = 0.

(3) In between scattering events, the electrons, which are
charge −e particles, respond to the externally applied
electric field E and magnetic field B.
The main equation of Drude theory is

dp p×B p
= −e E + − . (91)
dt m τ
In Eq. (91) by p we mean the averaged value of p.
A. Conductivity
Consider Drude equation in the absence of magnetic

field
dp p
= −eE − . (92)
FIG. 17. Phonon dispersion relations in the [111] direction in dt τ
germanium at 80 K. The two TA phonon branches are hor- The steady state solution is
izontal at the zone boundary position, Kmax = 2π 1 1 1

a
, , .
2 2 2
The LO and TO branches coincide at K = 0; this also is a mv = p = −eτ E. (93)
consequence of the crystal symmetry of Ge. The results were
obtained with neutron inelastic scattering. Hence the electric current is
ne2 τ
j= E, (94)
m
and the conductivity σ is
ne2 τ
σ= . (95)
m
Let E = E0 eiωt . Then setting p = p0 eiωt we obtain
ne2 /m 0 Ω2p
σ= 1 = 1 . (96)
τ + iω τ + iω
Problem A.1 Plot real σ1 (ω) and imaginary part σ2 (ω)

of the conductivity given by Eq. (96).
FIG. 18. Paul Karl Ludwig Drude (1863 1906) was a Ger- In steady state in the presence of magnetic field
man physicist of Jewish origin specializing in optics. He wrote 1 m
a fundamental textbook integrating optics with Maxwell’s E= j×B+ 2 j (97)
theories of electromagnetism. In 1900 he developed a powerful
ne ne τ
model to explain the thermal, electrical, and optical proper- We can introduce resistivity matrix ρ̂ by the equation
ties of metals.
E = ρ̂j (98)
IV. ELECTRONS IN METALS: DRUDE m

ρxx = ρyy = ρzz = (99)
THEORY ne2 τ
and if B is oriented in the z direction
The defining characteristic of a metal is that it con- B
ducts electricity. At some level the reason for this con- ρxy = −ρyx = . (100)
duction boils down to the fact that electrons are mobile ne
in these materials. Hence Hall coefficient RH , which is defined as
In 1900 Paul Drude realized that he could apply Boltz- ρyx
mann’s kinetic theory of gases to understanding electron RH = , (101)
|B|
motion within metals. Drude made three assumptions
about the motion of electrons in Drude theory is given by
(1) Electrons have a scattering time τ . The probability 1
of scattering within a time interval dt is dt/τ . RH = − . (102)
ne
(2) Once a scattering event occurs, we assume the elec-
12
FIG. 20. Edwin Hall’s 1879 experiment. The voltage mea-

sured perpendicular to both the magnetic field and the cur-
rent is known as the Hall voltage which is proportional to B
and inversely proportional to the electron density (at least in
Drude theory).
FIG. 19. Frequency dependence of the (a) real and (b) imagi-
nary parts of the optical conductivity of silicon weakly doped
FIG. 21. Comparison of the valence of various atoms to the
by holes (1.1×1015 cm−3 , open circles) or electrons (4.2×1014
valence predicted from the measured Hall coefficient. Here
cm−3 , solid circles). The experiments were performed at
natomic is the density of atoms in the metal and RH is the mea-
T = 300 K, the full lines correspond to σ1 (ω) and σ2 (ω) as
sured Hall coefficient. In Drude theory, the middle column
obtained by the Drude model.
should give the number of electrons per atom, i.e., the valence.
For monovalent atoms, the agreement is fairly good. But for
divalent atoms, the sign can even come out wrong! The ∗
Problem A.2 Estimate the magnitude of the Hall volt- next to Be indicates that its Hall coefficient is anisotropic.
age for a specimen of sodium in the form of a rod of Depending on which angle you run the current you can get
rectangular cross-section 5mm by 5mm carrying a cur- either sign of the Hall coefficient!
rent of 1A down its long axis in a magnetic field of 1T
perpendicular to the long axis. The density of sodium
atoms is roughly 1 gram/cm3 , and sodium has atomic where jq is the heat current density.
mass of roughly 23. You may assume that there is one From kinetic theory of gas follows
free electron per sodium atom (sodium has valence 1). κ ∼ ncv v̄λ, (104)
where λ = v̄τ is the scattering length,
Problem A.3 Define the resistivity matrix ρ̂ as
3
E = ρ̂J. cv = kB (105)
2
is the heat capacity per particle, and
Use Drude theory to derive an expression for the matrix r
ρ̂ for a metal in a magnetic field. (You may assume B kB T
parallel to the z axis. Invert this matrix to obtain an v̄ ∼ (106)
m
expression for the conductivity matrix σ.
is the average velocity of a particle. Hence
2
nτ kB T
κ∼ . (107)
B. Thermal Transport m
We may look at the ratio of thermal conductivity to elec-
The thermal conductivity κ is defined by trical conductivity, known as the Lorenz number
κ
jq = κ∇T (103) L= , (108)
Tσ
13
which according to Drude theory is equal to low resistivity, because running a current through a ma-
2 terial with resistivity R will result in power I 2 R being
kB 2 dissipated, thus heating it up.
L= ∼ 10−8 Watt · Ohm/K . (109)
e
FIG. 22. Lorenz numbers κ/(T σ) for various metals in units FIG. 24. Single couple of a thermoelectric generator.
of 10−8 Watt·Ohm/K2 . The prediction of Drude theory is
that the Lorentz number should be on the order of 1 × 10−8
Watt·Ohm/K2 .
V. SOMMERFELD (FREE ELECTRONS)

The so-called Peltier effect is the fact that running elec-
THEORY
trical current through a material also transports heat.
The Peltier coefficient Π is defined by
A. Basic Fermi-Dirac Statistics
jq = Πj. (110)
In kinetic theory Given a system of free electrons with chemical poten-
tial µ the probability of an eigenstate of energy being
jq ∼ cv T nv. (111) occupied is given by the Fermi factor
Thus 1
nF () = . (114)
kB T eβ(−µ) + 1
Π∼− (112)
e The concentration of electrons is
so the ratio, (known as thermopower, or Seebeck coeffi- Z ∞
dg()
cient) S = Π/T is given by n= , (115)
0 1 + exp µ−
T
Π kB
S= ∼− ∼ −10−3 V/K. (113) where g()
T e
√
2 2 dk 2m3 √
g() = 3
4πk = . (116)
(2π) d π 2 ~3
At T = 0 we have
p p
2 2m3 µ3 2 2m3 EF3
n= = . (117)
3π 2 ~3 3π 2 ~3
Since we know roughly how many free electrons there are
in a metal (say, one per atom for monovalent metals such
as sodium or copper), we can estimate the Fermi energy,
FIG. 23. Seebeck coefficients of various metals at room tem- which, say for copper, turns out to be on the order of 7
perature, in units of 10−6 V/K. Note that the magnitude of eV, corresponding to a Fermi temperature (EF = kB TF )
the Seebeck coefficient is roughly one hundredth of the value of about 80,000 K(!).
predicted by Drude theory. For Cu and Be, the sign comes
out wrong as well! Problem A.1 Assuming that the free electron theory is
applicable to copper:
The Peltier effect is used for thermoelectric refrigera- (a) calculate the values of both Drude’s v̄ and vF for
tion devices. Running electricity through a thermoelec- copper at 300K in an electric field of 1 V m−1 and com-
tric material forces heat to be transported through that ment on their relative magnitudes.
material. You can thus transport heat away from one (b)estimate λ for copper at 300K and comment upon
object and towards another. A good thermoelectric de- its value compared to the mean spacing between the copper
vice has a high Peltier coefficient, but must also have a atoms. You will need the following information: copper is
14
B. Electronic Heat Capacity
The grand canonic potential is

X
Ω = −T ln 1 + eβ(µ−k )
k
√
2V T m3/2 ∞ √
Z
=− 2 3
d ln 1 + eβ(µ−)
π ~ 0
√
2 2V m3/2 ∞ d3/2
Z
=− . (118)
3π 2 ~3 0 eβ(−µ) + 1
FIG. 25. Enrico Fermi (1901 - 1954) was an Italian physicist High density of electrons results in extremely high
noted for his contributions to the development of quantum Fermi energy and Fermi velocity. Thermal and electric
theory, nuclear and particle physics, and statistical mechanics. excitations are small redistributions of electrons around
Fermi was awarded the Nobel Prize in Physics in 1938 for the Fermi surface. Consider
his work on induced radioactivity and is acknowledged as a Z ∞
unique physicist who was highly accomplished in both theory f ()d
I= β(−µ) + 1
. (119)
and experiment. 0 e
After substitution we get
Z ∞ Z µ/T
f (µ + T z) f (µ − T z)
I= T dz = T dz
−µ/T e z +1
0 e−z + 1
Z ∞
f (µ + T z)
+T T dz. (120)
0 ez + 1
In the first integral we write
1 1
=1− z (121)
e−z +1 e +1
to get
µ µ/T
f (µ − T z)
Z Z
I= f ()d − T dz
0 0 ez + 1
FIG. 26. Arnold Johannes Wilhelm Sommerfeld (1868- Z ∞
1951) was a German theoretical physicist who pioneered de-
f (µ + T z)
+T dz. (122)
velopments in atomic and quantum physics, and also educated 0 ez + 1
and mentored a large number of students for the new era of
theoretical physics. He served as PhD supervisor for many In the second integral we change upper limit to ∞ to
Nobel Prize winners in physics and chemistry. obtain
Z µ Z ∞
f (µ + T z) − f (µ − T z)
I= f ()d + T dz.
0 0 ez + 1
(123)
monovalent, meaning there is one free electron per atom. Expanding we get
The density of atoms in copper is n = 8.45 × 1028 m−1 . Z µ Z ∞
zdz
The conductivity of copper is σ = 5.9 × 107 Ω−1 m−1 at I= 2 0
f ()d + 2T f (µ) + ...
z
e +1
300K. 0
Z µ
0
π2 2 2 0
= f ()d + k T f (µ) + . . . (124)
0 6 B
Hence
√
2 2 2µm3/2
Ω = Ω0 − V kB T . (125)
Problem A.2 Estimate the value of EF for sodium [The 6~3
density of sodium atoms is roughly 1 gram/cm3 , and According to theorem of small increments, which claims
sodium has atomic mass of roughly 23. You may assume that
that there is one free electron per sodium atom (sodium
has valence one)] (δF )T,V,N = (δΩ)T,V,µ , (126)
15
we get of T /TF ∼ 0.01, we can return to re-examine some of the

above Drude calculations of thermal transport. We had
π 2 kB T 2 found that Drude theory predicts a thermopower that is
F = F0 − N . (127)
4 TF too large by a factor of 100. Now it is clear that the
(If you start from the free energy instead of the grand reason for this error was that we used in this calculation
canonical potential, you have to worry about the fact that the heat capacity per electron for a classical gas, which
the chemical potential shifts as a function of temperature, is too large by roughly TF /T ∼ 100. If we repeat the cal-
which makes the calculations much more complicated.) culation using the proper heat capacity, we will now get
Thus a prediction for thermopower which is reasonably close
in magnitude to what is actually measured in experiment
π2 T for many metals.
S= N kB , (128)
2 TF We also used the heat capacity per particle in the
and Drude calculation of the thermal conductivity too large
by a factor of TF /T , but on the other hand the value of v̄ 2
π2 T that he used was too small by roughly the same factor.
C= N kB (129)
2 TF Thus Drudes prediction for thermal conductivity came
out roughly correct (and thus the Wiedemann-Franz law
correctly holds).
Problem B.1 (a) Show that the kinetic energy of a free

electron gas in three dimensions is E = 53 EF N .
(b) Calculate the pressure P = −∂E/∂V , and then the
bulk modulus B = −V ∂P/∂V .
(c) Given that the density of atoms in sodium is 2.53 ×
1022 cm−3 and that of potassium is 1.33 × 1022 cm−3 , and
given that both of these metals are monovalent (i.e., have
one free electron per atom), calculate the bulk modulus
associated with the electrons in these materials. Compare
your results to the measured values of 6.3 GPa and 3.1
FIG. 27. Heat capacity divided by temperature of silver at GPa respectively.
very low temperature plotted against temperature squared.
At low enough temperature one can see that the heat capacity
is actually of the form C = γT + αT 3 . If the dependence were
purely T 3 , the curve would have a zero intercept. The cubic C. Magnetic Susceptibility
term is from the Debye theory of specific heat. The linear
term is special to metals. For a small magnetic field, the magnetization of a sys-
tem M (magnetic moment per unit volume) is typically
related linearly to the applied magnetic field H by a
(magnetic) susceptibility χ. We write for small fields H,
M = χH. (130)
Both M and H are measured in amperes per meter; χ is

therefore a dimensionless quantity.
The susceptibility we defined above (and will use both
in this Section and in Section XV) is the so called volume
susceptibility, sometimes designated as χv . There are
FIG. 28. Low-temperature heat capacity coefficient for some
two other measures of susceptibility, the mass magnetic
metals. All of these metals have heat capacities of the form
C = γT + αT 3 at low temperature. This table gives the
susceptibility (χmass or χg , sometimes χm ), measured in
measured experimental (exp) value and the Sommerfeld the- m3 kg−1 and the molar magnetic susceptibility (χmol )
oretical (th) predictions for the coefficient γ in units of 10−4 J measured in m3 mol−1 that are defined below, where ρ
mol−1 K−2 . The theoretical value is obtained by setting the is the density and M is molar mass in kg mol−1 .
electron density equal to the atomic density times the valence
(number of free electrons per atom). χv
χmass =
ρ
Realizing now that the heat capacity of the electron M χv
gas is reduced from that of the classical gas by a factor χmol = M χmass = (131)
ρ
16
FIG. 30. Filling of electronic states up to the Fermi energy.

FIG. 29. Wolfgang Ernst Pauli (1900 - 1958) was an Aus- Left: Before the magnetic field is applied the density of states
trian theoretical physicist of Jewish origin and one of the pi- for spin-up and spin-down are the same g↑ (E) = g↓ (E) =
oneers of quantum physics. In 1945 Pauli received the Nobel g(E)/2. Right: When the magnetic field is applied, the states
Prize in Physics for his ”decisive contribution through his dis- with up and down spin are shifted in energy by +µB B and
covery of a new law of Nature, the exclusion principle or Pauli −µB B respectively as shown. Hence up spins pushed above
principle”. the Fermi energy can lower their energies by flipping over to
become down spins. The number of spins that flip (the area of
the approximately rectangular sliver) is roughly g↑ (EF )µB B.
D. Pauli Paramagnetism
Another property we can examine about the free elec-

tron gas is its response to an externally applied magnetic
field. There are several ways that the electrons can re-
spond to the magnetic field. First, the electron motion
can be curved due to the Lorentz force. Secondly, the
electron spins can flip over due to the applied magnetic
field, which also changes the magnetic moment of the
electron gas: this is the effect we will focus on here.
p2
H= + gµB B · σ, (132)
2m FIG. 31. Experimentally measured magnetic susceptbili-
where g = 2 is the g–factor of the electron, B is the mag- ties χexp of various metals compared to Pauli’s theoreti-
netic field, σ is the spin of the electron which takes eigen- cal predictionχP auli . In both cases the susceptibility is di-
values ±1/2, and µB = e~/2me = .67kB (Kelvin/Tesla) mensionless and is listed here in units of 10−6 (e.g., Li has
is the Bohr magneton. From Fig. 30 follows that χexp = 3.4 × 10−6 ). The theoretical calculation assumes the
bare mass of the electron for m.
dM
χP = = µ0 µ2B g(EF ). (133)
dH
Problem D.1 The Wilson ratio is defined as with specific, characteristic orientations. Examples of
large crystals include snowflakes, diamonds, and table
π 2 kB
2
T χP salt. Most inorganic solids are not crystals but polycrys-
Rw = 2 , (134)
3µ0 µB Ce tals, i.e. many microscopic crystals fused together into a
single solid. Examples of polycrystals include most met-
where Ce is the electron heat capacity. Calculate Wilson als, rocks, ceramics, and ice. A third category of solids
ratio for Sommerfeld model. is amorphous solids, where the atoms have no periodic
structure whatsoever. Examples of amorphous solids in-
clude glass, wax, and many plastics.
VI. CRYSTALS
A crystal or crystalline solid is a solid material whose

A. Bravais Lattices
constituents (such as atoms, molecules, or ions) are ar-
ranged in a highly ordered microscopic structure, form-
ing a crystal lattice that extends in all directions. In In geometry and crystallography, a Bravais lattice,
addition, macroscopic single crystals are usually identi- named after Auguste Bravais (1850), is an infinite array
fiable by their geometrical shape, consisting of flat faces of discrete points generated by a set of discrete transla-
17
tion operations described by (in 3d): • Primitive (P): lattice points on the cell corners only
(sometimes called simple)
R = n1 a1 + n2 a2 + n3 a3 , (135) • Base-centered (A, B, or C): lattice points on the
cell corners with one additional point at the center
with a1 , a2 and a3 being the primitive lattice vectors and of each face of one pair of parallel faces of the cell
n1 , n2 and n3 being integers. (sometimes called end-centered)
A unit cell is a region of space such that when many
• Body-centered (I): lattice points on the cell corners,
identical units are stacked together it tiles (completely
with one additional point at the center of the cell
fills) all of space and reconstructs the full structure. A
• Face-centered (F): lattice points on the cell corners,
primitive unit cell is a unit cell containing exactly one
with one additional point at the center of each of
lattice point. Primitive cell may be chosen following this
the faces of the cell
procedure: (1) draw lines to connect a given lattice point
to all nearby lattice points; (2) at the midpoint and nor- Problem A.1 Consider a face-centered-cubic cell. Con-
mal to these lines, draw new lines or planes. The smallest struct a primitive cell within the conventional cell, and
volume enclosed in this way is the Wigner-Seitz primitive compare the two. How many atoms are in the primitive
cell. All space may be filled by these cells. Given a lat- cell, and how does this compare with the number in the
tice point, the Wigner - Seitz cell is the set of all points original cell? Do the same for a body-centered-cubic cell.
in space which are closer to that given lattice point than
to any other lattice point. A conventional unit cell is the Not all combinations of lattice systems and centering
smallest lattice cell that has the overall symmetry of a types are needed to describe all of the possible lattices, as
crystal lattice. It can be a non-primitive unit cell. it can be shown that several of these are in fact equivalent
to each other. For example, the monoclinic I lattice can
be described by a monoclinic C lattice by different choice
of crystal axes. Similarly, all A- or B-centred lattices can
be described either by a C- or P-centering. This reduces
the number of combinations to 14 conventional Bravais
lattices, shown in Fig. 181.
FIG. 32. Some unit cells for the hexagonal lattice.
A lattice system is a class of lattices with the same

set of lattice point groups. In 3d there are seven lattice
systems: triclinic, monoclinic, orthorhombic, tetragonal,
rhombohedral, hexagonal and cubic.
There are 14 Bravais lattices. These are obtained by FIG. 33. In two-dimensional space, there are 5 Bravais lat-
combining one of the seven lattice systems with one of tices: 1 - square (tetragonal), 2 - hexagonal, 3 - oblique (mon-
the centering types. The centering types identify the oclinic), 4 -rectangular primitive (orthorhombic), and 5 - cen-
tered rectangular (orthorhombic).
locations of the lattice points in the conventional unit
cell as follows:
Although the simple cubic lattice is conceptually the of atoms as small spheres that weakly attract each other
simplest of all lattices, in fact, real crystals of atoms are and therefore try to pack close together. When you as-
rarely simple cubic. To understand why this is so, think semble spheres into a simple cubic lattice you find that it
18
FIG. 34. The 14 Bravais lattices
is a very inefficient way to pack the spheres together - the Calculate the packing fraction for
packing fraction is π6 ≈ .52. The bcc and fcc lattices leave (a) a simple cubic lattice.
much less open space between the spheres than packing (b) a bcc lattice.
the spheres in√a simple cubic lattice; they correspond to (c) a fcc lattice.
the fraction π 8 3 ≈ .68 and 3√ π
2
≈ .74 respectively. Cor- (d) a rhombohedral lattice.
respondingly, bcc and fcc lattices are realized much more (e) a hcp structure (see below).
frequently in nature than simple cubic (at least for met- The fact that it is impossible to pack spheres more
als). For example, the elements Al, Ca, Au, Pb, Ni, Cu, densely than you would get by placing the spheres at the
Ag (and many others) are fcc whereas the elements Li, vertices of an fcc lattice was first conjectured by Johannes
Na, K, Fe, Mo, Cs (and many others) are bcc. Kepler in 1611, but was not mathematically proven until
1998! (It is infinite number of packing realizing the same
Problem A.2 Packing Fractions. Consider a lattice packing fraction.)
with a sphere at each lattice point. Choose the radius of
the spheres to be such that neighboring spheres just touch.
The packing fraction is the fraction of the volume of all B. Crystal Structures
of space which is enclosed by the union of all the spheres
(i.e., the ratio of the volume of the spheres to the total The structure of all crystals can be described in terms
volume). of a Bravais lattice, with a group of atoms attached to ev-
19
ery lattice point. The group of atoms is called the basis; Problem B.1 Calculate the number of atoms in the di-
when repeated in space it forms the the whole crystal. amond conventional cell.
In other words, the basis consists of a primitive cell, con-
taining one single lattice point.
1. Diamond Structure
The diamond structure is the structure of the semicon-

ductors silicon and germanium and is related to the struc-
ture of several important semiconductor binary com-
pounds. Other element in group 14 which adopts this
structure as solidifies is grey (α) tin. 2. Hexagonal Close Packed Structure
There is another crystal structure, the hexagonal close

packed structure (hcp) which achieves precisely the same
packing density for spheres as the fcc lattice. The hcp
structure is characterised by two nestedhexagonal lattice
that are shifted by the vector 32 , 13 , 12 (in the conven-
tional unit cell basis) against each other. The underlying
lattice is not a Bravais lattice since the individual lattice
points are not equivalent with respect to their environ-
ments. But it can be looked at as a hexagonal Bravais
FIG. 35. Diamond structure, showing the tetrahedral bond lattice with a two-atomic basis with the atoms sitting at
arrangement. The space lattice of diamond is face-centered the positions (0, 0, 0) and 32 , 13 , 21 . We are forced to de-
cubic. The primitive basis has two identical atoms at coordi- scribe the hcp and diamond lattices as lattices with bases
nates 000 and 14 , 14 , 14 . There is no way to choose a primitive by the intrinsic geometrical arrangement of the lattice

cell such that the basis of diamond contains only one atom. points.
C. Allotropy ical elements to exist in two or more different forms, in

the same physical state, known as allotropes of these el-
In 1910 British polar explorer Robert Scott hoped to ements. Allotropes are different structural modifications
be the first to reach the South Pole, but was beaten by of an element; the atoms of the element are bonded to-
Norwegian explorer Roald Amundsen. On foot, the expe- gether in a different manner. For example, the allotropes
dition trudged through the frozen deserts of the Antarc- of carbon include diamond (the carbon atoms are bonded
tic, marching for caches of food and kerosene deposited together in a tetrahedral lattice arrangement), graphite
on the way in. In early 1912, at the first cache, there (the carbon atoms are bonded together in sheets of a
was no kerosene; the cans soldered with tin were empty. hexagonal lattice), graphene (single sheets of graphite),
The cause of the empty tins could have been related to and fullerenes (the carbon atoms are bonded together in
tin pest. spherical, tubular, or ellipsoidal formations).
At 13.2◦ C (about 56◦ F) and below, pure tin trans- Graphite has a layered, planar structure. The individ-
forms from the silvery, ductile metallic allotrope of β form ual layers are called graphene. In each layer, the carbon
white tin (with a tetragonal structure) to the brittle, non- atoms are arranged in a honeycomb lattice with sepa-
metallic, α form grey tin with a diamond cubic structure. ration of 0.142 nm, and the distance between planes is
The transformation is slow to initiate due to a high acti- 0.335 nm.
vation energy but the presence of germanium (or crystal Buckminsterfullerene is a type of fullerene with the for-
structures of similar form and size) or very low tempera- mula C60 . It has a cage-like fused-ring structure (trun-
tures of roughly −30◦ C aids the initiation. There is also cated icosahedron) that resembles a soccer ball (football),
a large volume increase of about 27 % associated with made of twenty hexagons and twelve pentagons, with a
the phase change to the nonmetallic low temperature al- carbon atom at each vertex of each polygon and a bond
lotrope. This frequently makes tin objects (like buttons) along each polygon edge. In solid buckminsterfullerene,
decompose into powder during the transformation, hence the C60 molecules adopt the fcc (face-centered cubic) lat-
the name tin pest. tice.
Allotropy or allotropism is the property of some chem- Carbon nanotubes (CNTs) are allotropes of carbon
20
solid inner core is believed to be composed of liquid iron

mixed with nickel and trace amounts of lighter elements.
Martensite is formed in carbon steels by the rapid cool-
ing (quenching) of the austenite form of iron at such a
high rate that carbon atoms do not have time to diffuse
out of the crystal structure in large enough quantities
to form cementite (Fe3 C). As a result of the quenching,
the face-centered cubic austenite transforms to a highly
strained body-centered tetragonal form called martensite
that is supersaturated with carbon. The shear deforma-
tions that result produce a large number of dislocations,
which is a primary strengthening mechanism of steels.
FIG. 36. The hcp conventional cell. The primitive cell has Plutonium normally has six allotropes and forms a sev-
a1 = a2 , with an included angle of 120◦ . The c axis (or a3 ) is
enth (zeta, ζ) under high temperature and a limited pres-
normal to the plane of a1 and a2 . The ideal hcp structure has
c = 1.633a. The two atoms of one basis are shown as solid
sure range. These allotropes have very similar energy lev-
circles. One atom of the basis is at the origin; the other atom els but significantly varying densities and crystal struc-
is at 32 , 13 , 21 , which means at the position r = 23 a1 + 31 a2 +

tures. This makes plutonium very sensitive to changes in
1
a . temperature, pressure, or chemistry, and allows for dra-
2 3
matic volume changes following phase transitions. Un-
like most materials, plutonium increases in density when
it melts, by 2.5%, but the liquid metal exhibits a lin-
ear decrease in density with temperature. Densities of
the different allotropes vary from 16.00 g/cm3 to 19.86
g/cm3 .
The presence of these many allotropes makes machin-
ing plutonium very difficult, as it changes state very read-
FIG. 37. Some elements forming hcp structure. In general, ily. For example, the α phase exists at room temperature
most (but not all) of the elements having in solid state hexag- in unalloyed plutonium. It has machining characteristics
onal lattice system, have hcp structure. similar to cast iron but changes to the β phase (beta
phase) at slightly higher temperatures. The reasons for
the complicated phase diagram are not entirely under-
with a cylindrical nanostructure. These cylindrical car- stood; recent research has focused on constructing ac-
bon molecules have unusual properties, which are valu- curate computer models of the phase transitions. The
able for nanotechnology, electronics, optics and other α phase has a low-symmetry monoclinic structure, hence
fields of materials science and technology. Owing to the its poor conductivity, brittleness, strength and compress-
material’s exceptional strength and stiffness, nanotubes ibility.
have been constructed with length-to-diameter ratio of Plutonium in the δ phase (delta phase) normally ex-
up to 132, 000, 000 : 1, significantly larger than for any ists in the 310 ◦ C to 452 ◦ C range but is stable at
other material. room temperature when alloyed with a small percent-
Among the metallic elements that occur in nature in age of gallium, aluminium, or cerium, enhancing worka-
significant quantities (56 up to U, without Tc and Pm), bility and allowing it to be welded in weapons applica-
almost half (27) are allotropic at ambient pressure. tions. The delta phase has more typical metallic charac-
Iron represents perhaps the best-known example for ter, and is roughly as strong and malleable as aluminium.
allotropy in a metal. At atmospheric pressure, three al- In fission weapons, the explosive shock waves used to
lotropic forms of iron exist: alpha iron (α), gamma iron compress a plutonium core will also cause a transition
(γ) (also known as austenite), and delta iron (δ). At very from the usual delta phase plutonium to the denser al-
high pressure, a fourth form exists, called epsilon iron pha phase, significantly helping to achieve supercritical-
() hexaferrum (with HCP crystal structure). Some con- ity. The plutonium-gallium alloy is the most common
troversial experimental evidence exists for another high- δ-stabilized alloy.
pressure form that is stable at very high pressures and
temperatures.
VII. CHEMICAL ELEMENTS
The phases of iron at atmospheric pressure are im-
portant because of the differences in solubility of carbon,
forming different types of steel. The high-pressure phases A. Periodic Table
of iron are important as models for the solid parts of
planetary cores. The inner core of the Earth is generally The periodic table (see Fig. 44), proposed originally in
assumed to consist essentially of a crystalline iron-nickel 1869 by Dmitri Mendeleev, is a tabular arrangement of
alloy with structure. The outer core surrounding the the chemical elements, ordered by their atomic number,
21
FIG. 38. Crystal structures of the elements.
whose structure shows periodic trends. Table rows are a group have similar physical or chemical characteristics.
commonly called periods and columns are called groups. The reason for the chemical similarities stems from the
Six groups have accepted names as well as assigned num- details of the fillings of atomic orbital shells. To a large
bers: for example, group 17 elements are the halogens; extent chemistry is determined by the electrons in the
and group 18 are the noble gases. Also displayed are outermost shell of an atom. So for example, the fact
four simple rectangular areas or blocks associated with that the chemistries of C, Si, and Ge are similar is due
the filling of different atomic orbitals. The elements in to the fact that each has only two electrons in a partially
filled p-shell.
Problem A.1 a) Add the missing element names to the A period is a horizontal row in the periodic table. Al-
Table 44. though groups generally have more significant periodic
b) Of the 118 chemical elements, 19 are connected with trends, there are regions where horizontal trends are more
the names of 20 people. 15 elements were named to honor significant than vertical group trends, such as the f-block,
16 scientists. Four other elements have indirect connec- where the lanthanides and actinides form two substantial
tion to the names of non-scientists. Write a short para- horizontal series of elements. Moving left to right across
graph about each of three of the above mentioned people a period, ionization energy (the energy required to re-
of your choice. move an electron from an atom), electron affinity, and
22
FIG. 41. Lattice structure of two Fe phases regulated by two

different temperatures (730 and 920 ◦ C). The alpha iron has
no spaces for carbon atoms to reside, while the gamma iron
is open to free movement of small carbon atoms.
FIG. 39. Graphite, buckyball, and nanotube.
FIG. 42. Anomalous Length Changes in Plutonium

Plutonium is a unique element in exhibiting six different crys-
tallographic phases at ambient pressure (it has a seventh
phase under pressure). In addition, unlike most metals, plu-
tonium contracts on melting. Transformations to different
crystal structures occur readily and are accompanied by very
large volume changes. By comparison, aluminums behavior
FIG. 40. Low-pressure phase diagram of pure iron. is predictable and uneventful. It expands monotonically on
heating in the solid phase, and it also expands on melting.
The dashed lines show that thermal contraction on cooling
the liquid (L) phase of plutonium extrapolates to that of the
electronegativity (a chemical property that describes the β-phase; the thermal contraction on cooling the -phase ex-
tendency of an atom to attract a shared pair of elec- trapolates to that of the γ-phase.
trons towards itself) usually increase, atomic radius de-
creases. This occurs because each successive element has
an added proton and electron, which causes the electron
to be drawn closer to the nucleus. the periodic table, has no group numbers and comprises
Specific regions of the periodic table can be referred lanthanides and actinides.
to as blocks in recognition of the sequence in which the
electron shells of the elements are filled. Each block is
named according to the subshell in which the ”last” elec-
tron notionally resides. The s-block comprises the first
B. Electrons in Atom
two groups (alkali metals and alkaline earth metals) as
well as hydrogen and helium. The p-block comprises the
last six groups, which are groups 13 to 18, and contains, In the self consistent field approximation, in an isolated
among other elements, all of the metalloids. The d-block atom there exist individual electron states, which can be
comprises groups 3 to 12 and contains all of the transi- labeled by four quantum numbers: n - principal, ` - an-
tion metals. The f-block, often offset below the rest of gular momentum, m - magnetic, and σz - spin quantum
23
(b) Element 118 has recently been discovered, and is

expected to be a noble gas, i.e., is in group VIII. (No
real chemistry tests have been performed on the ele-
ment yet, as the nucleus decays very quickly.) Assuming
that Madelung’s rule continues to hold, what should the
atomic number be for the next noble gas after this one?
Problem B.2 Exceptions to Madelungs Rule Although
Madelungs rule for the filling of electronic shells holds
extremely well, there are a number of exceptions to the
rule. Here are a few of them:
FIG. 43. Dmitri Ivanovich Mendeleev (1834 - 1907) was Cu = [Ar] 4s1 3d10
a Russian chemist and inventor. He formulated the Periodic Pd = [Kr] 5s0 4d10
Law, created a farsighted version of the periodic table of ele-
Ag = [Kr] 5s1 4d10
ments, and used it to correct the properties of some already
discovered elements and also to predict the properties of eight Au = [Xe] 6s1 4f14 5d10
elements yet to be discovered. What should the electron configurations be if these ele-
ments followed Madelungs rule and the Aufbau principle?
numbers;
A lattice with a basis is also necessary when more than
n = 1, 2, 3, . . . one kind of atom or ion is present in a compound, even
` = 0, 1, 2, . . . , n − 1 when the atoms or ions are located only at the points of
m = −`, . . . , ` some Bravais lattice, because the crystal structure nev-
σz = −1/2, 1/2. (136) ertheless lacks the full translational symmetry of that
Bravais lattice.
The angular momentum shells with ` = 0, 1, 2, 3 are Ionic radius is the radius of an atom’s ion in ionic crys-
sometimes known as s,p,d,f, . . . respectively in atomic tals structure. The radii of the cation Na+ and the anion
language. These shells can accommodate 2, 6, 10, 14, . . . F− as given in Tables 48 would lead to 1.02Å + 1.33Å =
electrons respectively including both spin states. 2.35Å for the interatomic separation in the crystal NaF,
As we pass from one element to another of next higher as compared with the observed 2.32Å.
atomic number, one electron is added each time to the The covalent radius is a measure of the size of an atom
atom. The maximum number of electrons in any shell that forms part of one covalent bond. The interatomic
is 2n2 , where n is the principal quantum number. The distance between C atoms in diamond is 1.54Å; one-half
maximum number of electrons in a subshell (s, p, d or of this is 0.77Å. In silicon, which has the same crystal
f) is equal to 2(2` + 1). Thus these subshells can have a structure, one-half the interatomic distance is 1.17Å. In
maximum of 2, 6, 10 and 14 electrons respectively. In the SiC each atom is surrounded by four atoms of the op-
ground state the electronic configuration can be built up posite kind. If we add the C and Si radii just given,
by placing electrons in the lowest available orbitals until we predict 1.95Å for the length of the C-Si bond, in fair
the total number of electrons added is equal to the atomic agreement with the 1.89Å observed for the bond length.
number. This is called the Aufbau principle (the German This is the kind of agreement (a few percent) that we
word Aufbau means building up). Thus orbitals are filled shall find in using tables of atomic radii.
in the order of increasing energy, using two general rules
to help predict electronic configurations:- Metallic radius is defined as one-half of the distance
1. Electrons are assigned to orbitals in order of increasing between the two adjacent metal ions in the metallic lat-
value of (n + `). tice. This radius depends on the nature of the atom as
2. For subshells with the same value of (n + `), electrons well as its environmentspecifically, on the coordination
are assigned first to the sub shell with lower n. number (CN).
The concepts of atomic radii is fruitful in predicting
Problem B.1 Madelung’s Rule. interatomic spacing. The existence and probable lat-
(a) Use Madelung’s rule to deduce the atomic shell tice constants of phases that have not yet been synthe-
filling configuration of the element tungsten (symbol W) sized can be predicted from the additive properties of the
which has atomic number 74. atomic radii.
VIII. CHEMICAL BOND. I cally transferred from one atom to the other, leaving two
oppositely charged ions, which then attract each other,
A. Ionic Solids
The general idea of an ionic bond is that sometimes

it is energetically favorable for an electron to be physi-
24
FIG. 44. The periodic table of the elements. The organization of the table is such that each column has similar chemical
properties. Each element is listed with its atomic number. The structure of the periodic table reflects Aufbau Prinnciple rule
which dictates the order in which shells are filled.
FIG. 45. Ordering of filling orbitals in atoms (Madelung’s

rule).
FIG. 46. Ionization energies.

forming chemical compound.
The total energy change from transferring an electron
from atom A to atom B is In addition, there is also
∆EA+B→A+ +B − = (Ionization Energy)A Cohesive Energy = Energy gained from A+ + B − → AB.

−(Electron Affinity)B . (137) (138)
25
FIG. 47. Electron affinity values. Electron affinity is defined

as the change in energy of a neutral atom (in the gaseous
phase) when an electron is added to the atom to form a neg-
ative ion.
So finally
∆EA+B→AB = (Ionization Energy)A (139)
−(Electron Affinity)B − Cohesive Energy of AB.
Problem A.1 From the data presented on Fig. 51 cal-
culate cohesive energy of NaCl in kJ/mol.
One can write a much simplified equation for a lattice
energy for a solid
X Qi Qj
ELattice = − , (140)
i<j
4π0 |Ri − Rj |
where Qi is the charge on the ith ion, and Ri is its po-

sition. This sum is sometimes known as the Madelung
Energy. For the case of the charges of the ions ±e Eq.
(140) can be rewritten as
e2 X (±) e2
ELattice = − ≡− M, (141)
4π0 R i<j pij 4π0 R
where R is the nearest-neighbor separation in the crystal,

and we put |Ri − Rj | = Rpij ; M is called Madelung
constant.
For the simple cubic lattice Madelung constant is given
by the series
∞
X (−1)j+k+`
M= (142)
j,k,`=−∞
(j 2 + k 2 + `2 )1/2
( the term j = k = ` = 0 is to be left out).

Problem A.2 (a) Calculate Madelung constant for a 1d
lattice
∞
X (−1)j
M =2 . (143)
j=1
j
(b) Calculate Lattice energy for a hypothetical one-

dimensional NaCl crystal presented on Fig. 52.
FIG. 48. Metallic, Covalent and Ionic Radii. Covalent radii
are in parentheses.
26
attractions because you have eight chlorides around the

sodium rather than six is more than countered by the new
repulsions between the chloride ions themselves. When
sodium chloride is 6:6-co-ordinated, there are no such
repulsions - and so that is the best way for it to organise
itself.
Which structure a simple 1:1 compound like NaCl or
CsCl crystallises in depends on the radius ratio of the
positive and the negative ions. The empirical rule is that
if the radius of the positive ion is bigger than 73% of
that of the negative ion, then 8:8-co-ordination is possi-
ble. Less than that (down to 41%) then you get 6:6-co-
ordination.
Problem B.2 (a) Do the lattices of NaCl and CsCl cor-

respond to the empirical rule presented above? (b) Check
FIG. 49. Visual representation of the Tables 48. The neutral up a few other compounds from the Tables 56.
atoms are colored gray, cations red, and anions blue.
Problem B.3 Take Madelung constant for the fcc lattice
as M = −1.75 and calculate the lattice energy for NaCl.
In three dimensions the series presents much greater How well it agrees with the data on Fig. 51.
difficulty. The following expansion is often found in the
literature:
√ √ C. Zinc Blende and Fluorite Structures
M = −6 + 12/ 2 − 8/ 3 + 6/2 − · · · = −1.747 . . . .
(144)
ZnS exists in two main crystalline forms, and this du-
alism is often a salient example of polymorphism. In each
B. Why are the Caesium Chloride and Sodium
form, the coordination geometry at Zn and S is tetrahe-
Chloride Structures Different? dral. The more stable cubic form is known also as zinc
blende or sphalerite. The hexagonal form is known as
the mineral wurtzite, although it also can be produced
Caesium chloride or cesium chloride is the inorganic
synthetically. As shown in Fig. 57, the zincblende crys-
compound with the formula CsCl. This colorless solid
tal is a an fcc lattice with a basis given by a Zn atom at
is an important source of caesium ions in a variety of
[0, 0, 0] and an S atom at [ 41 , 14 , 14 ].
niche applications. Its crystal structure forms a major
Fluorite (also called fluorspar) is the mineral form of
structural type where each caesium ion is coordinated by
calcium fluoride, CaF2 . The Ca atoms form a fcc lattice
8 chlorine ions. The structure is shown in Fig. 53.
and the F’s fill in the holes as shown in Fig. 59.
Sodium chloride, also known as salt, is an ionic com-
Fluorite is a colorful mineral, both in visible and ultra-
pound with the chemical formula NaCl, representing a
violet light, and the stone has ornamental and lapidary
1:1 ratio of sodium and chloride ions. The sodium chlo-
uses. Industrially, fluorite is used as a flux for smelt-
ride, NaCl, structure is shown in Figs. 54 and 55.
ing, and in the production of certain glasses and enam-
Problem B.1 Check up whether the lattice constants els. The purest grades of fluorite are a source of fluoride
from the Tables 56 agree with the ion radii from the Ta- for hydrofluoric acid manufacture, which is the interme-
bles 48. diate source of most fluorine-containing fine chemicals.
Optically clear transparent fluorite lenses have low dis-
When attractions are set up between two ions of op- persion, so lenses made from it exhibit less chromatic
posite charges, energy is released. That means that to aberration, making them valuable in microscopes and
gain maximum stability, you need the maximum number telescopes. Fluorite optics are also usable in the far-
of attractions. In CsCl lattice (see Fig. 53) a positive ultraviolet and mid-infrared ranges, where conventional
ion like caesium with eight chloride ions. So why doesn’t glasses are too absorbent for use.
sodium chloride do the same thing (see Fig. 54)?
Now imagine what would happen if you replaced the
caesium ion with the sodium ion, which has fewer layers D. Perovskite Structures
of electrons around it and hence, is smaller. You still
have to keep the chloride ions in contact with the sodium. The mineral perovskite (CaTiO3 ) is named after a Rus-
The effect of this would be that the whole arrangement sian mineralogist, Count Lev Aleksevich von Perovski,
would shrink, bringing the chloride ions into contact with and was discovered and named by Gustav Rose in 1839
each other - and that introduces repulsion. Any gain in from samples found in the Ural Mountains [where the
27
FIG. 50. Visual representation of the Tables 48.
FIG. 52. a hypothetical one-dimensional NaCl crystal.
FIG. 51. The energy per molecule unit of a crystal of sodium

FIG. 53. The cesium chloride crystal structure. The space
chloride is 7.9 − 5.1 + 3.6 = 6.4 eV lower than the energy of +
lattice is simple cubic, and the
basis has one Cs ion at 000
separated neutral atoms. The lattice energy with respect to
and one Cl− ion at 12 , 21 , 12 . The ionic diameters here are
separated ions is 7.9 eV per molecule unit. All values on the
reduced in relation to the cell in order to clarify the spatial
figure are experimental.
arrangement.
Author of these Notes was born]. The perovskite struc- than 10, while others, such as titanium dioxide (TiO2 ),
ture has the general stoichiometry ABX3 , where A and have values between 20 and 70.
B are cations and X is an anion. Barium atoms are located at the corners of the unit
Barium titanate (Ba2− Ti4− O2+ 3 ) with the structure of cell and oxygens at the face center positions. Both bar-
perovskite is a dielectric ceramic used in capacitors, with ium and oxygen have ionic radii of about 1.4Å and to-
dielectric constant values as high as 7, 000. Over a narrow gether they make up a face centered cubic array having
temperature range, values as high as 15, 000 are possible; a lattice parameter near 4Å. Octahedrally coordinated
most common ceramic and polymer materials are less titanium ions located at the center of the perovskite
28
FIG. 54. We may construct the sodium chloride crystal struc-

ture by arranging Na+ and Cl− ions alternately at the lattice
points of a simple cubic lattice. In the crystal each ion is
surrounded by six nearest neighbors of the opposite charge. FIG. 56. Some compounds with: the cesium chloride struc-
−
The space lattice is fcc, and the basis has one Cl ion at ture (above); the sodium chloride structure (below).
000 and one Na+ ion at 12 , 12 , 21 . The figure shows one con-
ventional cubic cell. The ionic diameters here are reduced in
relation to the cell in order to clarify the spatial arrangement.
A more true to life picture of NaCl structure is presented on
Fig. 55.
FIG. 57. Zincblende conventional unit cell. This is fcc with

a basis given by a Zn atom (light) at [0, 0, 0] and a S atom
FIG. 55. Left: Small units reproduced periodically to form an (dark) at [ 14 , 14 , 41 ].
fcc crystal. This particular figure depicts NaCl (table salt),
with the larger spheres being Cl− ions and the smaller spheres
being Na+ ions. Right: The macroscopic morphology of a tanium, and silicon, are also possible. The most known
crystal often will reflect the underlying microscopic structure. example is spinel: MgAl2 O4 , after which this class of
These are large crystals of salt (also known as halite). minerals is named. A and B can also be the same metal
with different valences, as is the case with magnetite,
Fe3 O4 (as Fe2+ Fe3+ 2−
2 O4 ), which is the most abundant
unit cell are the active ions in promoting ferroelectric- member of the spinel group.
ity. The low-lying d orbitals of Ti lead to displacive
phase transformations and large electronic polarizability.
Cubic Ba2− Ti4− O2+3 is presented on Fig. 61. It’s crys- IX. CHEMICAL BOND. II
tal structure was reliably determined in 1945 from X-ray
diffraction data. A. Covalent Solids
A covalent bond is a bond where electrons are shared

E. Spinel Structures
roughly equally between two atoms. A caricature model
that can be used to describe the covalent bond is Particle
The spinels are any of a class of minerals of gen- in a Box Picture presented on Fig. 63.
eral formulation AB2 X4 which crystallise in the cubic
(isometric) crystal system, with the X anions (typically
chalcogens, like oxygen and sulfur) arranged in a cubic 1. Molecular Orbitals and Valence Bonds
close-packed lattice and the cations A and B occupying
some or all of the octahedral and tetrahedral sites in the Much more true to life is Molecular Orbital or Tight
lattice. Although the charges of A and B in the pro- Binding Theory. For simplicity, let us consider a single
totypical spinel structure are +2 and +3, respectively electron and two identical positive nuclei. We write the
(A2+ B3+ 2−
2 X4 ), other combinations incorporating diva- Hamiltonian as
lent, trivalent, or tetravalent cations, including magne-
sium, zinc, iron, manganese, aluminium, chromium, ti- H = K + VA + VB (145)
29
FIG. 60. Left: Perovskite. Right: The Samarian Spinel is

a 500-carat (100 g) spinel gemstone that is the largest of its
FIG. 58. Left: Sphalerite mineral, which is the chief ore of
kind in the world. It is part of the Iranian Crown Jewels.
zinc. Actually, the mineral has composition (Zn, Fe)S. Right:
Deep green isolated fluorite crystal resembling a truncated
octahedron.
FIG. 59. The fluorite structure. This is fcc with a basis given
by Ca at [0, 0, 0] and F at [ 14 , 14 , 14 ] and [ 41 , 41 , 34 ].
p 2 FIG. 61. The crystal structure of barium titanate. The pro-

with K = 2m being the kinetic energy of the electron totype crystal is calcium titanate (perovskite). The structure
and Vi is the Coulomb interaction energy between the is cubic, with Ba2+ ions at the cube corners, O2− ions at
electron and the nucleus at position Ri . Let us write a the face centers, and a Ti4+ ion at the body center. Be-
trial wavefunction as low the Curie temperature the structure is slightly deformed,
with Ba2+ and Ti4+ ions displaced relative to the O2− ions,
|ψ >= αψA + βψB , (146)
thereby developing a dipole moment.
where ψA,B are atomic orbitals (or tight binding) or-
bitals. The form of Eq. (146) is frequently known as
a linear combination of atomic orbitals or LCAO. The t is called hopping. Diagonalizing the two-by-two matrix
orbitals which we use here can be taken as the ground- we obtain eigenenergies
state solution of the Schroedinger equation when there is
E± = 0 + Vcross ± t (151)
only one nucleus present:
(K + VA )ψA = 0 ψA where the lower energy orbital is the bonding orbital
whereas the higher energy orbital is the antibonding. The
(K + VB )ψB = 0 ψB . (147) corresponding wavefunctions are
We will now make a rough approximation that ψA > and 1
ψB are orthogonal so we can choose a normalization such ψb = √ (ψA ± ψB )
2
that < ψi ψj >= δij .
1
An effective Schroedinger equation takes the form ψab = √ (ψA ∓ ψB ), (152)
2
α α
Ĥ =E , (148) where b and ab stand for bonding and antibonding. The
β β
signs ± and ∓ depend on the sign of t, where the lower
where energy one is always called the bonding orbital and the
H11 = H22 = 0 + Vcross higher energy one is called antibonding.
H12 = −t, H21 = −t∗ , (149) The two nuclei interact, and to a first approximation,
their Coulomb repulsion cancels the attractive energy be-
and tween the nucleus and the electron on the opposite orbital
Vcross =< ψA |VB |ψA >=< ψB |VA |ψB > Vcross . Thus, including this energy we will obtain
t = − < ψA |VB |ψB >= − < ψA |VA |ψB >; (150) Ẽ± ≈ 0 ± t. (153)
30
FIG. 62. Atomistic positions in cubic perovskites.

FIG. 65. Hydrogen molecule orbitals.
FIG. 66. He2 hypothetic molecule electron configuration. The

four electrons occupy one bonding orbital at lower energy,
FIG. 63. Particle in a box picture of covalent bonding. Two
and one antibonding orbital at higher energy than the atomic
separated hydrogen atoms are like two different boxes each
orbitals.
with one electron in the lowest eigenstate. When the two
boxes are pushed together, one obtains a larger box, thereby
lowering the energy of the lowest eigenstate which is known the molecule we calculated above is
as the bonding orbital. The two electrons can take opposite
spin states and can thereby both fit in the bonding orbital. Ψ(1, 2) = ψb (1)ψb (2) (φ↑ (1)φ↓ (2) − φ↑ (2)φ↓ (1)) ,
The first excited state is known as the antibonding orbital.
(154)
where φ↑,↓ are spin functions. Alternative to molecu-
As the nuclei get closer together, the hopping term |t| lar orbitals theory is the valence bond theory. In the
increases, giving an energy level diagram as shown in left framework of this theory the two electron wavefunction
side of Fig. 64. This picture is obviously unrealistic, as it is approximated as
suggests that two atoms should bind together at zero dis-
Ψ(1, 2) = (ψA (1)ψB (2) + ψA (2)ψB (1))
tance between the nuclei. The problem here is that our
assumptions and approximations begin to break down as (φ↑ (1)φ↓ (2) − φ↑ (2)φ↓ (1)) . (155)
the nuclei get closer together (for example, our orbitals
are no longer orthogonal, Vcross does not exactly cancel
the Coulomb energy between nuclei, etc.). A more realis- 2. Hybridization
tic energy level diagram for the bonding and antibonding
states is given in the right side of Fig. 64. Chemist Linus Pauling first developed the hybridisa-
tion theory in 1931 to explain the structure of simple
molecules such as methane (CH4 ) using atomic orbitals.
Pauling pointed out that a carbon atom forms four bonds
by using one s and three p orbitals, so that ”it might be
inferred” that a carbon atom would form three bonds at
right angles (using p orbitals) and a fourth weaker bond
using the s orbital in some arbitrary direction. In real-
ity, methane has four bonds of equivalent strength sep-
arated by the tetrahedral bond angle of 109.5◦ . Pauling
explained this by supposing that in the presence of four
hydrogen atoms, the s and p orbitals form four equiva-
FIG. 64. Left: model tight binding energy levels as a function lent combinations or hybrid orbitals, each denoted by sp3
of distance between the nuclei of the atoms. Right: more
to indicate its composition, which are directed along the
realistic energy levels.
four C-H bonds. This concept was developed for such
simple chemical systems, but the approach was later ap-
The two electron wave function of the ground state of plied more widely, and today it is considered an effective
31
FIG. 69. sp3 hybridization. The wavefunctions are of the

form zf (r), xf (r) and yf (r) and are called the pz , px , py
orbitals, respectively.
FIG. 67. Linus Carl Pauling (1901-1994) was an American
chemist and biochemist, one of the founders of the fields of
quantum chemistry and molecular biology. His contributions In graphite atoms in the plane are bonded covalently,
to the theory of the chemical bond include the concept of or- with only three of the four potential bonding sites satis-
bital hybridisation and the first accurate scale of electroneg- fied. The fourth electron is free to migrate in the plane,
ativities of the elements. For his scientific work, Pauling was making graphite electrically conductive. However, it does
awarded the Nobel Prize in Chemistry in 1954. not conduct in a direction at right angles to the plane.
Bonding between layers is via weak van der Waals bonds,
which allows layers of graphite to be easily separated, or
heuristic for rationalising the structures of organic com- to slide past each other.
pounds.
FIG. 68. Methane molecule.
The valence electrons in the ground state of the free

atoms have the configuration ns2 np2 , with n = 2, 3, 4
for diamond, Si, and Ge, respectively. In the crystal
the ground state is formed from the configuration nsnp3 .
Using the language of chemistry, we say that the valence
electrons form directed sp3 tetrahedral bonding orbitals FIG. 70. sp2 hybridization.
of the form
s + px + py + pz ; s + px − py − pz
s − px − py + pz ; s − px + py − pz . (156) 3. Fractionally Ionic Covalent Bonds
Each atom in the diamond structure is at the center of

Problem A.1 To understand molecular orbitals in a
a tetrahedron, with the nearest-neighbor atoms at the
heteronuclear molecule consider the model with two dif-
vertices. The four orbitals just enumerated have lobes
ferent atoms A and B having atomic orbitals with two
pointing in the tetrahedral directions. Tetrahedral (sp3 )
different energies A and B . Assume that the Hamilto-
configuration almost ubiquitous among semiconductors:
nian has the form
Si, Ge, III-V (GaAs, AlAs, InP), II-VI (CdS, CdTe).
Covalently bonded materials tend to be strong and

A −t
tend to be electrical semiconductors or insulators H= . (157)
−t∗ B
(roughly since electrons are tied up in the local bonds.
The directionality of the orbitals makes these materials Find eigenenergies and eigenfunctions. Show that for
retain their shape well (non-ductile) so they are brittle. |A − B | |t we effectively get ionic bond.
They do not dissolve in polar solvents such as water in
the same way that ionic materials do. Percent ionic character is the amount of or measure of a
32
compound’s ionic and covalent character. It is calculated

by dipole moment as
Actual dipole moment/theoretical dipole moment= per-
cent ionic character.
Electronegativity, symbol χ, is a chemical property
that describes the tendency of an atom to attract a
shared pair of electrons (or electron density) towards it-
self.[1] An atom’s electronegativity is affected by both
its atomic number and the distance at which its valence
electrons reside from the charged nucleus. The higher
the associated electronegativity number, the more an el-
ement or compound attracts electrons towards it. The
difference in electronegativity between atoms A and B is
given by:
p
|χA − χB | = (eV)−1/2 Ed (AB) − [Ed (AA) + Ed (BB)]/2,
(158)
FIG. 73. HF orbitals. The σ−π notation differentiates molec-
where the dissociation energies, Ed , of the AB, AA ular orbitals with respect to their symmetry.
and BB bonds are expressed in electronvolts, the factor
(eV)−1/2 being included to ensure a dimensionless result.
this case the electrons become delocalized throughout the
crystal. We should think of the delocalized free electrons
as providing the glue that holds together the positive ions
that they have left behind. Since the electrons are com-
pletely delocalized, the bonds in metals tend not to be
directional. Metals are thus often ductile and malleable.
Since the electrons are free, metals are good conductors
of electricity as well as of heat.
C. Crystals of Inert Gases
The inert gases form the simplest crystals. The crys-

tals are transparent insulators, weakly bound, with low
FIG. 71. Fractional ionic character of bonds in binary crys- melting temperatures. The crystal structures are all cu-
tals. bic close-packed (fcc), except 3 He and 4 He. Cohesive
energy is the energy required to form separated neutral
atoms in their ground electronic state from the solid. The
properties of inert gas crystals are presented in table on
Fig. 74.
FIG. 72. sp hybridization.
In the case of HF, the energies of the hydrogen and

the fluorine orbitals are different. Fluorine is more elec-
tronegative than hydrogen, 2p has lower energy than H
1s. As a result, the bonding orbital is mostly composed
of the orbital on the F atom - meaning that the bonding
electrons are mostly transferred from H to F - forming a
more ionic bond. FIG. 74. Properties of inert gas crystals.
When two atoms (or two molecules) are very far apart
B. Metallic Solids from each other, there remains an attraction between
them due to what is known as van der Waals forces, some-
Metallic bonds are similar to covalent bonds in the times known as fluctuating dipole forces, or molecular
sense that electrons are shared between atoms, but in bonding. In short, both atoms have a dipole moment,
33
which may be zero on average, but can fluctuate mo-

mentarily due to quantum mechanics. If the first atom
obtains a momentary dipole moment, the second atom
can polarize also obtaining a dipole moment to lower its
energy. As a result, the two atoms (momentarily dipoles)
will attract each other.
d1 · d2 − 3(d1 · n)(d2 · n)
V (R) = , (159)
R3
where n is the unit vector in the direction from one atom
to the other. In the second order of perturbation theory
we obtain
1
V (R) ∼ . (160)
R6
FIG. 76.
This type of bonding between atoms is very typical of
inert atoms (such as noble gases: He, Ne, Kr, Ar, Xe)
whose electrons do not participate in covalent bonds or D. Hydrogen Bonding
ionic bonds. This bonding is weak compared to covalent
or ionic bonds, but it is also long ranged in comparison The hydrogen atom is extremely special due to its very
since the electrons do not need to hop between atoms. small size. As a result, the bonds formed with hydrogen
atoms are qualitatively different from other bonds. When
the hydrogen atom forms a covalent or ionic bond with
a larger atom, being small, the hydrogen nucleus (a pro-
ton) simply sits on the surface of its partner. This then
makes the molecule (hydrogen and its partner) into a
dipole. These dipoles can then attract charges, or other
dipoles, as usual. What is special about hydrogen is that
when it forms a bond, and its electron is attracted away
from the proton onto (or partially onto) its partner, the
unbonded side of the proton left behind is a naked posi-
tive charge unscreened by any electrons in core orbitals.
As a result, this positive charge is particularly effective
in being attracted to other clouds of electrons.
FIG. 75. Typical shape of an interatomic pair potential.
The Lennard-Jones interatomic potential is

σ 12 σ 6
U (R) = 4ε − , (161)
R R
where ε is the depth of the potential well and σ is the

distance at which the potential crosses zero. The term
proportional to 1/r6 in the potential can be motivated
from a classical or quantum mechanical description of FIG. 77. One of crystal structures of ice. A hydrogen on
the interaction between induced electric dipoles. This one water molecule is attracted to an oxygen on another
potential seems to be quite accurate for noble gases, and molecule forming a weak hydrogen bond. These bonds are
is widely used for systems where dipole interactions are strong enough to form ice below the 273.15 K. H is positively
significant, including in chemistry force fields to describe charged but also very small: another real bond cannot be
intermermolecular interactions. established without overlap of electron clouds (the distances
between water molecules are far too big in this drawing).
Problem C.1 Show that the minimum of the Lennard-
Jones potential U (R) occurs at R/σ = 21/6 ≈ 1.12.
Problem D.1 Water ice is ordinarily found in the
34
hexagonal form, but can also exist in the cubic, rhom- To construct the reciprocal lattice, let us first write the
bohedral, and many other forms. Which form of ice is points of the direct lattice in the form
presented on Fig. 77?
R = n1 a1 + n2 a2 + n3 a3 (163)
with n1 , n2 , and n3 integers, and with a1 , a2 , and a3
being primitive lattice vectors of the direct lattice.
The primitive lattice vectors of the reciprocal lattice
(which we will call b1 , b2 , and b3 ) are defined to have
the following property:
ai · bj = 2πδij . (164)
We can certainly construct vectors bi to have the de-
sired property of Eq. (164), as follows:
2πa2 × a3
b1 =
a1 · [a2 × a3 ]
2πa3 × a1
b2 =
a1 · [a2 × a3 ]
2πa1 × a2
b3 = . (165)
a1 · [a2 × a3 ]
We can write an arbitrary point in reciprocal lattice as
G = m1 b1 + m2 b2 + m3 b3 . (166)
Problem A.1 Given that the primitive basis vectors of

a lattice are a = (a/2)(i + j), b = (a/2)(j + k) and
c = (a/2)(k+i), where i, j, and k are the usual three unit
vectors along cartesian coordinates, what is the Bravais
lattice?
B. Lattice Planes and Miller Indices
FIG. 78. Bonding. A lattice plane (or crystal plane) is a plane contain-
ing at least three non-collinear (and therefore an infinite
number of) points of a lattice. A family of lattice planes is
an infinite set of equally separated parallel lattice planes
which taken together contain all points of the lattice.
X. WAVE SCATTERING BY CRYSTALS. I
The families of lattice planes are in one-to-one corre-
spondence with the possible directions of reciprocal lat-
A. The Reciprocal Lattice tice vectors, to which they are normal. Further, the spac-
ing between these lattice planes is
The reciprocal lattice is the Fourier transform of a Bra-
vais lattice, known in this context as the direct lattice. d = 2π/|Gmin |, (167)
While the direct lattice exists in real-space and is what
where Gmin is the minimum length reciprocal lattice vec-
one would commonly understand as a physical lattice, the
tor in this normal direction.
reciprocal lattice exists in reciprocal space (also known
Any reciprocal lattice vector can be written as
as momentum space or k-space).
Given a direct lattice of points R, a point G is a point G = hb1 + kb2 + `b3 , (168)
in the reciprocal lattice if and only if
where bi are the basis of the reciprocal lattice vectors.
eiG·R = 1 (162) Hence these three integers h, k and ` determine a family
of lattice planes; in this context they are called Miller in-
for all points R of the direct lattice. The reciprocal lattice dices and written (hk`). By convention, negative integers
of a reciprocal lattice, then, is the original direct lattice are written with a bar, as in 3̄ for −3. The integers are
again, since the two lattices are Fourier transforms of usually written in lowest terms, i.e. their greatest com-
each other. mon divisor should be 1. The notation {hk`} denotes
35
evidence was supported by the discovery in 1912 of X-ray

diffraction by crystals, when Laue developed the theory
of X-ray diffraction by a periodic array, and his cowork-
ers reported the first experimental observation of x-ray
diffraction by crystals. The importance of x-rays for this
task is that they are waves and have a wavelength compa-
rable with the length of a building block of the structure.
Such analysis can also be done with neutron diffraction
and with electron diffraction, but x-rays are usually the
tool of choice.
FIG. 79. Some lattice planes in cubic crystals. Exposing a solid to a wave in order to probe its prop-
erties is an extremely useful thing to do. The most com-
the set of all planes that are equivalent to (hk`) by the monly used probe is X-rays. Another common, more
symmetry of the lattice. modern, probe is neutrons. It can hardly be overstated
If one chooses ai to be the real (direct) space primi- how important this type of experiment is to science.
tive lattice vectors, then bi will be the primitive lattice
vectors for the reciprocal lattice. In this case, any set of
integer Miller indices (hk`) represents a reciprocal lattice
vector. On the other hand, if one chooses ai to describe
the edges of some non-primitive (conventional) cell, the
corresponding bi will not be reciprocal lattice vectors.
As a result not all integer sets of Miller indices corre-
spond to a lattice plane. (Each (hk`) represents a family
of planes but not necessarily a family of lattice planes,
i.e., there are some planes that do not intersect lattice
points).
For any cubic lattice (simple cubic, fcc, or bcc) it is
conventional to choose ai to be ax̂, aŷ, and aẑ with a the
cube edge length. The lattice vectors, which are edge vec-
tors of the conventional (cube) unit cell, are thus orthog-
onal and of equal length, as are those of the reciprocal lat-
tice. In this common case, the Miller indices (hk`) simply
denote normals/directions in Cartesian coordinates. Cor-
respondingly, i are the vectors 2πx/a, 2πy/a, and 2πz/a.
The spacing d between adjacent (hk`) lattice planes is
FIG. 80. Wavelength versus particle energy, for photons, neu-
a trons, and electrons.
d(hk`) = √ . (169)
h + k 2 + `2
2
For the simple cubic lattice any set of (hk`) represents a

lattice plane, for the bcc, h + k + ` must be even, and for
the fcc, h, k, ` must be either all odd or all even.
The reciprocal lattice plays a fundamental role in most
analytic studies of periodic structures, particularly in the
electron theory. In neutron and X-ray diffraction, due to
the Laue conditions, the momentum difference between
incoming and diffracted X-rays of a crystal is a reciprocal
lattice vector. The diffraction pattern of a crystal can be
used to determine the reciprocal vectors of the lattice.
Using this process, one can infer the atomic arrangement
of a crystal.
FIG. 81. A generic scattering experiment
C. Fermi’s Golden Rule

According to Fermi’s golden rule the transition rate
Γ(k0 , k) per unit time for the particle scattering from k
The original experimental evidence for the periodicity to k0 is given by
of the crystal structure rests on the discovery by min-
eralogists that the index numbers that define the orien- 2π
tations of the faces of a crystal are exact integers. This Γ(k0 , k) = | < k0 |V |k > |2 δ (Ek − Ek0 ) . (170)
~
36
If V (r) is a periodic function

V (r + R) = V (r), (171)
then
Z X Z
dreik·r V (r) = eik·R dxeik·x V (x), (172)
R Ω
where Ω is the volume of the unit cell. Thus the intensity

of scattering

X 0

2
I(k0 , k) ∝ ei(k−k )·R |S(k − k0 )| , (173)

R
where
Z
S(k) = dxeik·x V (x), (174)
Ω
FIG. 82. The points on the right-hand side are reciprocal-
lattice points of the crystal. The vector k is drawn in the
Hence the lattice and the basis in the description of scat- direction of the incident x-ray beam, and the origin is chosen
tering should be treated separately. The lattice, as we such that k terminates at any reciprocal lattice point. We
will see, determines the positions (angles) of diffracted draw a sphere of radius k = 2π/λ about the origin of k. A
beams, and the basis determines the intensity of a given diffracted beam will be formed if this sphere intersects any
beam. other point in the reciprocal lattice. The sphere as drawn
intercepts a point connected with the end of k by a reciprocal
lattice vector G. The diffracted X-ray beam is in the direction
D. The Laue Condition k0 = k + G. The angle θ is the Bragg angle.
We have
X (2π)D X D
eik·R = δ (k − G), (175)
Ω
R G
where in the L.H.S. the sum is over lattice points R of the
direct lattice, and in the R.H.S. it is a sum over points G
of the reciprocal lattice; D is the number of dimensions
(1, 2 or 3) and the δ D is a D-dimensional delta function.
Returning to Eq. (173) we understand that the scat-
tering must vanish unless k0 − k is a reciprocal lattice
vector. This condition
k0 − k = G (176)
is known as the Laue equation (or Laue condition) and is FIG. 83. Max Theodor Felix von Laue (1879-1960) was
precisely the statement of the conservation of crystal mo- a German physicist who won the Nobel Prize in Physics in
mentum. When the waves leave the crystal, they should 1914 for his discovery of the diffraction of X-rays by crystals.
He contributed to optics, crystallography, quantum theory,
have
superconductivity, and the theory of relativity. During Na-
|k0 | = |k|, (177) tional Socialist regime, Laue and his close friend Otto Hahn
secretly helped Jewish scientific colleagues to emigrate from
which is just the conservation of energy, which is enforced Germany.
by the delta function in Fermi’s golden rule.
Equation (176) may be expressed in another way to
give what are called the Laue equations. These are valu- E. The Bragg Condition
able because of their geometrical representation. Taking
G = hb1 + kb2 + `b3 (178) It turns out that this Laue condition is nothing more
and calculating the scalar products of both sides of Eq. than the scattering condition associated with a diffrac-
(176) successively with a1 , a2 , a3 we get tion grating. This description of the scattering from crys-
tals is known as the Bragg formulation of (X-ray) diffrac-
a1 · ∆k = h, a2 · ∆k = k, a3 · ∆k = `. (179) tion.
Ewald construction, exhibited in Fig. 82, helps us visu- Consider the configuration shown in Fig. 84. An in-
alize the nature of the accident that must occur in order coming wave is reflected off of two adjacent layers of
to satisfy both Eq. (176) and Eq. (177) in three dimen- atoms separated by a distance d. The wave is deflected by
sions. 2θ in this diagram. The additional distance traveled by
37
spacing between lattice planes is d = 2π/|G|. Just from

geometry we have
k̂ · Ĝ = sin θ = −k̂0 · Ĝ, (183)
where the hats over vectors indicate unit vectors. Sup-
pose the Laue condition is satisfied. That is, k0 − k = G
with |k0 | = |k| = 2π/λ with λ the wavelength. We can
rewrite the Laue equation as
2π 0
(k̂ − k̂) = G. (184)
λ
FIG. 84. Bragg scattering off of a plane of atoms in a crystal.
The excess distance traveled by the wave striking the lower Now let us dot this equation with Ĝ to give
plane is 2d sin θ. 2π 0
Ĝ · (k̂ − k̂) = Ĝ · G
λ
2π
(sin θ0 − sin θ) = |G|
λ
2π
(2 sin θ) = λ. (185)
|G|
Taking into account Eq. (167 we recover Brag condition
(181).
The diffraction condition being satisfied, one can write
the scattering intensity as
I(h,k,`) ∝ |S(h,k,`) |2 . (186)
If we think in terms of Laue condition, h, k, ` are the
components of the scattering vector G. If we think in
FIG. 85. William Lawrence Bragg (1890 1971) was an terms of Bragg condition, h, k, ` are the Miller indices of
Australian-born British physicist and X-ray crystallographer, the appropriate family of lattice planes.
discoverer of Bragg’s law of X-ray diffraction, which is basic
for the determination of crystal structure. He was joint winner
(with his father, William Henry Bragg) of the Nobel Prize in XI. WAVE SCATTERING BY CRYSTALS. II
Physics in 1915
A. Scattering Amplitudes
the component of the wave that reflects off of the further It is usually a very good approximation to assume that
layer of atoms is the scattering potential is the sum over the scattering
extra distance = 2d sin θ. (180) potentials of the individual atoms in the system, so that
we can write
In order to have constructive interference, this extra dis- X
V (x) = Vj (x − xj ), (187)
tance must be equal to an integer number n of wave-
atoms j
lengths. Thus we derive the Bragg condition for con-
structive interference, or what is known as Bragg’s law where Vj is the scattering potential from atom j and xj
is the position of the atom in unit cell. The form of the
nλ = 2d sin θ. (181) function Vj will depend on what type of probe wave we
taking into account Eq. (169) we obtain a very useful are using and what type of atom j is.
expression for the spacing between planes Since neutrons are uncharged, they scatter almost ex-
clusively from nuclei (rather than from electrons) via the
λ a nuclear forces. As a result, the scattering potential is
d(hk`) = =√ . (182) extremely short-ranged, and can be approximated as a
2 sin θ h2 + k 2 + `2
delta function. We thus have
X
V (x) = fj δ(x − xj ), (188)
F. Equivalence of Laue and Bragg Conditions atoms j
where xj is the position of the jth atom in the unit cell.

Consider again Fig. 84 and the reciprocal lattice vector Here, fj is known as the form factor or atomic form fac-
G which corresponds to the family of lattice planes. The tor, and represents the strength of scattering from that
38
particular nucleus. For the case of neutrons this quan- The point here is that the (100) planes have some par-
tity is proportional to11 the so-called nuclear scattering- ticular spacing but there are five other families of planes
length bj . (To be precise, fj = 2π~2 bj /m with m the with the same spacing: (010), (001), (1̄00), (01̄0), (001̄).
mass of the neutron.) Thus for neutrons we frequently In the powder diffraction method, the crystal orienta-
write tions are random, and here there would be six possible
X equivalent orientations of a crystal which will present the
V (x) ∼ bj δ(x − xj ), (189) right angle for scattering from one of these planes, so
atoms j there will be scattering intensity which is six times as
and large as we would otherwise calculatethis is known as
X the multiplicity factor. For the case of the 111 family,
S(G) ∼ bj eiG·xj . (190) we would instead find eight possible equivalent planes:
atoms j (111), (111̄), (11̄1), (1̄11), (11̄1̄), (1̄11̄), (1̄1̄1), (1̄1̄1̄). Thus,
we should replace Eq. (186) with
X-rays scatter from the electrons in a system. In the
simplest approximation for the form-factor of X-ray scat- I{h,k,`} ∝ M{h,k,`} |S{h,k,`} |2 . (196)
tering we obtain result similar to Eq. (190)
X
S(G) ∼ Zj eiG·xj , (191)
atoms j
where Zj is the number of electrons in atom (ion) j. We’ll

use general notation for Eqs. (190) and (191)
X
S(G) = fj eiG·xj , (192)
atoms j
For a simple cubic lattice

S(h,k,l) = f. (193)
For a bcc lattice
h 1 1 1
i
S(h,k,l) = f 1 + e2πi(h,k,`)[ 2 , 2 , 2 ] = f 1 + (−1)h+k+` .

(194)
From Eq. (194) we see that the structure factor, and FIG. 86. Selection rules for cubic, bcc, and fcc lattices.
therefore the scattering intensity, vanishes for h + k + `
being any odd integer. This phenomenon is known as a
Consider CsCl, which can be described as a simple cu-
systematic absence.
bic lattice with a basis given by : Cs, Position = [0, 0, 0],
The fcc crystal can be thought of as a sim-
Cl, Position = [ 21 , 12 , 12 ]. Thus the structure factor is given
ple cubic lattice with a basis given by the points
by
[0, 0, 0], [0, 12 , 12 ], [ 21 , 0, 12 ], [ 21 , 12 , 0]. As a result the struc-
ture factor of fcc lattice is given by 1 1 1
S(h,k,l) = fCs + fCl e2πi(h,k,l)[ 2 , 2 , 2 ]
S(h,k,l) = f 1 + (−1)h+k + (−1)h+l + (−1)k+l .

= fCs + fCl (−1)h+k+l . (197)
(195)
If we consider the fcc lattice to be a cubic lattice with
Problem A.1 Show that the expression in the R.H.S. of basis consisting of eight atoms having positions presented
Eq. (195) vanishes unless h, k and ` are either all odd or in Table 87. the structure factor for ZnS will be the sum
all even.
It is useful to write down a table (presented on Fig.
86) of possible lattice planes and the selection rules that
can occur for the smallest reciprocal lattice vectors. The
selection rules were given previously: simple cubic allows
scattering from any plane, bcc must have h+k+` be even,
and fcc must have h, k, ` either all odd or all even. On the FIG. 87. Basis for conventional unit cell of ZnS
table we have also added a column N which is the square
magnitude of the Miller indices. We have also added an of eight terms.
additional column labeled multiplicity. This quantity is
important for figuring out the amplitude of scattering. Problem A.2 Show that the structure factor for ZnS
39
can be factored to give

S(h,k,l) = 1 + (−1)h+k + (−1)h+l + (−1)k+l

h π
i
× fZn + fS ei 2 (h+k+l) . (198)
The first term in brackets in the R.H.S. of Eq. (198)

is precisely the same as the term we found for the fcc
crystal. Both Eq. (194) and Eq. (198) are the examples
of general theorem, that the structure factor of a crys-
tal system with a basis will always factorize into a piece
which comes from the underlying lattice structure times
a piece corresponding to the basis
lattice basis
S(hk`) = S(hk`) × S(hk`) . (199)
FIG. 89. Continuous rotation electron diffraction pattern of
a single crystal of silver. The axis of rotation is normal to the
B. Experimental Techniques paper.
1. The Rotating-Crystal Method

that particular wavelength which satisfies Bragg’s law at
the present orientation, and a diffracted beam emerges
Conceptually, perhaps the simplest method is to take a at the corresponding angle. The diffracted beam is then
large single crystal of a material, fire waves at it (X-rays, recorded as a spot on the film. But since the wavelength
say) from one direction, and measure the direction of the corresponding to a spot is not measured, one cannot de-
outgoing waves. However, given a single direction of the termine the actual values of the interplanar spacings-only
incoming wave, it is unlikely that you precisely achieve their ratios. Therefore one can determine the shape but
the Laue (Bragg) condition for any inverse lattice vector not the absolute size of the unit cell. A typical Laue
(set of lattice planes). In order to get diffraction, one photograph is shown in Fig. 90(b).
should rotate the crystal.
FIG. 90. The Laue method: (a) Experimental arrangement.

(b) Laue pattern for an Mg crystal, with the x-ray beam par-
allel to the 6-flold symmetry axis.
FIG. 88. A setup for a single crystal diffraction experiment.
3. Powder Diffraction
2. The Laue Method Powder diffraction, or the Debye - Scherrer method, is

the use of wave scattering on a sample which is not single
This method can be used for a rapid determination of crystalline, but is powdered. Because one does not need
the symmetry and orientation of a single crystal. The single crystals this method can be used on a much wider
experimental arrangement is shown in Fig. 90(a). A variety of samples.
white X-ray beam-i.e., one with a spectrum of continu- In this case, the incoming wave can scatter off of any
ous wavelength- is made to fall on the crystal, which has one of many small crystallites which may be oriented in
a fixed orientation relative to the incident beam. Flat any possible direction. In spirit this technique is similar
films are placed in front of and behind the specimen. to the rotating crystal method in that there is always
Since ,1. covers a continuous range, the crystal selects some angle at which a crystal can be oriented to diffract
40
the incoming wave. A schematic of the Debye - Scherrer

setup is shown in Fig. 91 and sample data is shown in
Fig. 92. Using Bragg’s law, given the wavelength of
the incoming wave, we can deduce the possible spacings
between lattice planes.
FIG. 91. Schematic of a Debye - Scherrer powder diffraction

experiment.
FIG. 92. Debye - Scherrer powder diffraction data exposed on FIG. 93.
photographic film. In modern experiments digital detectors
are used.
C. Examples
Comparison of X-ray diffraction on KCl and KBr (fcc

lattice) is presented on Fig. 93. In KCl the numbers of
electrons in K+ and Cl− ions are equal. The scattering
amplitudes f (K+ ) and f (Cl− ) are almost exactly equal,
so the crystal looks to X-rays as if it were a monoatomic
simple cubic lattice of lattice constant a/2. Only even
integers occur in the reflection indices when these are
based on a cubic lattice with lattice constant a. In KBr
the form factor of Br− is quite different to that of K+ FIG. 94. XRD pattern for aluminum.
and all the reflections of the fcc lattice are present.
Results of X-ray diffraction on Al are presented on Fig.
94. Analysis of the experimental data presented on Fig. Consider now the powder diffraction on PrO2 (the data
95 allows to determine both the crystal structue of alu- is shown on Fig. 100). Given the wavelength .123 nm,
minum and the lattice constant. (and tentatively assuming a cubic lattice of some sort)
Results of neuron diffraction on TiC are presented on we first would like to figure out the type of lattice and
Figs. 96. From analysis of the experimental data pre- the lattice constant. Averaging the numbers in the final
sented on Fig. 97 we realize that the Bravais lattice of the column in the table on Fig. 101 gives us lattice constant
material is fcc and find the lattice constant. The analysis a = .541 ± .002 nm.
of the diffraction peaks relative magnitudes, presented on The analysis thus far is equivalent to what one would
Fig. 98, allows us to chose between two possible crystal do for X-ray scattering. However, with neutrons, assum-
structures. ing the scattering length is independent of scattering an-
41
FIG. 95. Analysis of XRD pattern for aluminum. In the first FIG. 97. Analysis of neutron diffraction pattern for titan
two columns we just read the angles off of the given graph. carbide.
In the third column of the table we calculate the distance
between lattice planes for the given diffraction peak using
Bragg’s law. In the fourth column we have calculated the
squared ratio of the lattice spacing d for the given peak to
the lattice spacing for the first peak (labeled a) as a reference.
We then realize that these ratios are pretty close to whole
numbers divided by 3, so we try multiplying each of these
quantities by 3 in the next column. If we round these numbers
to integers (given in the next column), we produce precisely
the values of N = h2 + k2 + `2 expected for the fcc lattice, and
we thus conclude that we are looking at an fcc lattice. The
final column then calculates the lattice constant. Averaging
these numbers gives us a measurement of the lattice constant
a = .54 nm.
FIG. 98. Neutron diffraction pattern for titan carbide.
The scattering intensity of the peaks are then given in

terms of this structure factor and the peak multiplicities.
We thus can write for all of our measured peaks
π π
2
I{h,k,`} = CM{h,k,`} bPr + bO ei 2 (h+k+l) + ei 2 (h+k+3l) ,

(201)
FIG. 96. Neutron diffraction pattern for titan carbide. where the constant C contains other constant factors (in-
cluding the factor of 42 from the fcc structure factor).
Problem C.1 Show that the first multiplier in Eq.
gle (which is typically a good assumption) we can go a
(200)is the same for all possible (hkl) included in {hkl}.
bit further by analyzing the intensity of the scattering
peaks. Thus we can compile another table showing the predicted
It turns out that PrO2 has fluorite structure. Let us relative intensities of the peaks: From the analytic ex-
see what further information we might extract from the pressions in the third column we can immediately predict
data in Fig. 100. The structure factor for this crystal is that we should have
4
S(h,k,l) = 1 + (−1)h+k + (−1)h+l + (−1)k+l (200)

Id = 3Ia , Ic = 2If , Ie = Ib , (202)
h π π
i 3
× bPr + bO ei 2 (h+k+l) + ei 2 (h+k+3l) . which agrees very well with the Table 102.
42
FIG. 101. Analysis of data shown in Fig. 100.
FIG. 99. Titan carbide crystal structure.
FIG. 102. Predicted versus measured intensities of diffraction

peaks for scattering of neutrons from PrO2 . Here the predic-
tion is based purely on the scattering structure factor and the
scattering multiplicity.
b) Write down expressions for the structure factors

S(hkl) for neutron diffraction assuming NaH has
(i) the sodium chloride (NaCl) structure
(ii) the zinc blende (ZnS) structure.
c) Hence, deduce which of the two structure models
is correct for NaH. (Nuclear scattering length of Na =
0.363×105 nm; nuclear scattering length of H = −0.374×
FIG. 100. Powder diffraction of neutrons from PrO2 . The 105 nm.)
wavelength of the neutron beam is λ = .123 nm. d) How does one produce monochromatic neutrons for
use in neutron diffraction experiments?
e) What are the main differences between neutrons and
Problem C.2 From the Table 102 find the ratio of Pr
X-rays?
and O scattering lengths.
Problem C.3 Neutron Scattering. X-ray diffraction D. Scattering in Liquids and Amorphous Solids
from sodium hydride (NaH) established that the Na atoms
are arranged on a face-centered cubic lattice. A material need not be crystalline to scatter waves.
a) Why is it difficult to locate the positions of the H However, for amorphous solids or liquids, instead of hav-
atoms using X-rays? ing delta-function peaks in the structure factor at recip-
The H atoms were thought to be displaced from the Na rocal lattice vectors, the structure factor (defined as the
atoms either by [ 41 , 14 , 14 ] or by [ 12 , 12 , 12 ], to form the ZnS Fourier transform of the density) will have smooth behav-
(zincblende) structure or NaCl (sodium chloride) struc- ior with incipient peaks corresponding to 2π/d, where d
ture, respectively. To distinguish these models a neutron is roughly the typical distance between atoms. As the
powder diffraction measurement was performed. The in- material gets close to its freezing point, the peaks in
tensity of the Bragg peak indexed as (111) was found to the structure factor will get more pronounced, becom-
be much larger than the intensity of the peak indexed as ing more like the structure of a solid where the peaks are
(200). delta functions.
43
E. Scattering and Nobel Prizes
In addition to the Nobel Prizes in physics awarded in

1914 to M. von Laue for the discovery of X-rays by crys-
tals, in 1915 to W. H. Bragg and W. L. Bragg for the
determination of crystal structures using X-rays, and in
1994 to B. N. Brockhouse and C. G. Shull for pioneering
contributions to the development of neutron scattering
techniques for studies of condensed matter and of the
neutron diffraction technique there have about a dozen
Nobel Prizes that have relied on, or further developed
these techniques (also electron diffraction technique).
In 1962 a chemistry Nobel Prize was awarded to Perutz
and Kendrew for using X-rays to determine the struc-
ture of the biological proteins hemoglobin and myoglobin.
The same year, Watson and Crick were awarded the prize
in biology for determining the structure of DNA - which
they did with the help of X-ray diffraction data taken FIG. 103. The reciprocal lattices (dots) and corresponding
by Rosalind Franklin. Two years later in 1964, Dorothy first Brillouin zones of (a) square lattice and (b) hexagonal
Hodgkin won the prize for determination of the struc- lattice.
ture of penicillin and other biological molecules. Further
Nobel Prizes were given in chemistry for using X-rays
to determine the structure of boranes (Lipscomb, 1976),
for the development of direct methods for X-ray crys-
tallographic structure determination (H. A. Hauptman
and J. Karle, 1985), photosynthetic proteins (Deisen-
hofer, Huber, Michel, 1988), and ribosomes (Ramakrish-
nan, Steitz, Yonath, 2009).
The recent (2016) Nobel Prize in Physics awarded to
Kosterlitz, Thouless, and Haldane recognized work with
a broad impact in neutron scattering.
FIG. 104. Felix Bloch (1905-1983) was a Swiss physicist,

XII. BLOCH (ELECTRONS IN A PERIODIC working mainly in the U.S. He was awarded the 1952 No-
POTENTIAL) THEORY bel Prize for Physics for ”their development of new ways and
methods for nuclear magnetic precision measurements.” Fe-
A. Brillouin Zones and Bloch Theorem lix Bloch made fundamental theoretical contributions to the
understanding of electron behavior in crystal lattices, ferro-
magnetism, and nuclear magnetic resonance.
The reciprocal space can be broken up into Brillouin
zones. The first Brillouin zone is a uniquely defined prim-
itive cell in reciprocal space – the Wigner-Seitz cell (in the Bloch Theorem: An electron in a periodic potential
reciprocal space) of the reciprocal lattice point G = 0. has eigenstates of the form
All k points which are closer to 0 than to any other re-
ciprocal lattice point define the first Brillouin zone. The ψαk (r) = eik·r uα
k (r) (204)
boundaries of this cell are given by planes related to where uα k is periodic in the unit cell. The vector k is
points on the reciprocal lattice. The importance of the called the crystal momentum.
Brillouin zone stems from the Bloch wave description of The vector k can be chosen within the first Brillouin
waves. one. Such choice is known as mapping the bands in the
Similarly all k points where the point 0 is the second reduced zone scheme.
closest reciprocal lattice point to that point constitute We can repeat a given Brillouin zone periodically
the second Brillouin zone, and so forth. Zone boundaries through all of wavevector space. To repeat a zone, we
are defined in terms of this definition of Brillouin zones. translate the zone by a reciprocal lattice vector. If we
Consider a Hamiltonian which is periodic, that is can translate a band from other zones into the first zone,
we can translate a band in the first zone into every other
H(r) = H(r + R), (203) zone. In this scheme the energy (k) of a band is a peri-
odic function in the reciprocal lattice. Such procedure is
where R is any lattice vector. known as mapping the bands in the periodic zone scheme.
44
Different bands are drawn in different zones in

wavevector space. Such procedure is known as mapping
the bands in the extended zone scheme.
We consider as an example the low-lying free electron
bands of a simple cubic lattice. Suppose we want to
exhibit the energy as a function of k in the [100] direction.
For convenience, choose units such that ~2 /2m = 1. In
Table on Fig. 105 we show several low-lying bands in this
empty lattice approximation with their energies (000) at
k = 0 and (kx 00) along the kx axis in the first zone:
FIG. 105.
B. Nearly Free Electron Model
The band structure of a crystal can often be explained

by the nearly free electron model for which the band
electrons are treated as perturbed only weakly by the
periodic potential of the ion cores. This model answers
almost all the qualitative questions about the behavior
of electrons in metals.
We start with completely free electrons whose Hamil-
tonian is FIG. 106. Low-lying free electron energy bands of the empty
sc lattice, as transformed to the first Brillouin zone and plot-
p2 ted vs. (kx 00). The free electron energy is ~2 (k + G)2 /2m,
H0 = (205)
2m where the Gs are given in the second column of the Table on
The corresponding energy eigenstates, the plane waves Fig. 105. The bold curves are in the first Brillouin zone, with
−π/a ≤ kx ≤ π/a. Energy bands drawn in this way are said
|k >, have eigenenergies given by
to be in the reduced zone scheme.
~2 |k|2
0 (k) = . (206)
2m (at any order of perturbation theory) that the only effect
Now consider a weak periodic potential perturbation to of V0 is to shift the energies of all of the eigenstates by
this Hamiltonian this constant. Henceforth we will assume that V0 = 0 for
simplicity.
H = H0 + V (r), (207) At second order in perturbation theory we have
with V periodic, meaning 0
X | < k + G|V |k > |2
(k) = 0 (k) + (210)
V (r) = V (r + R), (208) 0 (k) − 0 (k + G)
G
where R is any lattice vector. The matrix elements of where the ’ means that the sum is restricted to have
this potential are zero unless k0 k is a reciprocal lattice G 6= 0. In this sum, however, we have to be careful. It is
vector. possible that, for some k it happens that 0 (k+G) is very
We now apply perturbation theory. At first order in close to 0 (k) or perhaps they are even equal, in which
the perturbation V we have case the corresponding term of the sum diverges and the
(k) = 0 (k)+ < k|V |k >= 0 (k) + V0 , (209) perturbation expansion makes no sense. This situation
needs to be handled with degenerate perturbation theory.
which is just an uninteresting constant energy shift to If two plane-wave states |k > and |k + G > are
all of the eigenstates. In fact, it is an exact statement of approximately the same energy (meaning that |k >
45
and |k + G > are close to zone boundaries), then we

must diagonalize the matrix elements of these states
first. Within this two-dimensional space we can write
any wavefunction as
|Ψ >= α|k > +β|k + G > . (211)
Using the variational principle to minimize the energy is
equivalent to solving the effective Schroedinger equation
VG∗

0 (k) α α
=E . (212)
VG 0 (k + G) β β
The characteristic equation determining E is then FIG. 108. Energy contours in a rectangular zone, calculated
in 2 d nearly free electron model.
(0 (k) − E) (0 (k + G) − E) − |VG |2 = 0. (213)
The simplest case we can consider is when k is precisely
on a zone boundary (and therefore k + G is also precisely
on a zone boundary). In this case 0 (k) = 0 (k + G), our
characteristic equation simplifies, and we get
E± = 0 (k) ± |VG |. (214)
Thus we see that a gap opens up at the zone boundary.
Whereas both k and k + G had energy 0 (k) in the ab- FIG. 109. Reduced zone scheme for Fig. 108.
sence of the added potential VG , when the potential is
added, the two eigenstates form two linear combinations
with energies split by ±|VG |. at each atomic site. The method is closely related to
the LCAO method (linear combination of atomic orbitals
method) used in chemistry. Consider the tight binding
model in one dimension, that is a chain of atoms as shown
in Fig. 111.
There is a single orbital on atom n which we call |n >.
For convenience we will assume that the system has pe-
riodic boundary conditions (i.e., there are N sites, and
site N is the same as site 0). Further, we assume that all
of the orbitals are orthogonal to each other
< n|m >= δnm . (215)
Let us now take a general trial wavefunction of the form

X
|Ψ >= φn |n > . (216)
n
The effective Schroedinger equation can be written as2

X
Hnm φm = Eφn , (217)
m
FIG. 107. Three energy bands of a nearly free electron model
in 1d plotted in (a) the extended, (b) reduced, and (c) re- where Hnm is the matrix element of the Hamiltonian
peated zone schemes.
Hnm =< n|H|m > . (218)
We write the Hamiltonian as

X
C. Tight Binding Model H=K+ Vj , (219)
j
In solid-state physics, the tight binding model is an
approach to the calculation of electronic band structure where
using an approximate set of wave functions based upon P2
superposition of wave functions for isolated atoms located K= , Vj = V (r − Rj ). (220)
2m
46
where we have defined 0 = atomic + V . The eigenfunc-

tion of the Hamiltonian are of the form
φn e−ikna , (226)
and the spectrum is
(k) = 0 − t eika + e−ika = 0 − 2t cos ka. (227)

1. s-state Bands
An s-bans arising from a single atomic s-level for a

simple cubic structure has the dispersion law
FIG. 110. Repeated first zone for Fig. 108. X
(k) = 0 − t eik·δ ≡ 0 − tγ(k), (228)
δ
where the sum is with respect to all nearest neighbors.

The function γ(k) always appears as the dispersion in s-
states tight binding models where the interaction is lim-
ited to nearest neighbors.
So for the cubic lattice we have
FIG. 111. The tight binding chain. There is one orbital on (k) = 0 − 2t(cos(kx a) + cos(ky a) + cos(kz a)).(229)
each atom, and electrons are allowed to hop from one atom
to the neighboring atom. Problem C.1 Find the dispersion law for the square 2-d
lattice.
With these definitions we have

X
H|m >= (K + Vm )|m > + Vj |m > . (221)
j6=m
If we take the tight-binding orbitals |m > to be the

atomic orbitals, we have
(K + Vm )|m >= atomic |m > . (222)
Thus we obtain
X
Hnm = atomic δn,m + < n|Vj |m > . (223)
j6=m
We now have to figure out what the final term of this FIG. 112. Constant energy lines in the section kz = 0 of
equation is. The diagonal matrix element is a Hartree Brillouin zone of a simple cubic lattice, for the assumed energy
term, where the electron on one orbital interacts with band (k) = 0 − 2t(cos(kx a) + cos(ky a) + cos(kz a)).
the potential energy of all other sites. The meaning of
the non-diagonal term is that, via the interaction with
The s-bans arising from a single atomic s-level for an
some nucleus which is not the mth , an electron on the
fcc structure with twelve nearest neighbors has the dis-
mth atom can be transferred (can hop) to the nth atom.
persion law
Generally this can only happen if n and m are very close
to each other. Thus, we assume
kx a

ky a

 (k) = 0 − 4t cos cos (230)
X  V0 n = m 2 2
< n|Vj |m >= −t n = m ± 1 (224)

ky a kz a kz a kx a
j6=m
 0 otherwise + cos cos + cos cos .
2 2 2 2
With the above matrix elements we obtain the Hamil-
Such model can describe, for example, alkaline metal.
tonian in the form
Hnm = 0 δn,m − t (δn−1,m + δn+1,m ) , (225) Problem C.2 Derive Eq. (230).
47
FIG. 115. d-orbitals.
FIG. 113. Constant energy (k) = 0 surface in the Brillouin

zone of a simple cubic lattice, for the assumed energy band
(k) = 0 − 2t(cos(kx a) + cos(ky a) + cos(kz a)). The filled
volume contains one electron per primitive cell.
2. p-state Bands
An example using p-states is a two-dimensional square

lattice of an imaginary material DO2 . Here O is oxygen
and D is a cation with p valence orbitals. This example is
inspired by the cuprate materials that are high temper- FIG. 116. (a) Two-dimensional lattice structure of DO2 . (b)
ature superconductors, where the D atoms are copper. px orbital points toward toward the D atom, with plus or-
Copper has bonding by d-orbitals. The present example bital in one direction and minus in the other. pz orbitals are
is simplified by having the D atoms containing a single perpendicular to the plane.
s-orbital.
orbitals give an effective Hamiltonian of

Ep 4w2 cos θx cos θy
H= , (231)
4w2 cos θx cos θy Ep
where θµ = kµ a/2. The overlap constant w2 should be
small, since it is a next nearest neighbor hopping. The
dispersion law of these two bands is
E(k) = Ep ± 4w2 cos θx cos θy . (232)
Three other bands are constructed from |s > orbitals
of the cation and the two orbitals on the oxygen pointing
towards the cation. The effective Hamiltonian is
 
Es 2iw1 sin θx 2iw1 sin θy
H =  −2iw1 sin θx Ep 0  (, 233)
−2iw1 sin θy 0 Ep
FIG. 114. High-temperature superconductor. and the dispersion law is

E(k) = Ep (234)
1
Problem C.3 Find chemical formula of the compound E(k) = [Es + Ep
2
with the elementary cell presented on Fig. 114. q
± (Es − Ep )2 + 16w12 sin2 θx + sin2 θy .

The s-state orbital of the D atom |s > and has an

energy Es . The oxygen atoms has orbitals |pµ >, where The final two bands are obtained from the overlap of
µ = (x, y, z). Their eigenvalue is Ep . |pz > orbitals have the oxygen p-orbitals pointing perpendiculat to the near-
nonzero matrix elements only with themselves. These est neighbor cation. They have nonzero second order
48
ovelaps w20 with each other. Their dispersion is down the energy of all states within the first zone and
pushes up energy of all states in the second zone. As a
E(k) = Ep ± 4w20 cos θx cos θy . (235)
result, states close to the boundary get filled up preferen-
This completes the discussion of the seven energy bands tially at the expense of states further from the boundary.
due to the overlaps of the seven orbitals of this imaginary This deforms the Fermi surface roughly as shown in the
material. middle of Fig. 118.
FIG. 118. Fermi sea of a square lattice of monovalent atoms

in two dimensions as the strength of the periodic potential
is varied. Left: In the absence of a periodic potential, the
Fermi sea forms a circle whose area is precisely half that of
the Brillouin zone (the black square). Middle: when a pe-
riodic potential is added, states closer to the zone boundary
are pushed down in energy deforming the Fermi sea. Right:
With strong periodic potential, the Fermi surface touches the
Brillouin zone boundary. The Fermi surface remains continu-
ous since the Brillouin zone is periodic. The area of the Fermi
sea remains fixed as the strength of the periodic potential is
changed.
The physics shown in Fig. 118 is quite similar to what

is seen in many real materials. In Fig. 119 we show
the Fermi surfaces for the monovalent metals potassium,
lithium, and copper. Potassium is an almost ideal free
electron metal, with an almost precisely spherical Fermi
FIG. 117.
surface. For lithium, the Fermi surface is slightly dis-
torted, bulging out near the zone boundaries. For copper,
the Fermi surface is greatly distorted touching the Bril-
louin zone boundary in tubes. In all three cases, however,
the Fermi seas fill exactly half the Brillouin zone volume
D. Fermi Surface
since the metals are monovalent.
Let us now consider the case of divalent atoms. In the
Of all the constant energy surfaces the most important absence of a periodic potential, the Fermi surface is still
for us is the Fermi surface. The Fermi surface is the circular, although it now crosses into the second Brillouin
boundary between the filled and unfilled states. zone. For the weak or intermediate periodic potential, in
It is useful to try to understand how the nearly-free spite of the gap at the Zone boundaries, the lowest band
electron model explains the shape of the Fermi surface and the second lowest one overlap; we still have partially
in two dimensions (which is actually a Fermi line). The filled bans and, hence, a metal (or a semimetal).
number of k-states in a single Brillouin zone is equal to
the number of unit cells in the entire system. Thus, for Problem D.1 For the Fermi surface in the absence of
monovalent atoms there is enough electrons to half fill a potential presented on Fig. 120 plot the part of the Fermi
single Brillouin zone. In the absence of a periodic poten- surface in the second zone on reduced and repeated zone
tial, the Fermi sea forms a circular disc as shown in the schemes.
left of Fig. 118. The area of this disc is precisely half the
area of the zone. When a periodic potential is added a For a strong enough potential, the gaps at the zone
gap opens at the zone boundary this gap opening pushes boundaries may be so large, that the entire lower band
49
FIG. 119. Fermi surfaces of monovalent metals potassium FIG. 121. Fermi surface of the (fcc) divalent metal calcium.
(left), lithium (middle) and copper (right). The wire frames Left: The first band is almost completely filled with electrons.
demark the first Brillouin zones, which are half filled. Left: The solid region is where the Fermi surface is inside the Bril-
The potassium Fermi surface is almost exactly spherical, cor- louin zone boundary (the zone boundary is demarcated by the
responding to nearly free electrons in very weak periodic po- wire frame). Right: A few electrons fill small pockets in the
tentials. Middle: The periodic potential for lithium is slightly second band. There are just enough electrons to completely
stronger, and correspondingly the Fermi surface is distorted fill the lowest band, but there are states in the lowest band
towards the zone boundaries. Right: The periodic potential which are of higher energy than some states in the second
is so strong in the case of copper that the Fermi surface inter- band, so that a few states are empty in the first band (left),
sects the Brillouin zone boundary. In copper (like in two other and a few states are filled in the second band (right).
noble metals, Au and Ag) the free electron sphere bulges out
in < 111 > directions to make contact with the hexagonal
zone faces.
some amount of current.
FIG. 122. Band scheme for a band insulator.
FIG. 120. Fermi sea of a square lattice of divalent atoms in

two dimensions. In the absence of a periodic potential, the
Fermi sea forms a circle whose area is precisely that of the
Brillouin zone (the black square). When an intermediately XIII. ELECTRON DYNAMICS
strong periodic potential is added, the Fermi sea is a distorted
circle. When a sufficiently strong periodic potential is added, A. Quasi-Classical Dynamics of Electrons
states inside the zone boundary are pushed down in energy
so that all of these states are filled and no states outside the
first Brillouin zone are filled. Since there is a gap at the zone
Under the one band approximation the Bloch electron
boundary, this situation is an insulator. (The area of the can be described as a particle with the classical Hamilton
Fermi sea remains fixed.) function
H(p, r) = (p) − eφ(r), p = ~k. (236)

is filled and the upper band is empty. In general, when Here (p) is the Bloch dispersion law and φ(r) is the
the Fermi energy is in the gap, the highest filled band is scalar potential. To take account of magnetic field one
known as the valence band and the lowest empty band is should replace the momentum p by the kinematic one
known as the conduction band. Because of the gap be-
tween the bands, at zero temperature, a sufficiently small p → P = p + eA, (237)
electric perturbation will not create any excitations the
system does not respond at all to electric field. Thus, where A is the vector-potential which is connected with
systems of this type are known as (electrical) insulators the magnetic field B by the relation
(or more specifically band insulators). If the band gap
B = ∇ × A. (238)
is below about 4 eV, these type of insulators are called
semiconductors, since at room temperature a few elec- Consequently, we have
trons can be thermally excited into the conduction band,
and these electrons then can move around freely, carrying H(p, r) = (p + eA) − eφ(r). (239)
50
The classical Hamilton equations are

∂H ∂H
ṙ = v = , ṗ = − . (240)
∂p ∂r
Finally we obtain the Lorentz force equation
~k̇ = −e(E + v(k) × B), (241)
where
∂(k) FIG. 124. Motion in a magnetic field of the wavevector of an
v(k) = . (242) electron on the Fermi surface, in (a) and (b) for electron and
∂k hole Fermi surfaces respectively. In (a) the wavevector moves
Consider the case of magnetic field only. around the orbit in a clockwise direction; in (b) the wavevec-
tor moves around the orbit in a counter-clockwise direction.
e
k̇ = − v(k) × B. (243) The direction in (b) is what we expect for a free electron of
~ positive charge; the smaller k values have the lower energy, so
Eq. (243) means that the change of vector k that the filled electron states lie inside the Fermi surface. We
(i) normal to the direction of B, call the orbit in (b) electronlike. The sense of the motion in
(ii) normal to v, which is itself normal to the energy a magnetic field is opposite in (a) to that in (b), so that we
refer to the orbit in (a) as holelike. A hole moves as a particle
surface.
of positive charge. In (c) for a rectangular zone we show the
It follows immediately that the component of k along the motion on an open orbit in the periodic zone scheme. An
magnetic field and the energy (k) are both constants of open orbit is topologically intermediate between a hole orbit
motion. Hence electrons move along curves given by the and an electron orbit.
intersection of surfaces of constant energy with planes
perpendicular to magnetic field.
FIG. 125. Indicating only a few of the surprisingly many types

of orbits an electron can pursue in k-space when magnetic field
is applied to a noble metal. The figure displays (a) a closed
FIG. 123. Orbit of an electron in magnetic field.
particle orbit; (b) a closed hole orbit; (c) an open orbit.
B. Boltzmann Equation and the scattering term is

∂fk gk
Boltzmann equation has the form =− , (247)
∂t scatt τ
∂fk ∂fk ∂fk ∂fk
= + + . (244)
∂t ∂t dif f ∂t f ield ∂t scatt where gk = fk − fk0 .
Here the diffusion term is Consider homogeneous linear conductivity in the static
electric field and in the absence of magnetic field. We
∂fk ∂fk obtain Boltzmann equation in the form
= −vk = −vk ∇fk , (245)
∂t dif f ∂r
∂f 0

gk
the acceleration term is = −evk · E − . (248)
τ ∂
∂fk ∂fk ∂fk
= −k̇ = e(E + v(k) × B) , (246)
∂t f ield ∂k ∂k After solving Boltzmann equation we can calculate the
51
current
Z
J=2 −evk gk dk (249)
∂f 0 dS
Z
1 2
= e τ v k (v k · E) − d
4π 3 ∂ ~vk
1 e2 τ
Z
vk vk dSF
= 3
· E.
4π ~ vk
(250)
For crystals with cubic symmetry, the conductivity is
1 e2 τ 1
Z
σ= ΛdSF , (251)
4π 3 ~ 3
where we introduced the mean free path
Λ = vτ. (252)
Problem B.1 Show that for quadratic dispersion law
Eq. (251) takes the form
ne2 τ
σ= . (253)
m
Problem B.2 Obtain analog of Eq. (251) fot the fre-
quency dependent conductivity.
FIG. 126. Carrier concentrations for metals, semimetals, and

XIV. SEMICONDUCTORS semiconductors. The semiconductor range may be extended
upward by increasing the impurity concentration, and the
range can be extended downward to merge eventually with
Carrier concentrations representative of metals, the insulator range.
semimetals, and semiconductors are shown in Fig. 126.
Semiconductors are generally classified by their electrical
resistivity at room temperature, with values in the range
of 10−2 to 109 ohm-cm, and strongly dependent on purity
of the sample, temperature and other physical factors.
A. Optical Properties of Semiconductors
Devices based on semiconductors include transistors,
switches, diodes, photovoltaic cells, detectors, and ther-
mistors. These may be used as single circuit elements
or as components of integrated circuits. We discuss in Band insulators cannot absorb photons which have en-
this chapter the central physical features of the classical ergies less than their band-gap energy. The reason for
semiconductor crystals, particularly silicon, germanium, this is that a single such photon does not have the en-
and gallium arsenide. ergy to excite an electron from the valence band into the
Some useful nomenclature: the semiconductor com- conduction band. Since the valence band is completely
pounds of chemical formula AB, where A is a trivalent filled, the minimum energy excitation is of the band-gap
element and B is a pentavalent element, are called III- energy, so a low energy photon creates no excitations at
V (three-five) compounds. Examples are indium anti- all. As a result, these low-energy photons do not get ab-
monide and gallium arsenide. Where A is divalent and B sorbed by this material at all, and they simply pass right
is hexavalent, the compound is called a II-VI compound; through the material.
examples are zinc sulfide and cadmium sulfide. Silicon
and germanium are sometimes called diamond-type semi- While the band gap determines the minimum energy
conductors, because they have the crystal structure of di- excitation that can be made in an insulator (or semicon-
amond. Diamond itself is more an insulator rather than a ductor), this is not the complete story in determining
semiconductor. Silicon carbide SiC is a IV-IV compound. whether or not a photon can be absorbed by a material.
A highly purified semiconductor exhibits intrinsic con- It turns out to matter quite a bit at which values of k
ductivity, as distinguished from the impurity conductiv- the maximum of the valence band and the minimum of
ity of less pure specimens. An electronic band scheme the conduction band lies. If the value of k for the valence
leading to intrinsic conductivity is indicated in Fig. 122. band maximum is the same as the value of k for the con-
The conduction band is vacant at absolute zero and is duction band minimum, then we say that it is a direct
separated by an energy gap Eg from the filled valence band gap. If the values of k differ, then we say that it is
band. an indirect band gap.
52
FIG. 128. In (a) the lowest point of the conduction hand

occurs at the same value of k as the highest point of the va-
lence band. A direct optical transition is drawn vertically
with no significant change of k, because the absorbed pho-
ton has a very small wavevector. The threshold frequency ωg
for absorption by the direct transition determines the energy
gap Eg = ~ωg . The indirect transition in (b) involves both
a photon and a phonon because the hand edges of the con-
duction and valence bands are widely separated in k space.
The threshold energy for the indirect process in (h) is greater
than the true band gap. The absorption threshold for the in-
FIG. 127. Electrical resistivities vary over many orders of direct transition between the hand edges is at ~ω = Eg + ~Ω,
magnitude for metals, semiconductors, and insulators. where Ω is the frequency of an emitted phonon of wavevec-
tor K ≈ −kg . At higher temperatures phonons are already
present; if a phonon is absorbed along with a photon, the
B. Electrons and Holes threshold energy is ~ω = Eg − ~Ω. Note: The figure shows
only the threshold transitions. Transitions occur generally
between almost all points of the two hands for which the
Suppose we start with an insulator or semiconductor wavevectors and energy can he conserved.
and we excite one electron from the valence band to the
conduction band, as shown in the left of Fig. 130. This
excitation may be due to absorbing a photon, or it might (with the derivative being taken in any direction for an
be a thermal excitation. (For simplicity in the figure we isotropic system). Correspondingly, the (group) velocity
have shown a direct band gap. When the electron has is given by
been moved up to the conduction band, there is an ab- ~k
sence of an electron in the valence band known as a hole. v= . (255)
me
Since a completely filled band is inert, it is very conve-
nient to only keep track of the few holes in the valence Analogously we can define an effective mass for holes.
band (assuming there are only a few) and to treat these We define the effective mass of a hole as
1 ∂2E

holes as individual elementary particles. The electron 1
can fall back into the empty state that is the hole, emit- =− 2 (256)
mh ~ ∂k 2 k=0
ting energy (a photon, say) and annihilating both the
electron from the conduction band and the hole from the Correspondingly, the (group) velocity is given by
valence band. While the electrical charge of an electron ~k
is negative the electrical charge of a hole (the absence of v= . (257)
mh
an electron) is positiveequal and opposite to that of the
electron. In the semiclassical picture, we can write simple New-
tons equations for electrons and holes
dv e
C. Effective Mass =− (E + v × B) (258)
dt me
dv e
= (E + v × B). (259)
It is useful to describe the states near the edge of a dt mh
band in terms of an effective mass. Let the lower edge of
the conduction band is at the point Γ. We then define
the effective mass of an electron to be given by 1. Silicon and Germanium
∂2E

1 1 It often occurs that the conduction band has more than
= 2 (254)
me ~ ∂k 2 k=0
one minimum at different points kmin in the Brillouin
53
FIG. 129. Optical absorption of several semiconductors (Si,

Ge, GaAs) as a function of photon wavelength or energy. FIG. 131. Band structure of GaAs.
For energies below the bandgap, the absorption is extremely
small. (Note the absorption is on a log scale!) GaAs has a
(direct) gap energy 1.44 eV where the absorption drops very
rapidly. For Ge the drop at 0.8 eV reflects the direct band
gap energy. There is still some weak absorption at longer
wavelengths due to the smaller energy indirect band gap. Si
has a somewhat more complicated band structure with many
indirect transitions and direct gap of 3.4 eV.
FIG. 132. Effective masses of electrons and holes in direct-gap

semiconductors.
In many semiconductors it has been possible to de-

termine by cyclotron resonance the effective masses of
carriers in the conduction and valence bands near the
band edges. The determination of the energy surface is
equivalent to a determination of the effective mass ten-
FIG. 130. Electrons and holes in a semiconductor. Left: A sor. Cyclotron resonance in a semiconductor is carried
single hole in the valence band and a single electron in the out with centimeter wave or millimeter wave radiation at
conduction band. Right: Moving the hole to a momentum
low carrier concentration.
away from the top of the valence band costs positive energy
- like pushing a balloon under water. As such, the effective For silicon the conduction band minima lie on the six
mass of the hole is defined to be positive. equivalent ∆ -lines along < 100 >, and occur at about
0.85% of the way to the zone boundary. The effective
mass of these anisotropic minima is characterized by
zone with exactly the same energy (such a situation oc- a longitudinal mass along the corresponding equivalent
curs due to the symmetry of the crystal), and the dis- (100) direction and two transverse masses in the plane
persion law in the vicinity of a minimum is anisotropic, perpendicular to the longitudinal direction. The longi-
as for electrons in Si or Ge.. We then say that there are tudinal electron mass is ml = 0.98m and the transverse
multiple valleys in the band structure. It is easy to gen- electron masses are mt = 0.19m, where m = 9.11×10−31
eralize Eqs. (254), (258) to take account the anisotropy. kg is the free electron rest mass. Two of the three valence
We introduce the components of the reciprocal effective band maxima occur at 0 eV. These bands are referred to
mass tensor as the light and heavy hole bands with a light hole mass
of mlh = 0.16m and a heavy hole mass of mhh = 0.46m.
2
1 1 ∂ E
= 2 (260) In addition there is a split-off hole band with its max-
me µν ~ ∂kµ ∂kν k=kmin
imum at Ev,so = −0.044 eV and a split-off hole mass
dvµ 1 of msoh = 0.29m. In germanium the conduction band
= −e (E + v × B)ν . (261)
dt me µν minima are at the L points.
54
FIG. 135. Constant energy ellipsoids for electrons in silicon,

drawn for m∗l /m∗t = 5.
FIG. 133. Standard labels of the symmetry points and axes

of the Brillouin zone of the fcc lattice. The zone center is Γ.
The boundary point at (2π/a)(100) is X; the boundary point
at 2π/a 21 , 12 , 12 is L; the line ∆ runs between Γ and X.

FIG. 136. Band structure of germanium.
proton and an extra electron. Since the valence band is

FIG. 134. Band structure of silicon. already filled this additional electron must go into the
conduction band. The As atom is known as a donor (or
electron donor) in silicon since it donates an electron to
D. Doping the conduction band. It is also sometimes known as an n-
dopant, since n is the symbol for the density of electrons
in the conduction band.
In a pure band insulator or semiconductor, if we excite
electrons from the valence to the conduction band (ei- Analogously, we can consider boron, the element be-
ther with photons or thermally) we can be assured that longing to the group directly to the left of Si on the pe-
the density of electrons in the conduction band (typi- riodic table. In this case, the boron dopant provides one
cally called n, which stands for negative charges) is pre- fewer electron than Si, so there will be one electron miss-
cisely equal to the density of holes left behind in the va- ing from the valence band. In this case B is known as an
lence band (typically called p, which stands for positive electron acceptor, or equivalently as a p-dopant, since p
charges). However, in an impure semiconductor or band is the symbol for the density of holes.
insulator this is not the case.
Without impurities, a semiconductor is known as in-
trinsic. The opposite of intrinsic, the case where there 1. Impurity States
are impurities present, is sometimes known as extrinsic.
Let us now examine the extrinsic case more carefully. Let us consider what happens when we add dopants.
Consider for example, silicon (Si), which is a semicon- For definiteness let us consider adding an n-dopant such
ductor with a band gap of about 1.1 eV. Now imagine as P to a semiconductor such as Si. Once we add a
that a arsenic (As) atom replaces one of the Si atoms in single n-dopant to an otherwise intrinsic sample of Si,
the lattice as shown in Fig. 137. This As atom, being we get a single electron above the gap in the conduction
directly to the right of Si on the periodic table, can be band. This electron behaves like a free particle with mass
thought of as nothing more than a Si atom plus an extra me . However, in addition, we have a single extra posi-
55
Because the dielectric constant of semiconductors is typ-

ically high (on the order of 10 for most common semi-
conductors) and because the effective mass is frequently
low (a third of m or even smaller), the effective Ryd-
berg Ryef f can be tiny compared to the real Rydberg.
Thus this donor impurity forms an energy eigenstate with
energy just below the bottom of the conduction band
(Ryef f below the band bottom only). At zero temper-
FIG. 137. Left: Charges associated with an arsenic impurity ature this eigenstate will be filled, but it takes only a
atom in silicon. Arsenic has five valence electrons, but silicon small temperature to excite a bound electron out of a
has only four valence electrons. Thus four electrons on the hydrogenic orbital and into the conduction band.
arsenic form tetrahedral covalent bonds similar to silicon, and
the fifth electron is available for conduction. The arsenic atom
is called a donor because when ionized it donates an electron
to the conduction band. Right: Boron has only three valence
electrons; it can complete its tetrahedral bonds only by taking
an electron from a Si-Si bond, leaving behind a hole in the
silicon valence band. The positive hole is then available for
conduction. The boron atom is called an acceptor because
when ionized it accepts an electron from the valence band.
At 0 K the hole is bound.
tive charge +e at some point in the crystal due to the P

nucleus. The free electron is attracted back to this posi-
tive charge and forms a bound state that is similar to a
hydrogen atom. There are two main differences between
a real hydrogen atom and this bound state of an electron FIG. 138. Energy diagram of a doped semiconductor (left)
in the conduction band and the impurity nucleus. First with donor impurities, or (right) with acceptor impurities.
of all, the electron has effective mass me which can be The energy eigenstates of the hydrogenic orbitals tied to the
very different from the real (bare) mass of the electron impurities are not all the same because each impurity is per-
(and is typically smaller than the bare mass of the elec- turbed by neighbor impurities. At low temperature, the donor
impurity eigenstates are filled and the acceptor eigenstates are
tron). Secondly, instead of the two charges attracting
empty. But with increasing temperature, the electrons in the
each other with a potential V = e2 /(4π0 r) they attract donor eigenstates are excited into the conduction band, and
each other with a potential V = e2 /(4πr 0 r), where r is similarly the holes in the acceptor eigenstates are excited into
the relative permittivity (or relative dielectric constant) the valence band.
of the material.
The eigenenergies of the hydrogen atom are
Ry
En = −
, (262)
n2 E. Statistical Mechanics of Semiconductors
where Ry is the Rydberg constant given by
me4 From Eq. (116) we obtain in isotropic case
Ry = = 13.6eV. (263)
820 h2 p
2m3 √
The corresponding radii of these states are rn = n2 aB gc ( ≥ c ) = 2 3e − c (267)
with the Bohr radius given by pπ ~
2m3 √
gv ( ≤ v ) = 2 3h v − . (268)
4π20 ~2 π ~
aB = = .51 × 10−10 m. (264)
me2
Problem E.1 Prove that for the case of two transverse
The analogous calculation for a hydrogenic impurity and one longitudinal masses Eq. (267) becomes
state in a semiconductor gives precisely the same expres-
sion, only 0 is replaced by r 0 and m is replaced by me . p
2m2t ml √
One obtains gc ( ≥ c ) = − c . (269)
π 2 ~3
me 1
Ryef f = Ry , (265) If the chemical potential is well above the bottom of the
m 2r
conduction band (i.e., if β(µ − c ) 1), then we recover
and Eq. (117) with µ substituted by µ − c . Such electron

m gas is called degenerate. The situation can occur only in
aef
B
f
= aB r . (266)
me heavily doped semiconductors.
56
Back in Section IV we studied Drude theory which

did not treat the Pauli exclusion principle properly: it
neglected the fact that in metals the high density of elec-
trons makes the Fermi energy extremely high. However,
in semiconductors or band insulators, when only a few
electrons are in the conduction band and/or only a few
holes are in the valence band, then we can consider this
to be a low-density situation, and to a very good approxi-
mation, we can ignore Fermi statistics. As a result, when
there is a low density of conduction electrons or valence
holes, it turns out that Drude theory works extremely
well.
FIG. 139. A band diagram of a semiconductor near the top
of the valence band (mostly filled) and the bottom of the con- Problem E.2 Consider electron and hole gases men-
duction band (mostly empty). This diagram shows a direct tioned in the caption to Fig. 19. Are the gases degenerate
band gap, but the considerations of this section apply to in- or non-degenerate?
direct gaps as well.
XV. MAGNETIC PROPERTIES OF SOLIDS

Further on we’ll consider the case of undoped, or lightly
doped semiconductor, for which the opposite limit is rel- A paramagnet is a material where χ > 0 (i.e., the
evant: the chemical potential is well below the bottom of resulting magnetization is in the same direction as the
the conduction band (i.e., β(c − µ) 1). In this case applied field). A diamagnet is a material where χ < 0
1 (i.e., the resulting magnetization is in the opposite direc-
= e−β(−µ) , (270) tion from the applied field). Both paramagnetism and
eβ(−µ) −1
diamagnetism were discovered by M. Faradey.
that is we get Maxwell-Boltzmann statistics. For the A ferromagnet is a material where M can be nonzero,
concentration we obtain even in the absence of any applied magnetic field. Fer-
p
2m3
Z ∞
√ romagnet mineral Fe3 O4 called magnetite was known to
n(T ) = 2 3e e−β(c −µ) d − c e−β(−c ) the ancients.
π ~ c In materials that exhibit antiferromagnetism, the mag-
3/2
netic moments of atoms or molecules, usually related to

1 2me kB T
= e−β(c −µ) (271) the spins of electrons, align in a regular pattern with
4 π~2
neighboring spins (on different sublattices) pointing in
Such electron gas is called non-degenerate (or classical). opposite directions. This is, like ferromagnetism and
The chemical potential is also well above the top of the ferrimagnetism, a manifestation of ordered magnetism.
valence band, so for holes we obtain This type of antiperiodic ground state is sometimes
3/2 known as a Neel state after Louis Neel who proposed
1 2mh kB T in the 1930s that such states exist.
p(T ) = e−β(µ−v ) . (272)
4 π~2
Thus the law of mass action A. Atomic Energy Levels
3
1 kB T 3/2
n(T )p(T ) = (me mh ) e−βEgap , (273) In the non-relativistic approximation, for a system of
2 π~2
particles in a centrally symmetric external field the to-
is valid independently of doping of the material. tal angular momentum L is conserved. Every stationary
For an intrinsic (i.e., undoped) semiconductor the state of an atom is also characterised by the total spin S
number of electrons excited into the conduction band of the electrons. The energy level having definite values
must be equal to the number of holes left behind in the of L and S is called a term and is degenerate to a degree
valence band, so p = n. We obtain then equal to the number of possible directions of space of L
1 3 times that of S.
µ= (c + v ) + (kB T ) log(mh /mh ), (274) In fact, the electromagnetic interactions of the elec-
2 4
trons contains relativistic effects, which depend on their
and spin. These effects lead in particular to the fact that
3/2 L and S are not conserved separately. Only the total
1 kB T
n(T ) = p(T ) = √ (me mh )3/4 e−βEgap /2 . angular momentum J = L + S is conserved. For this
2 π~2 reason, the exact energy levels must be characterised by
(275) the values J of the total angular momentum. A set of
57
FIG. 140. Magnetic type of the elements in the periodic table.

For elements without a color designation magnetism is even
smaller or not explored.
levels of different J values arises out of a given L and

S. These levels are split by the spin-orbit interaction.
This splitting is called a fine structure (or a multiplet
splitting). Each individual level is still 2J + 1 degenerate
with respect to the direction of vector J. A combination
of S, L, J and MJ determines a single state.
There is a generally accepted notation to denote the
atomic energy levels. Terms with different values of the
total angular momentum L are denoted by capital Latin
letters similar to the individual electron states: L = 0 is FIG. 141. Low lying carbon levels. (Check up that you un-
called S level, L = 1 is called P level etc. Above and to derstand why there are 15 states for the 2s2 2p2 electron con-
the left of this letter is placed the number 2S + 1 called figuration.)
the multiplicity of the term. Below and to the right of the
letter is placed the value of the total angular momentum
J. Thus the symbols 2 P1/2 , 2 P3/2 denote levels with The exclusion principle prevents two electrons of the
L = 1, S = 1/2, J = 1/2 and J = 3/2. same spin from being at the same place at the same time.
In atomic physics, Hund’s rules refers to a set of rules Thus electrons of the same spin are kept apart, further
that German physicist Friedrich Hund formulated around apart than electrons of opposite spin. Because of the
1927, which are used to determine the level that cor- coulomb interaction the energy of electrons of the same
responds to the ground state of a multi-electron atom. spin is lowerthe average potential energy is less positive
These three rules are: for parallel spin than for antiparallel spin.
1. For a given electron configuration, the term with
maximum multiplicity has the lowest energy. The mul-
tiplicity is equal to 2S + 1, where S is the total spin B. Atom in Magnetic Field
angular momentum for all electrons. Therefore, the term
with lowest energy is also the term with maximum S.
2. For a given multiplicity, the term with the largest For a uniform magnetic field, we may take A = 12 B × r
value of the total orbital angular momentum quantum and write down the Hamiltonian of electrons in atoms in
number L, has the lowest energy. the presence of magnetic field as
3. For a given term, in an atom with outermost sub- X p2 e2

a e 2
shell half-filled or less, the level with the lowest value H= + pa · [B × ra ] + |B × ra |
of the total angular momentum quantum number J lies a
2m 2m 8m
lowest in energy. If the outermost shell is more than half- + V (r) + 2µB B · S, (276)
filled, the level with the highest value of J, is lowest in
energy. where the Bohr magneton µb is defined as e~/2mc in CGS
The first Hund’s rule has its origin in the exclusion and e~/2m in SI. It is closely equal to the spin magnetic
principle and the coulomb repulsion between electrons. moment of a free electron; µB = 5.79 eV/T. The second
58
2µB B · σ to give the Hamiltonian

e2 X 2
H = H0 + µB B · (L + 2S) + |B × ra | ,(278)
8m a
where H0 is the Hamiltonian in the absence of the applied

magnetic field.
The second term in the R.H.S. of Eq. (278), known
sometimes as the paramagnetic term, is clearly just the
coupling of the external field to the total magnetic mo-
ment of the electron. This term tends to aligns this mo-
ment with the B-field and results in paramagnetism.
The third term in the R.H.S. of Eq. (278) is known
as the diamagnetic term of the Hamiltonian, and will
be responsible for the effect of diamagnetism. Since this
term causes an increase in the total energy of the atom
when the magnetic field is applied (at least in the first
order of perturbation theory), it has the opposite effect
from that of the above-considered paramagnetic term.
C. Lande g-factor
We shall assume that the magnetic field is so weak

that the second and the third term can be considered as
perturbations. Using the first order perturbation theory
and ignoring the third term we get
∆E = µB B· < L + 2S >= µB B· < J + S > (279)
(< · · · >≡< SLJ| . . . |SlJ >). Taking into account that
the averaged value of spin is in the direction of J, we get

<J·S>
∆E = µB B · J 1 +
J2
J + S − L2
2 2

= µB B · J 1 + (280)
2J2
Thus an atom with angular momentum J has 2J + 1
equally spaced energy levels.
Em = mgµB B, (281)
where m is the azimuthal quantum number and has the
values J, J − 1, . . . , J, and the g factor is given by the
Lande equation
J(J + 1) + S(S + 1) − L(L + 1)
g =1+ . (282)
2J(J + 1)
FIG. 142. Basic levels of the atoms.
It is convenient to introduce the effective number of Bohr
magnetons p, defined as
term in Eq. (276) can be rewritten as p = g[J(J + 1)]1/2 . (283)
e X
pa · [B × ra ] = µB B · L, (277) D. Rare Earth Ions
2m a
The chemical properties of the trivalent rare earth ions

where L is the orbital angular momentum quantum num- ions are similar because the outermost electron shells
ber. This can then be grouped with the Zeeman term are identically in the 5s2 5p6 configuration, like neutral
59
xenon. The radii of the trivalent ions contract fairly E. Iron Group Ions
smoothly as we go through the group from 1.11 Å at
cerium to 0.94 Å at ytterbium. This is known as the Consider an example of how the application of Hund’s
lanthanide contraction. rules allows to understand the second column of the Ta-
ble 144. The ion Mn2+ has five electrons in the 3d shell,
What distinguishes the magnetic behavior of one ion
which is therefore half-filled. The spins can all be par-
species from another is the number of 4f electrons com-
allel if each electron enters a different orbital, and there
pacted in the inner shell with a radius of perhaps 0.3 Å.
are exactly five different orbitals available, characterized
In lanthanum, just before the rare earth group begins,
by the orbital quantum numbers m = 2, 1, 0, −1, . . . , −2.
the 4f shell is empty; at cerium there is one 4f electron,
These will be occupied P by one electron each. We expect
and the number of 4f electrons increases steadily through
S = 5/2, and because m = 0 the only possible value
the group until we have 4f 13 at ytterbium and the filled
of L is 0, as observed.
shell 4f 14 at lutecium. Even in the metals the 4f core
However, the experimental magneton numbers for salts
retains its integrity and its atomic properties: no other
of the iron transition group of the periodic table are in
group of elements in the periodic table is as interesting.
poor agreement with Eq. (282). The values often agree
Consider a couple of examples of how the application quite well with magneton numbers p = 2[S(S + 1)]1/2
of Hund’s rules allows to understand the second column calculated as if the orbital moment were not there at all.
of the Table 143. The ion Ce3+ has a single f electron;
an f electron has l = 3 and s = 1/2. Because the f
shell is less than half full, the J value by the preceding
rule is |L − S| = 3 − 1/2 = 5/2. The ion Pr3+ has two
f electrons; one of the rules tells us that the spins add
to give S = 1. Both f electrons cannot have m = 3
without violating the Pauli exclusion principle, so that
the maximum L consistent with the Pauli principle is
not 6, but 5. The f shell is less than half full. Hence the
J value is |L − S| = 5 − 1 = 4.
The calculated magneton numbers are compared with
g values from the Lande result. The agreement between
the theory and the experiment is quite good. The dis- FIG. 144. Effective magneton numbers for iron group ions.
crepancy is marked only for Eu3+ and Sm3+ ions. For
these ions it is necessary to consider the influence of the The difference in behavior of the rare earth and the
high states of the L-S term, as the intervals between suc- iron group salts is that the 4f shell responsible for para-
cessive levels of the term are not large compared to kB T magnetism in the rare earth ions lies deep inside the ions,
at room temperature. within the 5s and 5p shells, whereas in the iron group ions
the 3d shell responsible for paramagnetism is the outer-
most shell. The 3d shell experiences the intense inhomo-
geneous electric field, called the crystal field, produced
by neighboring ions. This interaction has two major ef-
fects: The coupling of L and S vectors is largely broken
up, so that the states are no longer specified by their J
values; further, the 2L + 1 sublevels belonging to a given
L which are degenerate in the free ion may now be split
by the crystal field. This splitting diminishes or com-
pletely quenches the contribution of the orbital motion
to the magnetic moment.
F. Paramagnetism
1. Langevin Paramagnetism
Paul Langevin first described paramagnetism as being

FIG. 143. Effective magneton numbers for trivalent lan- due to the presence of unpaired electrons in atomic or
thanide group ions. molecular orbitals in a 1905 publication. Consider ions
(atoms) with a partly filled inner shell: transition ele-
ments, rare earth and actinide elements. Paramagnetism
60
FIG. 147. John Hasbrouck Van Vleck (1899 – 1980) was

an American physicist and mathematician, awarded the 1977
FIG. 145. Energy levels split by an octahedral crystal field. Nobel Prize in Physics, for his contributions to the under-
standing of the behavior of electrons in magnetic solids. J.
H. Van Vleck established the fundamentals of the quantum
mechanical theory of magnetism and the crystal field theory
(chemical bonding in metal complexes).
Problem F.1 (a) Plot on one graph B1/2 (x), B1 (x) and
B3/2 (x). (b) Prove that for JgµB B/kB T 1 we have
moment = χB, (288)
where
C
χ= (289)
T
FIG. 146. Paul Langevin (1872 – 1946) was a prominent
(this is known as Curie’s law), and
French physicist who developed Langevin dynamics and the
Langevin equation. Langevin is noted for his work on para- g 2 µ2B J(J + 1) p2 µ2B
magnetism and diamagnetism; he devised the modern inter- C= = . (290)
pretation of this phenomenon in terms of spins of electrons 3kB 3kB
within atoms.
2. Van Vleck Temperature Independent Paramagnetism

is exhibited by many of these ions even when incorpo-
rated into solids, but not invariably. Van Vleck paramagnetism occurs when J = 0 even
The partition function of an atom in magnetic field is though L = S is non-zero (one can use Hunds rules to
show that this occurs if a shell has one electron fewer
exp 2J+1 2J+1

2J x − exp − 2J x than being half filled). This paramagnetism is due to
Z= x x , (284)
exp 2J − exp − 2J the second order perturbation theory contribution to the
energy of the paramagnetic term of Eq. (278). In fact,
where x = JgµB B/kB T . The free energy is for S = L 6= 0 the non-diagonal matrix elements of Lz , Sz
for the transitions S, L, J → S, L, J ± 1 are not in general
2J + 1 2J + 1 zero (we took axis ẑ in the direction of magnetic field).
F = − kB T ln exp x − exp − x
2J 2J The contribution to the energy we are talking about is
h x x i
+ kB T ln exp − exp − . (285) |< SLJ = 1|Lz + 2Sz |SLJ = 0 >| 2 2
2
2J 2J ∆E = µB B .
ESLJ=0 − ESLJ=1
The magnetization is
(291)
∂F
M =− = gJµB BJ (x), (286) Thus the magnetic moment of the atom is
∂B
where the Brillouin function BJ is defined by dE
moment = − (292)
dB
2J + 1 (2J + 1)x 1 x 2
|< SLJ = 1|Lz + 2Sz |SLJ = 0 >| 2
BJ (x) = ctanh − ctanh . =2 µB B.
2J 2J 2J 2J ESLJ=1 − ESLJ=0
(287)
61
G. Diamagnetism
A classic situation in which diamagnetism occurs is

for atoms with filled shell configurations, like the noble
gases where L = S = J = 0. In this case the param-
agnetic term of Eq. (278) does not give contribution to
the energy in any order of perturbation theory. We thus
need to consider the effect of the final term in Eq. (278),
the diamagnetic term.
If we imagine that B is applied in the ẑ direction, the
expectation of the diamagnetic term can be written as
e2 X D 2
E e2 B 2 X
x2a + ya2 .

∆E = |B × ra | =
8m a 8m a FIG. 148. Lev Davidovich Landau (1908 – 1968) was a
(293) Soviet physicist of Jewish origin who made fundamental con-
tributions to many areas of theoretical physics. His accom-
Using the fact that the atom is rotationally symmetric, plishments include the quantum mechanical theory of dia-
magnetism, the theory of second-order phase transitions, the
we can write
Ginzburg - Landau theory of superconductivity, the theory of
X 2X Fermi liquid. He received the 1962 Nobel Prize in Physics for
< x2a + ya2 >= < ra2 >, (294) his development of a theory of superfluidity.
a
3 a
so we have
e2 B 2 X Conduction electrons in a metal are responsible for the
∆E = < ra2 > . (295) so-called Landau diamagnetism
12m a
Problem G.1 Try to understand qualitatively why the 1

χLandau = − χP auli (298)
contribution due to Van Vleck paramagnetism (if it exists 3
in a given atom) always exceeds the contribution due to
Larmor diamagnetism.
which combines with the Pauli paramagnetism to reduce
Thus the magnetic moment of the atom is the total paramagnetism of the conduction electrons by
" # 1/3. Both Larmor and Landau diamagnetism are tem-
dE e2 X perature independent.
2
moment = − =− < ra > B. (296)
dB 6m a Larmor diamagnetism is often strong enough to over-
whelm Pauli paramagnetism in metals (corrected by the
Here the magnetic moment is proportional to the area Landau effect). This is particularly true in heavy ele-
< r2 > enclosed by the orbit of the electron. Note that ments where there are many core electrons that can con-
the magnetization we get for a current loop is also pro- tribute to the diamagnetism). If there are free spins in
portional to the area of the loop. This reminds us of the material, then Curie paramagnetism occurs which is
Ampere’s molecular currents hypothesis. always stronger than any diamagnetism.
The result described by Eq. (330) is known as Larmor
diamagnetism. For most atoms, < r2 > is on the order of
a few Bohr radii squared. Thus for noble gases we obtain Problem G.2 Analyse to what extend Eq. (297) corre-
sponds to the right column of the Table 149.
µ0 M Zne2 µ0 < r2 >
χLarmor = =− , (297)
B 6m
A substance that is diamagnetic repels a magnetic
where µ0 = 4π × 10−7 N/A2 is the vacuum permeability, field. This repulsion can be used to levitate very light
n is the the concentration of atoms, and Z is the number pieces of pyrolytic graphite or bismuth above a moder-
of electrons per atom. ately strong permanent magnet. As water is predomi-
At low temperature, noble gas atoms form very weakly nantly diamagnetic, this technique has been used to lev-
bonded crystals (with the exception of the case of helium itate water droplets and even live animals, such as a
which does not crystalize but rather forms a superfluid grasshopper, frog and a mouse. Paramagnets can not lev-
at low temperature), and the same calculation continues itate, because they are attracted to field maxima, which
to apply . In fact, for any material, diamagnetism of do not exist in free space. Diamagnets (which induce
electrons in core orbitals is described by Eq. (297) usually a negative moment) are attracted to field minima, and
fairly accurately. there can be a field minimum in free space.
62
FIG. 151. James Clerk Maxwell (1831 - 1879) was a Scot-

tish theoretical physicist and mathematician. His most signifi-
FIG. 149. Magnetic susceptibilities. Diamagnetism is also cant achievement was the development of the classical electro-
observed in some ionic solids, for example, in NaCl. magnetic theory. His set of equations – Maxwell’s equations
– demonstrated that electricity, magnetism and even light are
all manifestations of the same phenomenon: the electromag-
netic field.
to electric field and magnetization is proportional to mag-

netic field
P = χe E
M = χm H, (304)
where χe is electric susceptibility and χm is magnetic
susceptibility. Thus Eqs. (303) can be written as
D = E
B = µH, (305)
FIG. 150. A live frog levitates inside a 32 mm (1.26 in) di-
ameter vertical bore of a Bitter solenoid in a magnetic field where
of about 16 Teslas at the Nijmegen High Field Magnet Lab-
oratory. = 1 + 4πχe
µ = H + 4πχm . (306)
XVI. DIELECTRICS AND FERROELECTRICS

B. Dielectric Constant and Polarizability
A. Electromagnetic Fields in Material Medium
Electric susceptibility can be related to the polarizabil-
Maxwell equations in a material medium are ity of a single atom (molecule) α
1 p = αE, (307)
∇ × E = − Ḃ (299)
c
where p is the dipole moment of the atom (molecule).
∇·B=0 (300)
For example, for hydrogen atom in a ground state
∇ · D = 4πρ (301)
9 3
1 4π α= a . (308)
∇ × H = Ḋ + j, (302) 2 B
c c
In dilute medium the relation is
where
χe = N α, (309)
D = E + 4πP
where N is the concentration of atoms (molecules).
B = H + 4πM; (303) For most liquids which are not too complicated in
structure, we could expect that an atom finds itself, on
P is the polarization of the material (the electric dipole
the average, surrounded by other atoms in what would
moment per unit volume), measured in coulombs per me-
be a good approximation to a spherical hole. If we call
ter square, and M is the magnetization of the material
E the field in the uniform dielectric, we can write
(the magnetic dipole moment per unit volume), measured
in amperes per meter. Often polarization is proportional E = Ehole + Eplug , (310)
63
C. Ferroelectric Crystals
If we imagine a lattice of unit cells of barium titanate

like that on Fig. 61, it is possible to pick up chains of
ions along vertical lines. Let the dipole moment of each
atom is p. Hence the field created at the position of any
given dipole by its neares neighbor is
2p
E= . (319)
r3
At any given atom, the dipoles at equal distances above
and below it give fields in the same direction, so for the
field created by the whole chain we get

2p 2 2 2 4.81p
E = 3 2+ + + + ... = .
a 8 27 64 a3
(320)
FIG. 152. The electric field of a uniformly polarized sphere.
It is possible to show that if our model were like a com-
pletely cubic crystal, the number 4.81 would be changed
where Ehole is the field in the hole, and Eplug is the field to 4π/3 = 4.19. In other words, if the next lines were
inside a sphere which is uniformly polarized. The fields at a distance a they would contribute −15% to our sum.
due to a uniformly polarized sphere are shown on Fig. However the next main chain we are considering is at the
152. The electric field inside the sphere is uniform and distance 2a, and the field from a periodic structure dies
its value is off exponentially with the distance.
4π So we have two equations: (307) and (320). There are
Eplug = − P. (311)
3 two solutions: E and p both zero, or
Using Eq. (310) we obtain
a3
4π α= , (321)
Ehole = E + P. (312) 4.81
3
Hence with E and p both finite. For BaTiO3 the spacing a
is 2 × 10−8 cm. We can compare this with the known
4π polaizabilities of individual atoms.
P = Nα E + P . (313)
3 Barium titanate exists in one of five polymorphs de-
pending on temperature. From high to low temperature,
N αE
P= . (314) the crystal symmetries of the five polymorphs are hexag-
1 − 4π
3 N αP onal, cubic, tetragonal, orthorhombic, and rhombohedral
Thus we get Clausius-Mossotti relation crystal structure. All of these phases exhibit the ferro-
r − 1 4π electric effect apart from the cubic phase.
= N α. (315)
r + 2 3
Let us see what happens when N grows (for example, D. Polaritons
because of decrease of the temperature) in such a way
3
that N α approaches 4π . We call Tc the critical tem- The coupling of the electric field E of the photon with
3
perature at which N α is exactly 4π . Near the critical the dielectric polarization P of the TO phonon is de-
temperature scribed by the electromagnetic wave equation:
3
Nα = − β(T − Tc ), (316) c2 k 2 E = ω 2 (E + 4πP ) . (322)
4π
where β is a constant of the same order of magnitude as At low wavevectors the TO phonon frequency ωT is in-
the thermal expansion coefficient, or about 10−6 to 10−5 dependent of k. The polarization is proportional to the
K−1 . Substituting (316) into (315) we obtain displacement of the positive ions relative to the negative
3 − 4πβ(T − Tc ) ions, so that the equation of motion of the polarization
r − 1 = . (317) is like that of an oscillator and may be written as
4πβ(T − Tc )/3
Since we have assumed that β(T − Tc ) is small compared −ω 2 P + ωT2 P = (N q 2 /M )E, (323)
with one, we can approximate this formula by where there are N ion pairs of effective charge q and
9 reduced mass M , per unit volume. For simplicity we
r − 1 = . (318)
4πβ(T − Tc ) neglect the electronic contribution to the polarization.
64
FIG. 154. BaTiO3 dielectric permittivity.
FIG. 155. Displacement of Ti atoms in ferroelectric phases.
in accord with the definition of (∞) as the optical di-

electric constant, obtained as the square of the optical
refractive index. We set ω = 0 to obtain the static di-
electric function:
(0) = (∞) + 4πN q 2 /M ωT2 , (327)
which is combined with (325) to obtain (ω) in terms of
accessible parameters:
2
ωL − ω2
(ω) = (∞) 2 . (328)
ωT − ω 2
The zero of (ω) defines the longitudinal optical phonon
FIG. 153. Ion polarizability data. frequency ωL , as the pole of (ω) defines ωT . The zero
gives
2
(∞)ωL = (0)ωT2 . (329)
The equations (322) and (323) have a solution when
2
ω − c2 k 2 4πω 2
N q 2 M ω 2 − ωT2 = 0 (324)

XVII. LANDAU THEORY OF PHASE
TRANSITIONS
The dielectric function obtained from (323) is
4πN q 2 /M We can obtain a consistent formal thermodynamic the-
(ω) = 1 + . (325) ory of the behavior of a ferroelectric crystal by consider-
ωT2 − ω 2
ing the form of the expansion of the energy as a function
If there is an optical electronic contribution to the po- of the polarization P . We assume that the Landau7 free
larization from the ion cores, this should be included. In energy density in one dimension may be expanded for-
the frequency range from zero up through the infrared, mally as
we write 1 1 1
F̂ (P, T, E) = −EP + g0 + g2 P 2 + g4 P 4 + g6 P 6 + . . . ,
4πN q 2 /M 2 4 6
(ω) = (∞) + 2 . (326)
ωT − ω 2 (330)
65
zero at some temperature T0 :

g2 = γ(T − T0 ), (332)
where γ is taken as a positive constant and T0 may be
equal to or lower than the transition temperature. A
small positive value of g2 means that the lattice is soft
and is close to instability. A negative value of g2 means
that the unpolarized lattice is unstable. The variation of
g2 with temperature is accounted for by thermal expan-
sion and other effects of anharmonic lattice interactions.
A. Second-Order Transition
If g4 in (330) is positive, nothing new is added by the

term in g6 , and this may then be neglected. The polar-
FIG. 156. Spontaneous polarization projected on cube edge ization for zero applied electric field is found from (331):
of barium titanate, as a function of temperature.
γ(T − T0 )Ps + g4 Ps3 = 0, (333)
so that either Ps = 0, or Ps2 = (γ/g4 )(T0 − T ). For
T > T0 the only real root of (333) is at Ps = 0, because
γ and g4 are positive. Thus T0 is the Curie temperature.
For T < T0 the minimum of the Landau free energy in
zero applied field is at
|Ps | = (γ/g4 )1/2 (T0 − T )1/2 . (334)
The phase transition is a second-order transition because
the polarization goes continuously to zero at the transi-
tion temperature.
Problem A.1 Dielectric constant below transition
temperature. In terms of the parameters in the Lan-
dau free energy expansion, show that for a second-order
phase transition the dielectric constant below the transi-
tion temperature is
4π∆P
=1+ = 1 + 2π/γ(Tc − T ). (335)
E
This result may be compared with (338) above the transi-
FIG. 157. A plot of the observed energies and wavevectors of
tion.
the polaritons and of the LO phonons in GaP. The theoretical
dispersion curves are shown by the solid lines. The dispersion
curves for the uncoupled phonons and photons are shown by
the short, dashed lines. B. First-Order Transition
The transition is first order if g4 in (330) is negative.

where the coefficients gn depend on the temperature. We must now retain g6 and take it positive in order to
The value of P in thermal equilibrium is given by the restrain from going to minus infinity. The equilibrium
minimum of F̂ as a function of P ; the value of F̂ at condition for E = 0 is given by (331):
this minimum defines the Helmholtz free energy F (T, E).
γ(T − T0 )Ps − |g4 |Ps3 + g6 Ps5 = 0, (336)
The equilibrium polarization in an applied electric field
E satisfies the extremum condition so that either Ps = 0 or
∂ F̂ γ(T − T0 ) − |g4 |Ps2 + g6 Ps4 = 0. (337)
= 0 = −EP + g2 P 2 + g4 P 4 + g6 P 6 + . . . .(331)
∂P
The dielectric constant is calculated from the equilib-
To obtain a ferroelectric state we must suppose that rium polarization in an applied electric field E and is
the coefficient of the term in P 2 in (330) passes through found from (331). In equilibrium at temperatures over
66
the transition, the terms in P4 and P6 may be neglected;

thus E = γ(T − T0 )P , or
(T > Tc ) = 1 + 4πP/E = 1 + 4π/γ(T − T0 ). (338)

FIG. 159.
The result applies whether the transition is of the first or
second order, but if second order we haveT0 = Tc ; if first
order, then T0 < Tc . Equation (332) defines T0 , but Tc A. Rules for Primary Solubility
is the transition temperature.
The conditions favoring primary solubility were stud-
Problem B.1 Saturation polarization at Curie
ied carefully by Hume-Rothery and coworkers, whose re-
point. In a first-order transition the equilibrium con-
sults are summarized by the following four rules.
dition (337) with T set equal to Tc gives one equation for
a) Atomic size eflect. The solute and solvent atoms
the polarization Ps (Tc ). A further condition at the Curie
should be close in size. The difference in diameter of
point is that F̂ (Ps , Tc ) = F̂ (0, Tc ). (a) Combining these atoms should not exceed 15%. For silver and gold, the
two conditions, show that Ps2 (Tc ) = 3|g4 |/4g6 . (b) Using difference is only 0.2%.
this result, show that Tc = T0 + 3g42 /16γg6 . b) Crystal structure effect. In order for there to be ex-
tensive solubility, the structures of the solute and solvent
metals should be similar. Both silver and gold, for ex-
ample, have an fcc structure.
c) Electronegative valence effect. The two elements must
have similar electrochemical characteristics. By contrast,
an electropositive element such as silver and an elec-
tronegative element such as bromine would form a chem-
ical compound, not an alloy.
d) Relative valence effect. This rule asserts that it is
easier to dissolve a metal of higher valence into one of
lower valence than the reverse. For instance, aluminum
dissolves more readily in copper than copper in alu-
minum because, apparently, in the former situation it
is relatively easy for the excess aluminum electrons to
detach themselves from their own atoms and accommo-
date themselves in the alloy. If copper is dissolved in
FIG. 158. Landau free energy function versus (polarization)2
in a first-order transition, at representative temperatures. At aluminum, however, there is a deficiency of conduction
Tc the Landau function has equal minima at P = 0 and at electrons at the copper sites, and the electrons that tend
a finite P as shown. For T below Tc the absolute minimum to neutralize this deficiency have high energy.
is at larger values of P ; as T passes through Tc there is a
discontinuous change in the position of the absolute minimum. Problem A.1 The atomic size factor favors solid solu-
The arrows mark the minima. bility for the alloys presented on Fig. 159. What is the
effect of the relative valency factor in each case?
B. The Phase Diagram

XVIII. METALLIC ALLOYS
The composition of the solid and liquid phases in the
Metals are rarely used in their pure form in industrial region between the solidus and liquidus lines can be eval-
applications because pure metals, among other draw- uated from the phase diagram. Because mass is con-
backs, are too soft and ductile. Thus carbon and other served, we may write
metals are added to iron to harden it to steel, and alu-
minum is strengthened by adding copper, silicon, and
other elements to it. Brass is an alloy of copper and zinc. c(S + L) = cS S + cL L, (339)
ln all these cases one deals with a metallic alloy in which
atoms of one or more elements are dissolved in a metal. where S and L are the amounts of solid and liquid. By
An alloy is therefore a solid solutior. It differs from a rearranging, we can write this equation as
chemical compound in that, in a solid solution, the range L cS − c
of concentration of the solute relative to the solvent may = , (340)
S c − cL
vary, while in a chemical compound this concentration is
fixed. which is known as the lever formula.
67
where p is the number of microstales corresponding to

the same macrostate of the system. Texts on thermody-
namics and statistical mechanics show that the two defi-
nitions for entropy are entirely equivalent, and therefore
they will be used interchangeably here.
D. Polymorphic Transformation
When a metal or alloy is heated, at some temperature

it undergoes a transformation to a new crystal structure
(or solid phase). This happens most frequently in the
transition metals and their alloys. A well-known example
is iron which, when heated to 910◦ C, makes a transition
from a bcc (α-iron) to an fcc (β-iron) structure. Other
transition metals show similar polymorphic transforma-
FIG. 160. Phase diagram for a binary alloy A-B. tions. The phenomenon can be understood in terms of
the free-energy principle.
Using Eq. (341), one can show that the free energy at
temperature T is given by the expression
Z T Z T !
CP (T 0 ) 0
F = E0 − T S0 − dT dT, (343)
0 0 T0
FIG. 161.
where E0 and S0 are the internal energy and entropy at
absolute zero, respectively.
Problem B.1 a) Construct the phase diagram for for Problem D.1 Establish the validity of Eq. 343) for the
the Cu-Ni alloy, using the data presented on Fig. 161. free energy.
b) Starting with a liquid alloy of 60% Ni and cooling it
gradually, state the composition of the solid that forms Let us compare two possible but different structures
first. for the system, A and B (Fig. 162). The structure A
c) How much solid per kilogram can be extracted from the has a lower E0 , and hence is more stable at low tem-
melt at 1300◦ C? perature than B. Since A is more tightly bound, it also
has a higher Einstein, or Debye, temperature, and con-
sequently a rather low specific heat Cp . It follows from
C. Thermodynamics: Free Energy and Entropy (343) that as the temperature increases, the free energy
FA , decreases at a lower rate than FB , and hence the
To gain a deeper understanding of phase diagrams curves for FA and FB , versus T will intersect at some
and related effects, we need the concept of free energy. temperature Tc , as shown in Fig. 162. Below Tc , FA , FB ,
ln thermodynamics there are two different free energies: and A is the more stable of the two structures, while
The Helmholtz energy, F = E − T S, and the Gibbs en- above Tc , the situation is reversed. Of course, the trans-
ergy, G = F + P V , where the symbols on the right sides formation is observed only if the transition temperature
of these equations have the usual meaning. Most exper- is below the melting point;otherwise the solid would melt
iments are performed at atmospheric pressure, which is before it had a chance to undergo the polymorphic trans-
so low that the P V term may be neglected, compared formation.
with the other terms, without serious error. Therefore Figure 162 can also be used to describe the melting
it is sufficient for our purpose to use the Helmholtz free transition of a metal, where A and B then refer to the
energy. A well-known principle in thermodynamics-the solid and liquid phases, respectively.
principle of minimum free energy-asserts that if a system
is allowed several alternative states, it will choose the one
E. The Mixing Entropy of a Substitutional Alloy
with the lowest free energy.
The thermodynamic definition of entropy is
When two metals at the same temperature are mixed
δQ CP dT to form an alloy, the entropy of the system increases by
dS = = . (341)
T T virtue of the mixing process. This increase is called the
In statistical mechanics, the entropy of a system is de- entropy of mixing, or entropy of disorder. We expect the
fined as increase intuitively because the system becomes more dis-
ordered. We shall now calculate the increase in entropy
S = kB ln p, (342) using the statistical definition.
68
free energies of the two phases at the same temperature

by writing
∆F = EL − ES − T (SL − SS ) = ∆E − T ∆S. (346)
The energy difference ∆E is positive because in the liquid

phase many of the atoms occupy interstitial positions,
which results in a high energy. The volume is also larger
in the liquid than in the solid phase, so that the atoms
are pulled away from each other with some expenditure
of energy. However, ∆S is also positive, because the
liquid phase, being more disordered than the solid, has
FIG. 162. Polymorphic transformation. Free energy F versus a higher entropy. For T < Tm , where Tm is the melting
T , for a system in two different solid phases, A and B. temperature, the term ∆E dominates, that is, ∆F >
0, and consequently no melting takes place, while for
T > Tm the entropy term dominates, and the solid melts
completely. At T = Tm , the energy and entropy terms
exactly balance each other, ∆F = 0, and the two phases
are in equilibrium with each other.
G. Free Energy of Alloy Phases
Suppose that the free energy of the alloy in its homo-

geneous solid phase is given by the solid U -shaped curve
of Fig. 164. We also suppose that in this phase the al-
FIG. 163. The mixing entropy of a substitutional alloy S loy is a primary solid solution throughout the concentra-
versus concentration c. The entropy has a maximum at c = tion range. We shall show that the homogeneous-solution
0.5, whose value is 1.4 cal/mole. phase is more stable than any other structure.
Referring to Fig. 164, note that at composition c
the free energy for the homogeneous-solution phase is F.
Compare this with another possibility, namely, that the
system breaks up into two coexisting solid phases, one
N! of concentration c0 and the other of concentration c00 . A
p= (344)
n!(N − n)! state of this type is called a phase mixture.
Using Stirling formula we obtain Problem G.1 Show that the free energy for a phase mix-
S = −N kB [c ln c + (1 − c) ln(1 − c)] (345) ture of components c0 and c00 varies with concentration
along the straight line F 0 F 00 as the concentration in-
The variation of entropy with composition is indicated creases from c0 to c00 .
in Fig. 163, where S has a maximum at c = 0.5, the
point of maximum disorder, and decreases on either side Therefore at concentration c the free energy of the phase
of this point; it reaches zero at the endpoints at which, in mixture is F1 . Since the free energy of the homogeneous
each case, a state of complete order prevails. Numerically phase F is less than that of the phase mixture F1 , the
the maximum entropy is 1.4 cal/mole. lt is important to former is the more stable structure. By choosing different
note that near the endpoints S increases very rapidly as c0 and c00 , one can change the energy F1 , but, for the type
the other element is added. This means that there is of free-energy curve of Fig. 164(a), one cannot make it
a strong tendency toward solution at low concentration, less than F . Therefore the homogeneous-single phase
regardless of other possibly unfavorable factors. is the stable structure. Examples of systems with free-
energy curves resembling this figure are the Ag-Au and
Cu-Ni alloys.
F. Melting and Structure
Why does a solid melt when it is heated, and why does 1. Microscopic Model
this take place at a certain fixed temperature, different
for each substance? The answer must be that above this Let us calculate the free energy of a substitutional solid
temperature the free energy of the liquid phase becomes solution, using a simple atomic model, and compare the
lower than that of the solid phase. We may compare the results with the free-energy curves we have discussed.
69
This expression now has to be inserted in (347), and the

result plotted versus c. You can verify that only curves of
the types shown in Fig. 164 are obtained. More specif-
ically, the U -shaped curve of Fig. 164(a) is obtained
FIG. 164. (a) A free-energy-versus-concentration diagram FIG. 165. (a) Free energies of solid and liquid phases of an
leading to a stable, homogeneous phase. (b) A free-energy- alloy below its melting range. (b) Free energies of solid and
versus-concentration diagram leading to a stable phase mix- liquid phases within the melting range of the alloy. (c) Free
ture in the concentration range c0 < c < c00 . energies of the two phases above melting point.
The free energy for the solution is

Z T T
CP (T 0 ) 0
Z
F = E − T S = E0 + CP dT − T dT when VAB ≤ (VAA + VBB )/2, while the type shown in
0 0 T0 Fig. 164(b) is obtained when VAB (VAA + VBB )/2 Thus
+ N kB [c ln c + (1 − c) ln(1 − c)], (347) the latter type holds true when the attraction between
the different atoms is less strong than the average attrac-
where the various terms of energy and entropy mean the tion between similar atoms. From this point of view, you
following: E0 is the energy at absolute zero, the first inte- can see why, in this case, like atoms prefer to segregate
gral is the increase in thermal energy, the second integral into two separate phases, as we discovered previously.
results from the thermal entropy [see (341)], and the last There is a range of primary solubility near the endpoints
term is the mixing entropy (345). We note that if the because the mixing entropy there increases very rapidly,
two types of atoms are not dissimilar, then the integral forcing a certain amount of solubility, limited though it
terms are insensitive to compositional changes, and may may be.
be ignored if we are interested only in the shape of the
curve F versus c.
We can calculate E0 as follows: If we call the energy
of an A-A bond VAA , then the total energy of the A-A
bonds in the whole crystal is
1 1
N (1 − c)Z(1 − c)VAA = N Z(1 − c)2 VAA , (348)
2 2
where Z is the coordination number of the crystal struc- H. The Phase Diagram and Free Energy
ture.
The result for the total internal energy E0 is therefore
1 1 1 The concept of free energy leads readily to the phase
E0 = N Z(1 − c)2 VAA + N Zc2 VBB + N Zc(1 − c)VAB ,
2 2 2 diagram (Fig. 160) for a binary alloy. This can be seen
(349) from Fig. 165, in which we plot the free energy for a
completely miscible binary alloy at three progressively
where VBB and VAB , are the energies for a B-B and A-B higher temperatures T , T 0 , and T 00 (T < T 0 < T 00 ). The
bond, respectively. This equation may be recast in the diagrams show the energy for both the solid and liquid
following useful form: phases at each temperature. At temperature T the solid’s
1 curve lies entirely below that of the liquid curve, and
E0 = E0 = N Z [cVAA + (1 − c)VBB therefore the alloy is in the solid phase. However, at
2
VAA + VBB
some higher temperature T 0 , the liquid’s curve crosses
+2c(1 − c) VAB − . (350) the solid curve (Fig. 165b).
2
It is now useful to infer the phase diagram for a system whose free energy, for the solid solution, is given by Fig.
70
FIG. 166. (a),(b),(c) and(d): Free energies of solid and liquid phases for a solid solution described by Fig. 164(b), at increasingly
higher temperatures (T < T 0 < T 00 < T 000 ). (e) Phase diagram for the system. (f) Phase diagram of Ag-Cu system.
164(b). Figure 166 plots the free energies for the solid and known example of this type of system is the Cu-Ag alloy
liquid phases at four different temperatures T, T 0 , T ”, and shown in Fig. 166(f).
T ”0 , near the melting range at which T < T 0 < T 00 < T 000 .
In Fig. 166(a) the system is either a primary solution of Problem H.1 Confirm that the free-energy diagrams of
phase α, rich in A, or a solution in phase β, rich in B, Figs. 166(a)-166(d) lead to the phase diagram 166(f ).
or a phase mixture of α and β, depending on the con- Indicate on this latter figure suitable values for the tem-
centration as indicated above. No liquid phase appears peratures T, T 0 , T 00 , and T 000 indicated in the former fig-
because the free energy of the liquid phase is too high. ures.
At a higher temperature T 0 , shown in Fig. 166(b), a
situation obtains in which the tangents of the α and β Problem H.2 The phase diagram for the Cu-Ag alloy is
phases also touch the liquid curve, and this gives rise to shown in Fig. 166(f ).
several possibilities, depending on the concentration. A a) Confirm that the atomic % and weight % scales indi-
particularly interesting one occurs when the composition cated are consistent with each other.
is equal to ce ; here the three phases-α, β, and the liquid b) Determine the atomic percentage of the α-phase at the
phase-coexist. Such a composition is called the eutec- eutectic concentration just after solidification.
tic composition, and the corresponding temperature is c) Determine the percentage of the same phase at the
called the eutectic temperature. At still higher temper- temperature 850◦ C, and the Cu concentration in atomic
atures, the curves appear as shown in Figs. 166(c) and %.
(d). The phase diagram resulting from this situation is
shown in Fig. 166(e). A characteristic feature of such a
phase diagram is that elements A and B show only lim- I. Intermediate Phases
ited solid solubility in each other. They tend to segregate
into phase mixtures or turn into a liquid phase. A well- ln our discussion of solid solutions, we have so far as-
sumed that the solution has the same crystal structure
71
ting this energy term. However, in alloys involving Cu,

Ag, or Au with other metals of higher valence, the elec-
tron/atom ratio changes with composition, and this leads
to interesting effects on the crystal structure. It was first
observed by Hume-Rothery that the α-phase (fcc) of such
alloys as CuZn, CuAl and AgMg becomes unstable when
theelectron/atom ratio approaches the value 1.4, and a
complete transformation to the β-phase (bcc) takes place
when the ratio is near 1.5.
FIG. 167. (a) Intermediate phases of a solid solution. (b)

Intermediate phases of the Mg-Pb system.
throughout the entire composition range. However, some

other solid phases may have a low free energy at interme-
diate compositions. This possibility is illustrated in Fig.
167(a) for three different solid phases, α, β, and γ. Using
the rules developed for minimizing free energy, one can
determine the possible phase structure at various values FIG. 168. Density of states, g(E) versus energy E for an fcc
of composition. structure and a bcc structure. Dashed line represents free-
When the temperature is raised, the positions of the electron model; cross-hatched area represents region occupied
by electrons.
various intermediate phases may change relative to each
other. Eventually, when the temperature is sufficiently
high, melting starts. The phase diagram for this system
Since the critical factor in this type of transforma-
can be inferred from observing the evolution of the free-
tion is the concentration of electrons, we look for an ex-
energy diagram with temperature, and using the rules
planation of the above transition in terms of the band
of minimization. The phase diagram of the Mg-Pb alloy
structure. Figure 168 shows the density of states for the
of Fig. 167(b) is a typical result. This diagram is more
free-electron model, as well as for the fcc and bcc struc-
complex than Fig. 166(e), and in fact may be viewed as
tures. The density-of-states function deviates from the
a set of two eutectic diagrams joined together. In most
free-electron value appreciably only when the Fermi sur-
practical alloys in which there are several elements and
face begins to touch the Brillouin zone, which occurs at
several intermediate phases, the phase diagram is very
the energies E 0 and E 00 for the fcc and bcc structures,
complex indeed.
respectively.
As the concentration of electrons increases, starting in
J. Electron Concentration and the Zone Theory of the fcc structure, EF also increases, until the energy E 0
Alloy Phases is reached. Beyond this point, any increase in electrons
would lead to a rapid increase in E0 [because g(E) of
curve I decreases rapidly], and hence to a rapid increase
ln our discussion of structural properties of alloys, the
in the energy of the electrons. To obviate this increase in
conduction electrons have so far played no role. The
energy, the system begins to transform partially into the
energy of these electrons should, in fact, be added to
bcc phase and thus lessens its energy, the transformation
the internal energy E0 to arrive at the total internal en-
being completed at E 00 .
ergy, but in an alloy in which the two elements are of the
same valence, such as the Ag-Au or Cu-Ag systems, the Problem J.1 Show that according to the free-electron
electron/atom ratio remains unchanged as the composi- model, the points E 0 and E 00 occur at electron/atom ra-
tion is varied, and consequently the energy ofconduction tios of 1.36 and 1.48, respectively, in close agreement with
electrons remains essentially unaffected throughout the the observed values.
composition range. This is why we were justified in omit-
View publication stats

Solid State Physics2 PDF

Uploaded by

Copyright:

Available Formats

Solid State Physics2 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Solid State Physics2 PDF

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

SOLID STATE PHYSICS Course by E. Kogan

Presentation · August 2018

Magnetic Resonances View project

Critical Phenomena View project

The user has requested enhancement of the downloaded file.

CONTENTS 1. Molecular Orbitals and Valence Bonds 28

VII. CHEMICAL ELEMENTS 20 XIV. SEMICONDUCTORS 51

E. Iron Group Ions 59 importantly, however, the physics of solids provides a

XVI. DIELECTRICS AND FERROELECTRICS 62

XVII. LANDAU THEORY OF PHASE

XVIII. METALLIC ALLOYS 66

specific heat of sodium chloride at room temperature is

TABLE I. Heat capacities Cv of some solids at room temper-

FIG. 2. Temperature variation of Cp and Cv in sodium chlo-

in three dimensions we get Hence

FIG. 6. Albert Einstein (1879 - 1955) was an ethni-

FIG. 7. Max Karl Ernst Ludwig Planck (1858-1947)

In three dimensions the space dependence is eik·r .

is called Riemann ζ function. In general

Expanding both sides we get

Problem D.1 Show that

Eq. (29) can be presented as 12π 4 T 3

FIG. 10. Potential between neighboring atoms (thin black).

FIG. 9. Heat capacity of silver compared to the Debye and

Problem D.3 Einstein versus Debye. In both the 1 ∂L 1

a) Why is this expansion of the exponent and the exten-

c) In what temperature range is the above expansion

B. Vibrations of A One-Dimensional Chain

For small oscillations we can ignore the higher order

The equation of motion is

FIG. 13. Paul Adrien Maurice Dirac (1902 - 1984) was a

Operator â is not Hermitian; Solving Eqs. (62) and (63) we obtain

N̂ â† |n >= (n + 1)â† |n >; (72) Fourier transforming the variables

we obtain the dispersion relation

M1 M2 ω 4 − 2κ(M1 + M2 )ω 2 + 2κ2 (1 − cos ka) = 0.

FIG. 15. The one-dimensional harmonic chain with diatomic

Problem D.2 Using Eq. (84) analyse behavior of the

Consider a chain with two atoms in the primitive cell.

tron returns to momentum p = 0.

Consider Drude equation in the absence of magnetic

Problem A.1 Plot real σ1 (ω) and imaginary part σ2 (ω)

IV. ELECTRONS IN METALS: DRUDE m

FIG. 20. Edwin Hall’s 1879 experiment. The voltage mea-

V. SOMMERFELD (FREE ELECTRONS)

B. Electronic Heat Capacity

The grand canonic potential is

we get of T /TF ∼ 0.01, we can return to re-examine some of the

Problem B.1 (a) Show that the kinetic energy of a free

Both M and H are measured in amperes per meter; χ is

FIG. 30. Filling of electronic states up to the Fermi energy.

Another property we can examine about the free elec-

A crystal or crystalline solid is a solid material whose

FIG. 32. Some unit cells for the hexagonal lattice.

A lattice system is a class of lattices with the same

FIG. 34. The 14 Bravais lattices

The diamond structure is the structure of the semicon-

There is another crystal structure, the hexagonal close

C. Allotropy ical elements to exist in two or more different forms, in