Lattice specific heat of solids: summary

𝛿𝑄 𝑑𝑈'())*+,
𝐶= 𝐶)<)(' = 𝐶'())*+, + 𝐶,',+)><?*+ + 𝐶@(A?,)*+ + ⋯ other degrees of freedom, if any 𝐶'())*+, =
𝛿𝑇 𝑑𝑇
𝐶(Dulong − Petit) = 3R (from classical Kinetic theory of gas)
1
𝑥f𝑒 Q ℏ𝜔 τ = 𝑘X 𝑇 ⟨𝑛 ⟩ =
𝐶'())*+, = 𝑘X ~ [ 𝑑𝜔𝐷 𝜔 𝑥= 𝑒Q − 1
(𝑒 Q −1)2 𝜏
•

a Debye temperature
Z Q_ `
𝑇 b 𝑥]𝑒 Q Z
𝐶R,ST, = 9𝑁𝑘X [ 𝑑𝑥 Q 𝑇 More suitable for acoustic phonon
𝜃 c (𝑒 −1)2 𝐶R,ST, (T< 𝜃/50) ≅ 234𝑁𝑘X
𝜃 spectrum activated at low temperatures.
Debye T3 law

f gh More suitable for optical phonon spectrum

𝐶d*?e),*? = 𝑁𝑘X 𝑥
(g h ij)k activated at high temperatures.
Van Hove singularities

q uvw
𝐷 𝜔 𝑑𝜔 = ∫ , for general (actual) case
(fr)s xy

Debye solid Actual crystal structure

 Band theory of solids

Free electron model

Nearly free electron model
Kronig-Penny model
 Free electron model

• Free electron Fermi gas or gas of free electrons subject to the Pauli exclusion principle: Ideally non-interacting electrons as
the atoms in ideal gas
• Applicable to metallic solids.

Conduction electrons in metals can be considered as non-interacting or ‘free’, because

(1) quantum mechanical treatment shows that the motion of an electrons is not deflected by an array of a perfectly
periodic lattice,
(2) conduction electrons rarely collide with other free electrons due to Pauli exclusion principle.

Drude model describes free electron Fermi gas.

• Kinetic energy of the conduction band electron mainly decides electron transport properties in free electron model. à
Electrical and thermal conductivity, Ohm’s law
 Energy levels in one dimension

Schrödinger equation 𝐻𝜓 = 𝐸𝜓
•k j u u ℏk u k
For a single free electron, 𝐻 ≡ ≡ −𝑖ℏ −𝑖ℏ =−
f„ f„ uQ uQ f„ uQ k

ℏf 𝑑f 𝜓‡
− = 𝐸‡ 𝜓 n denotes index of solutions or orbitals
2𝑚 𝑑𝑥 f

Boundary conditions: 𝜓 𝑥 = 0 = 0; 𝜓 𝑥 = 𝐿 = 0

Solutions: 𝜓‡ (𝑥) = Asin(𝑘‡ 𝑥) ℏf 𝑘‡f ‡r

𝜖‡ = With 𝑘‡ =
2𝑚 Š

ℏk •Žk ℏk •kk ℏk •sk

𝜖j = , 𝜖f = , 𝜖f = ,………..
f„ f„ f„
1
There is a maximum 𝑛• , because no. of electrons in a solid is limited. The corresponding 𝑘• is called 𝑛𝜆 = 𝐿
Fermi wave vector and the corresponding energy is called 𝜖• , or Fermi energy. 2
In other words, the Fermi energy is defined as the energy of the topmost filled level in the ground 2𝜋
state of the N electron system. 𝑘=
𝜆
Note that each energy level can occupy 2 electrons with opposite spins (spin-degeneracy).
f
ℏf 𝑘•f ℏf 𝑛• 𝜋 f ℏf 𝑁𝜋
𝜖• = = = The orbital model is exact if there are no interactions between electrons.
2𝑚 2𝑚 𝐿 2𝑚 2𝐿
 Fermi Dirac distribution
Ground state ⟷ situation at absolute zero temperature
According to statistical mechanics, electrons occupy in orbitals as per the Fermi-Dirac distribution at non-zero temperatures.

1
Probability of electron occupation at temperature T 𝑓 𝜖 = (”i•)
—j
𝑒 •– b

𝜇 is the chemical potential. 𝜇 (T=0) = 𝜖•

At all temperatures, 𝑓 𝜖 =1/2 when ϵ = 𝜇

(”i•)
At high energy tail of the distribution, ≫1, the FD
•– b
Distribution takes form of Maxwell or Boltzmann distribution.

(”i•) MB limit
i
𝑓 𝜖≫𝜇 =𝑒 •– b

Maxwell - Boltzmann limit

 Free electron Fermi Gas in 3D
ℏf f 𝜕 f
𝜕 f
𝜕 f
− 𝛻 𝜓 = 𝐸‡ 𝜓 𝛻f ≡ f + f + f
2𝑚 𝜕𝑥 𝜕𝑦 𝜕𝑧

‡h fr ‡ª fr ‡« fr
𝜓• (𝑟⃗ ) = 𝑒 ¡•.£⃗ 𝑘 = 𝑘Q 𝚤̂ + 𝑘¦ 𝚥̂+𝑘¨ 𝑘© 𝑘Q = , 𝑘¦ = , 𝑘¨ =
Š Š Š
Where 𝑛Q , 𝑛¦ , 𝑛¨ are integers
ℏf 𝑘 f ℏf f 𝑝⃗ = −𝑖ℏ 𝛻
𝜖= = (𝑘Q + 𝑘¦f + 𝑘¨f )
2𝑚 2𝑚 𝑝⃗𝜓• (𝑟⃗ ) = −𝑖ℏ 𝛻𝜓• 𝑟⃗ = ℏ𝑘𝜓• 𝑟⃗

ℏ•
The particle velocity in orbital labelled by 𝑘 is given by, 𝑣⃗ =
„

ℏf 𝑘•f
Fermi energy in 3D 𝜖• = Fermi surface in free electron model is spherical.
2𝑚

𝑘 space is discretized at an interval of 2𝜋/𝐿. At 𝜖• , the orbitals exhaust N electrons.

4 Z Fermi wave vector Fermi energy Fermi velocity Fermi temperature

𝜋𝑘• 𝑉 Z
2. 3 = 𝑘• = 𝑁 j/Z f/Z j/Z 𝜖•
3𝜋 f 𝑁 ℏf 3𝜋 f 𝑁 ℏ 3𝜋 f 𝑁
2𝜋 Z 3𝜋 f 𝑘• = 𝜖• = 𝑣• = 𝑇• =
𝐿 𝑉 2𝑚 𝑉 𝑚 𝑉 𝑘X
Just an energy scale,
𝑘• , 𝑣• and 𝜖• depend only on electron concentration (N/V) in a metal. NOT actual temperature.
𝑇• ~ 50,000 K
 Electron density of states in 3D

Number of orbitals per unit energy range is called density of states, 𝐷 𝜖 . It is density of one particle states or density
of orbitals in energy space. 𝜖
𝐷(𝜖) •

Number of orbitals (or electronic states) N filled up to energy 𝜖 are related by,

Z/f
𝑉 2𝑚𝜖
𝑁= f
3𝜋 ℏf 𝜖
𝑓(𝜖, 𝑇 = 0) 𝑓(𝜖, 𝑇 ≠ 0)

u° q f„ Z/f Z j/f 𝜇
𝐷 𝜖 ≡ = 𝜖
u” Zrk ℏk f

q f„ Z/f j/f 𝜖 𝜖
𝐷 𝜖 = 𝜖
frk ℏk 𝑁(𝜖)=𝑓(𝜖)𝐷 𝜖 𝑁(𝜖, 𝑇)=𝑓(𝜖, 𝑇)𝐷 𝜖

3𝑁
𝐷 𝜖 =
2𝜖

𝜖 𝜖
Density of occupied electrons at finite temperature = 𝑓(𝜖, 𝑇)𝐷 𝜖 𝑘X 𝑇
 Heat capacity of the electron gas or electronic heat capacity

𝐶)<)(' = 𝐶'())*+, + 𝐶,',+)><?*+ + 𝐶@(A?,)*+ + ⋯ other degrees of freedom, if any 𝑑𝑈,'

𝐶,' =
𝑑𝑇

Pauli exclusion principle and the Fermi distribution function do not allow all electron excitation but only those lying in close
vicinity to 𝜖• . In other words, the electrons buried deep inside (with ϵ ≪ 𝜖• ) do not have available states to get excited; and
hence do not contribute to heat capacity.
𝑑𝑈,³ 𝑁(𝜖, 𝑇)
𝑈g³ ≈ 𝑁(𝑇/𝑇• )𝑘X 𝑇 𝐶,' = ≈ 𝑁𝑘X (𝑇/𝑇• ) Experimentally, 𝐶,' ~ T
𝑑𝑇 𝜖• = 𝑘X 𝑇•

b
In typical metals, 𝑇• ~ 50,000 K, At T = 300 K, ~ 0.01
b¹

𝜖
u Z Z
Classically (in absence of Pauli principle and FD distribution), 𝐶,' ~ 𝑁𝑘X 𝑇 = 𝑁𝑘X 𝑘X 𝑇
ub f f
(~ 100 times larger than the experimental values and hence does not represent the real situation.)
à Quantum statistics is important.

1 f 1 f
From a rigorous derivation, 𝐶g³ = 𝜋 𝐷(𝜖• ) 𝑇 𝐶g³ = 𝜋 𝑁𝑘X 𝑇/𝑇•
3 2

3𝑁 3𝑁 3𝑁
𝐷 𝜖 = → 𝐷 𝜖• = =
2𝜖 2𝜖• 2𝑘X 𝑇•
 Effective mass of an electron
At very low temperature,

𝑇 ≪ 𝑇• and 𝑇 ≪ 𝜃 𝐶 = 𝛾𝑇 + 𝛽𝑇 Z

Electron contribution Phonon contribution

𝐶/𝑇 = 𝛾 + 𝛽𝑇 f
γ is also known as the Sommerfeld coefficient..

j
But experimental values of γ are usually not the same as expected from 𝐶g³ = 𝜋 f 𝐷 𝜖• 𝑇.
Z
q f„ Z/f
𝐷 𝜖• = 𝜖• j/f
Thermal effective mass frk ℏk
f/Z
𝑚½¾ 𝛾(<Se,>¿,À) ℏf 3𝜋 f 𝑁
= 𝜖• =
2𝑚 𝑉
𝑚 𝛾(Á>,,)
𝐷 𝜖• ~𝑚~𝛾
Band effective mass
„ÇÈ Present in real crystals, but not accounted
Usually, ≠1 Electron phonon interaction
„ in the free electron theory.
Electron-electron interaction

„
Heavy fermions: Metals with very large effective masses ( ÇÈ > 1000 ) due to strong electron-electron correlations are
„
termed as heavy fermions. In such systems motion of conduction electrons become sluggish due to weak overlap of f-orbitals
of neighboring ions. Examples: UBe13, CeCu2Si2, CeAl3.
 Drude’s model for electrical conductivity and Ohm’s law
For a free electron 𝑚𝑣⃗ = ℏ𝑘

𝑑𝑣⃗ 𝑑𝑘
𝐹⃗ = 𝑚 =ℏ = −𝑒(𝐸 + 𝑣⃗×𝐵) Lorentz force
𝑑𝑡 𝑑𝑡
Electric field Magnetic field

𝑑𝑘
For 𝐸 ≠ 0, and 𝐵 = 0 ℏ = −𝑒𝐸 𝐸
𝑑𝑡
The applied electric field shifts the Fermi sphere.
𝑒𝐸𝑡 𝑒𝐸𝑡
𝑘 𝑡 −𝑘 0 =− 𝛿𝑘 = − In steady stare, the shift in FS by 𝐸 is counterbalanced by
ℏ ℏ
electron scattering processes.

ℏÏ• gÐÑ
In time t = 𝜏 (collision time), the increment in velocity is given by, 𝑣⃗ = =−
@ „

For a system of n free electrons per unit volume in the electric field 𝐸, the current density, ⃗ȷ = 𝑛𝑞𝑣⃗ = −𝑛𝑒𝑣⃗

‡g k Ñ „
‡g k Ñ where, σ = , or, 𝜌 =
⃗ȷ = 𝐸 =σ𝐸 „ ‡g k Ñ
„
Ohm’s law conductivity resistivity
Current density = Current/area = charge per unit time/area = charge density × velocity = 𝑛𝑞𝑣⃗
electron concentration charge velocity
 Experimental Electrical resistivity in metals : Matthiessen’s rule
A perfectly periodic lattice offer no resistance for the motion of electrons.
A departure from periodicity of lattice is the main source of electrical resistivity in metals.

Thermal lattice vibrations (phonons)

Discussed in Bloch-Grüneisen theory, 𝜌Š ~𝑇 Õ at low T and 𝜌Š ~𝑇 for 𝑇 > 𝜃

Origin of electrical resistivity in metals

Lattice imperfections: Impurities, defects, dislocations
It furnishes temperature independent residual resistivity (𝜌¡ )

Matthiessen’s rule: It is an empirical rule, which says that different mechanisms of

electrical resistivity work independently. The total resistivity then can be expressed as
a linear sum of individual sources of electron scattering.
𝜌 = 𝜌Š + 𝜌¡

×(Zcc Ø)
Residual Resistivity Ratio (RRR) = is a measure of purity of a metallic sample.
×(b→c)
Larger RRR à better purity metallic sample 𝜌¡

Another characteristic of a good metal is large mean free path: 𝑙 = 𝑣• 𝜏

 Electron transport in magnetic field: Cyclotron motion
𝐹⃗ = −𝑒(𝐸 + 𝑣⃗×𝐵)
𝑑 𝑚𝑣⃗
𝑚𝑣⃗ = −𝑒 𝐸 + 𝑣⃗×𝐵 −
𝑑𝑡 𝜏 In steady state, there is no acceleration
𝑑
Friction term →0
𝑑𝑡 Cyclotron frequency (𝜔Ù )
𝑑 𝑚𝑣⃗
𝑚𝑣⃗ + = −𝑒(𝐸 + 𝑣⃗×𝐵) 𝑚𝑣Q 𝑒𝜏 𝑒𝐵
𝑑𝑡 𝜏 = −𝑒(𝐸Q + 𝑣¦ 𝐵) 𝑣Q = − 𝐸Q − 𝜏𝑣
𝜏 𝑚 𝑚 ¦
𝑑 1
𝑚 + 𝑣⃗ = −𝑒(𝐸 + 𝑣⃗×𝐵) 𝑚𝑣¦ 𝑒𝜏 𝑒𝐵
𝑑𝑡 𝜏 = −𝑒(𝐸¦ − 𝑣Q 𝐵) 𝑣¦ = − 𝐸¦ + 𝜏𝑣
𝜏 𝑚 𝑚 Q
Let 𝐵 = 𝐵𝑧̂ , 𝑒𝜏
𝑣⃗ = 𝑣Q 𝚤̂ + 𝑣¦ 𝚥̂+𝑣¨ 𝑘© 𝑚𝑣¨ 𝑣¨ = − 𝐸¨
= −𝑒𝐸¨ 𝑚
𝐸 = 𝐸Q 𝚤̂ + 𝐸¦ 𝚥̂+𝐸¨ 𝑘© 𝜏

𝑑 1
𝑚 + 𝑣 = −𝑒(𝐸Q + 𝑣¦ 𝐵) Coupled 𝑣Q and 𝑣¦ leads to helical motion of an electron.
𝑑𝑡 𝜏 Q

𝑑 1 It gives rise to a positive magnetoresistance in metals.

𝑚 + 𝑣 = −𝑒(𝐸¦ − 𝑣Q 𝐵)
𝑑𝑡 𝜏 ¦
𝜌 𝐻 −𝜌 𝐻 =0
𝑀𝑅 =
𝑑 1 𝜌 𝐻=0
𝑚 + 𝑣 = −𝑒𝐸¨
𝑑𝑡 𝜏 ¨
 Hall Effect
The Hall effect is the electric field developed across two faces of a conductor, in the direction 𝚥⃗×𝐵, when a current 𝚥⃗ flows
across a magnetic field 𝐵.

𝐸Q (applied externally)
+ + + + 𝑦
Hall effect is measured on a thin
laminar sample, with magnetic field 𝐸¦ 𝑗Q
𝐵 = 𝐵𝑧̂
perpendicular to it. 𝑧 𝑥
- - - -

Lorentz force let electron deflect on edge of the sample. This creates a transverse electric field
𝑒𝜏 𝑒𝐵
𝑣Q = − 𝐸Q − 𝜏𝑣 Ey, which counterbalance it in steady state.
𝑚 𝑚 ¦ 𝐸¦
In such an equilibrium situation, 𝑣¦ = 0. Hall coefficient is defined by 𝑅Ý =
𝑗Q 𝐵
𝑒𝜏 𝑒𝐵
𝑣¦ = − 𝐸¦ + 𝜏𝑣 𝑒𝜏 𝑒𝐵
𝑚 𝑚 Q 𝑣Q = − 𝐸Q 𝐸¦ − 𝜏𝐸Q 1
𝐸 𝑒𝐵 𝑅 = = 𝑚 = −
𝑚 ¦ Ý
𝑒𝜏 =− 𝜏 𝑗Q 𝐵 𝑛𝑒 f 𝜏 𝑛𝑒
𝑣¨ = − 𝐸¨ 𝑒𝐵 𝑒𝜏 𝐸Q 𝑚 𝐸 𝐵
𝜏𝑣Q = 𝐸¦ 𝑚 Q
𝑚
𝑚 𝑚 𝑒𝐵 𝐸¦ 1
𝐸¦ = − 𝜏𝐸Q 𝑅Ý = =−
𝑒𝜏 𝑚 𝑗Q 𝐵 𝑛𝑒
𝑣¨ = − 𝐸¨
𝑚 𝑅Ý let us determine experimentally,
‡g k Ñ (1) carrier concentration,
⃗ȷ = 𝐸 =σ𝐸
„ (2) carrier type, i.e., electron (-ve 𝑅Ý ) or hole(+ve 𝑅Ý )
(3) Carrier mobility (together with magnetoresistance)
 Thermal conductivity of metals
Thermal conductivity 𝑗ß
Thermal gradient
∆𝑄 𝑑𝑇
𝑗ß = = −𝐾
𝐴𝑡 𝑑𝑥
Heat flux 𝑇j 𝑇f
𝑇j > 𝑇f
(energy transmitted across per unit area per unit time)

Mechanism: diffusion of thermal energy through electron collisions

𝑑𝑇
∆𝑇 = 𝑙 𝑙Q =mean free path in x-direction, 𝑣Q =average velocity in x-direction
𝑑𝑥 Q

ub
energy transmitted per particle ∆q = 𝑐∆𝑇 = 𝑐𝑙Q 𝑐 = heat capacity per particle
uQ
ub j ub
energy transmitted across per unit area per unit time, 𝑗ß = ∆q×particle flux = ∆q×n𝑣Q = 𝑐𝑙Q ×n𝑣Q = 𝐶𝑣𝑙
uQ Z uQ

𝑑𝑇 1 𝑑𝑇
𝑗ß = −𝐾 = 𝐶𝑣𝑙
𝑑𝑥 3 𝑑𝑥
𝑛=particle concentration in equilibrium, 𝑣= average particle velocity in 3D, C = heat capacity per unit volume, l=
average mean free path in 3D 1
𝐾 = 𝐶𝑣𝑙 𝐶 = 𝐶'())*+, + 𝐶,',+)><?*+
3
In pure metals electronic contribution for K is dominant over phonons.
 Wiedemann-Franz Law

Since The same conduction electrons transport heat energy and electric charge and, the ratio of thermal conductivity and
electrical conductivity is a universal constant. This fact is called Wiedemann-Franz law.

1 𝜋 f 𝑛𝑘Xf 𝑇 𝜋 f 𝑛𝑘Xf 𝑇𝜏 𝑛𝑒 f 𝜏
𝐾g³ = 𝐶𝑣𝑙 = . .𝑣 .𝑙 = 𝜎=
3 3 𝑚𝑣•f • 3𝑚 𝑚

𝜋 f 𝑛𝑘Xf 𝑇𝜏
𝐾 3𝑚 𝜋 f 𝑘X
= = 𝑇
𝜎 𝑛𝑒 f 𝜏 3 𝑒f
𝑚

Lorenz number Ø r k •–
𝐿= = = 2.45×10iæ watt-ohm/deg2
åb Z gk

This is indeed observed experimentally universally in many metals and is considered as a validation of the free electron
model of metals.
1 f 1 f 1
1 f 𝜋 𝑛𝑘X 𝑇 𝜋 𝑛𝑘X 𝑇 𝜋 f 𝑛𝑘Xf 𝑇 𝜖• = 𝑚𝑣•f 𝑙 = 𝑣• 𝜏
𝐶 = 𝜋 𝐷 𝜖• 𝑇 = 2 =2 = 2
3 𝑇• 1 f 𝑚𝑣 f
𝑚𝑣• /𝑘X •
𝐶= heat capacity per unit volume
2
 Critique of free-electron model

Free electron models gives a crude picture of electrons in solid explaining partially experimental electrical, thermal,
magnetic, and electrodynamic properties in metals.
However, it does not resolve the following experimental facts:

(1) Distinction between metals, semimetals, semiconductors, and insulators.

(2) Positive Hall coefficient for some materials.
(3) Effective mass of many solids are quite different than that expected from free electron model.
(4) Magnetic field dependent magnetoresistance in solids.

Reasons:
(1) The effect of periodic potential set by the lattice is ignored in the free electron model.
(2) Electron phonon interaction has not been considered in the free electron model.
(3) Electron-electron interaction is also ignored.

A better picture of reality is provided by band theory of solids, which considers periodicity in the lattice affecting motion
of electrons in solids.
 Energy bands

Chemistry viewpoint (Linear combination of atomic orbitals or LCAO

picture): The electronic levels (1s2, 2s2, 2p6,…) are sharply defined for
individual atoms situated far apart. When the atoms condense to form
solid, the atomic orbitals overlap resulting in formation of regions of
allowed electronic states (bands) and forbidden gaps in energy space.

Physics viewpoint: different extent of involvement of lattice and atomic

bonding gives rise to different models describing band structure. Band
theory incorporates lattice periodicity in electron dynamics.
Energy

𝑇≠0

Insulator Metal Semimetal Semiconductor

 Nearly free electron model

In nearly free electron model, the band structure of a crystal are treated as perturbed only weakly by the periodic
potential of the ion cores. r
At first Brillouin zone boundary, (i.e. 𝑘 = ± )
è
f f f
ℏ 𝑘 ℏ 1 𝜋
𝜖• = = (𝑘Qf + 𝑘¦f + 𝑘¨f ) 𝑘=± 𝐺=± 𝐺=
f‡r
, 𝑛 = 1,2,…
2𝑚 2𝑚 2 𝑎 è

‡h fr ‡ª fr ‡« fr
𝑘Q = , 𝑘¦ = , 𝑘¨ = According to nearly free electron model, energy gap develops at the zone boundary
Š Š Š
r
(𝑘 = ± ) due to Bragg reflection of electron wave functions in crystalline solids. This
𝜓• (𝑟⃗ ) = 𝑒 ¡•.£⃗ è
leads to formation of standing waves and hence non-existence of travelling waves as
𝑝⃗ = ℏ𝑘 permissible solution for the Schrodinger equation.

1D case
Free electron
(parabolic bands) Monoatomic linear lattice
 𝑘 −𝑘 Origin of energy gap in nearly free electron model
r
¡ Q
r
i¡ Q 𝜋
𝜓 + = 𝑒è + 𝑒 è = 2 Cos( 𝑥)
𝑎 Standing waves as a result of Bragg’s reflection
r
¡ Q
r
i¡ Q 𝜋 at the BZ boundary (𝑘 = )
r
𝜓 − = 𝑒è − 𝑒 è = 2i Sin( 𝑥) è
𝑎
ì
¡ Q
Electron probability density is uniform for travelling wave solution 𝜓 = 𝑒 í as 𝜌 𝑥 = |𝜓|f = 𝜓 ∗ 𝜓 = 1

r r
𝜌— 𝑥 = |𝜓(+)|f = 4 Cos f ( 𝑥) 𝜌i 𝑥 = |𝜓(−)|f = 4 Sinf ( 𝑥) 𝜓 −
è è

𝜓 +

The energy gap 𝐸ð is the difference between potential energies

r
corresponding to charge distributions 𝜌— and 𝜌i at 𝑘 = .
è
 Magnitude of energy gap

𝜋 𝜋
Normalized wave functions at the zone boundary are 𝜓 + = 2 Cos( 𝑥) 𝜓 − = 2 Sin ( 𝑥)
𝑎 𝑎

Potential energy of an electron in 1D lattice is expected to have the same periodicity as the lattice.
2𝜋
𝑈 𝑥 = 𝑈 Cos 𝑥
𝑎
Fourier component of U(x)

f f
2𝜋 𝜋f f
𝜋
𝐸ð = [ 𝑑𝑥𝑈 𝑥 𝜓 + − 𝜓 − = 2 [ 𝑑𝑥𝑈Cos 𝑥 Cos ( 𝑥) − Sin ( 𝑥)
𝑎 𝑎 𝑎
𝜓 −

fr j 𝑈
=2 ∫ 𝑑𝑥𝑈 Cos f 𝑥 = 2. 𝑈 = 𝑈 𝜓 +
è f

𝐸ð = 𝑈

The energy gap is equal to the Fourier component of the crystal potential.
 Bloch functions

Solution of the Schördinger equation for a periodic potential must be of a special form as first pointed out by Bloch.
It is applicable to all sort of crystalline solids (not limited to metals).

ℏf f
𝐻𝜓 = − 𝛻 + 𝑈 𝑟⃗ 𝜓 = 𝐸𝜓 𝑈 𝑟⃗ + 𝑅 = 𝑈(𝑟⃗)
2𝑚
Where, 𝑅= lattice translation vector

Bloch theorem: The Eigen functions of the wave equation for a periodic potential are the product of a plane wave 𝑒 ¡•.£⃗
times a function 𝑢• (𝑟⃗) with the periodicity of the crystal lattice.

𝜓• 𝑟⃗ = 𝑒 ¡•.£⃗ 𝑢• 𝑟⃗ , with 𝑢• (𝑟⃗ + 𝑅) = 𝑢• (𝑟⃗)

𝜓• 𝑟⃗ are called Bloch functions.

Bloch functions can be decomposed into a sum of travelling waves.

Bloch functions can be assembled into localized wave packets to represent electrons that propagate freely through
the potential field of ion cores.
 Proof of Bloch theorem
𝜓 𝑥
Periodic boundary condition in a linear lattice of N atoms is equivalent to a ring of length Na. a
𝐶𝜓 𝑥
𝑈 𝑥 = 𝑈 𝑥 + 𝑠𝑎 , 𝑠 =1 to N

We look for solutions of the wave equation such that, 𝜓 𝑥 + 𝑎 = 𝐶𝜓 𝑥 ,

where C is a constant in complex domain.
𝐶f𝜓 𝑥
→ 𝜓 𝑥 + 𝑁𝑎 = 𝐶 ° 𝜓 𝑥

Since 𝜓 𝑥 is single valued, 𝐶 ° = 1 à C must be one of the N root of unity in complex plane. 𝐶Z𝜓 𝑥

𝐶 ° = 𝑒 ó(fr¡)
kìô
ó
𝐶=𝑒 õ , s = 0,1,2, ……, N-1

kìöh
¡
𝜓 𝑥 = 𝑢• 𝑥 𝑒 õí satisfies the condition 𝜓 𝑥 + 𝑎 = 𝐶𝜓 𝑥 when 𝑢• 𝑥 = 𝑢• 𝑥 + 𝑎 .

This proves the Bloch result.

 Kronig-Penney model
This models considers an array of 1D square well potential that simulates crystal lattice.

ℏf 𝑑 f
𝐻𝜓 = − +𝑈 𝑥 𝜓 = 𝜖𝜓
2𝑚 𝑑𝑥 f

0, 0<𝑥<𝑎
𝑈 𝑥 =÷
𝑈c , −𝑏 < 𝑥 < 0

In the potential well region (0 < 𝑥 < 𝑎)

2𝑚𝜖
𝜓 = 𝐴𝑒 ¡ØQ + 𝐵𝑒 i¡ØQ 𝐾=
ℏ
In the potential barrier region (−𝑏 < 𝑥 < 0)
2𝑚(𝑈c − 𝜖)
𝜓 = 𝐶𝑒 úQ + 𝐷𝑒 iúQ 𝑄=
ℏ
From Bloch theorem, 𝜓 𝑎 < 𝑥 < 𝑎 + 𝑏 = 𝜓 −𝑏 < 𝑥 < 0 𝑒 ¡•(è—û)

Note that 𝑘 represents wave vector of electrons in solid and is different than K in the solution of the Schördinger equation.
 uü
Boundary conditions: 𝜓 and are continuous at 𝑥 = 0 and 𝑥 = 𝑎.
uQ

𝐴+𝐵 =𝐶+𝐷
𝑖𝐾(𝐴 − 𝐵) = 𝑄(𝐶 − 𝐷)

𝐴𝑒 ¡Øè + 𝐵𝑒 i¡Øè = (𝐶𝑒 úû + 𝐷𝑒 iúû )𝑒 ¡•(è—û)

𝑖𝐾(𝐴𝑒 ¡Øè − 𝐵𝑒 i¡Øè = (𝐶𝑒 úû − 𝐷𝑒 iúû )𝑒 ¡•(è—û)

For a non-trivial solution, the determinant of the

coefficient matrix must be zero.
This, on simplification gives the following relation.

𝑄f − 𝐾 f
Sinh 𝑄𝑏 Sin 𝐾𝑎 + Cosh 𝑄𝑏 Cos 𝐾𝑎 = Cos (𝑘 𝑎 + 𝑏 )
2𝑄𝐾

ú k ûè
In the limit of delta function potential, i.e., 𝑈c → ∞ and 𝑏 → 0 such that = 𝑃 is finite, the above equation reduces to:
f
𝑃
Sin 𝐾𝑎 + Cos 𝐾𝑎 = 𝐶𝑜𝑠(𝑘𝑎)
𝐾𝑎
It does put constraint on allowed energy values of electrons in a periodic potential and gives energy gaps at zone boundaries.

2𝑚𝜖 2𝑚(𝑈c − 𝜖)
𝜓 = 𝐴𝑒 ¡ØQ + 𝐵𝑒 i¡ØQ inside well 𝐾= 𝜓 = 𝐶𝑒 úQ + 𝐷𝑒 iúQ inside barrier 𝑄=
ℏ ℏ
 𝑃 2𝑚𝜖
Sin 𝐾𝑎 + Cos 𝐾𝑎 = Cos(𝑘𝑎) 𝐾= while 𝑘 is introduced by Bloch function.
𝐾𝑎 ℏ

The RHS put a constraint on LHS to be ≤ 1.

à Only those values of K are allowed for which the −1 ≤ 1 LHS ≤ 1. Others
will correspond to the he forbidden energy gap.

Allowed bands
 Number of orbitals in a band

Allowed values of 𝑘 in the first Brillouin zone for a 1D crystal composed of N atoms are given by
2𝜋 4𝜋 𝑁𝜋 𝐿 = 𝑁𝑎
𝑘 = 0, ± ,± ,…….±
𝐿 𝐿 𝐿
𝑎
Total number of independent 𝑘-points (i.e. in the 1st BZ) = No. of primitive cells = 𝑁
Š °r r
= 𝑎 or =
Each primitive cell contributes exactly one independent value of 𝑘 ° Š è

to each energy band.

Since there are 2 independent orientations of electron spin, there are 2N

orbitals in each energy band.
Single valence electron per atom à half filled band
Two valence electron per atomà fully occupied band

𝑁𝜋
𝐿
A translation in 𝑘 space by any arbitrary 𝐺⃗ leaves the band invariant.
à All bands can be drawn within the first BZ (reduced band scheme).
f‡r
𝐺= , n=1,2,… 𝐺⃗
è
 Metals and Insulators (0 K)

Insulators: Valence electrons exactly fill one or more bands leaving others empty.
An external electric field will not cause current flow because there is no continuous way to change the total momentum of
electrons due to unavailability of free states to move in.
The number of valence electrons in primitive cell is even in insulators in nearly free electron model. However, it is necessary
but not a sufficient condition. The additional condition is that the top most filled bands should not overlap in energy.

At 0 K, semiconductors are insulators.

Metals: In metals, either the top most is partially occupied, or the two bands in energy overlap in energy (not necessarily at
the same k-point). A small overlap with involvement of only a few states gives rise to semi-metallic behavior.

𝑘¦ 𝑘¦

𝑘Q 𝑘Q

Metal (large overlap)

Metal 𝐸=0 𝐸≠0
Insulator Semi metal (small overlap)
 Fermi surface
The Fermi surface is the surface of constant energy 𝜖• in 𝑘-space, which separates the unfilled orbitals from the filled
orbitals, at absolute zero temperature.
Cu: [Ar]3d104s1 Al: [Ne]3s23p1
valence electrons

Nearly free electron model (𝑈 ≠ 0)

Free electron model (𝑈 = 0)
Half occupied

The deformation from spherical shape

of the FS is due to interaction of free
electrons with the lattice.
Fully occupied

A Fermi surface can exist only in metals. In other words, a metal can be defined as a solid with a Fermi surface.
The electrical properties are of the metal are determined by volume and shape of the Fermi surface.
 Mapping a band in reduced zone scheme

From Bloch theorem 𝜓• 𝑟⃗ = 𝑒 ¡•.£⃗ 𝑢• 𝑟⃗ , where 𝑢• (𝑟⃗ + 𝑅) = 𝑢• (𝑟⃗)

If 𝑘 % lies outsize the first BZ, one can always find a suitable 𝐺⃗ , such that
𝑘 = 𝑘 % + 𝐺⃗
$
𝜓• $ 𝑟⃗ = 𝑒 ¡• .£⃗ 𝑢• $ 𝑟⃗ = 𝑒 ¡•.£⃗ (𝑒 i¡#⃗ .£⃗ 𝑢• $ 𝑟⃗ ) = 𝑒 ¡•.£⃗ 𝑢• 𝑟⃗ = 𝜓• 𝑟⃗

Since 𝑢• $ 𝑟⃗ + 𝑅 = 𝑢• $ 𝑟⃗ and 𝑒 i¡#⃗ . £⃗—&

= 𝑒 i¡#⃗ .£⃗ ,
therefore 𝑢• 𝑟⃗ = 𝑒 i¡#⃗ .£⃗ 𝑢• $ 𝑟⃗ is also periodic in the crystal lattice

Any wave vector 𝑘 % lying outside the first Brillouin zone can be equivalently brought within the first Brillouin zone.
This transformation does not change any underlying physics.
Then the band structure is called in the reduced band scheme.

𝜓•—#⃗ 𝑟⃗ = 𝑒 ¡(•—#⃗ ).£⃗ 𝑢•—#⃗ 𝑟⃗ = 𝑒 ¡•.£⃗ 𝑒 ¡#⃗ .£⃗ 𝑢•—#⃗ 𝑟⃗ = 𝑒 ¡•.£⃗ 𝑢• $ 𝑟⃗

In other way,
à If 𝜓• 𝑟⃗ is solution of SE describing the band structure, so is 𝜓•—#⃗ 𝑟⃗ .
 Reduced zone scheme in free electron model

ℏf 𝑘 f
𝜖• =
2𝑚

The reduced zone scheme introduces band index 𝑛.

Each value of 𝑛 characterizes a different band.

𝜓‡• 𝑟⃗ = 𝑒 ¡•.£⃗ 𝑢‡• (𝑟⃗)

A finite lattice potential leads to appearance of energy 𝜓f•

gap at the zone boundaries due to Bragg reflection of
electron waves.

𝑈 ≠ 0 introduced 𝐸ð
at BZ boundaries 𝜓j•
 Schemes in band structure
Band structure: The electronic band structure (or 𝜋 𝜋
simply band structure) of a solid describes the range −
𝑎 𝑎
of energy levels that electrons may have within it, as well as
the ranges of energy that they may not have (called band
gaps).
There are three schemes to represent band structure.

Extended zone scheme: Different bands are drawn in

different zones in wave vector space

Reduced zone scheme: All bands are drawn in the first

Brillouin zone

Periodic zone scheme: Every band is drawn in every zone

𝜖Ø = 𝜖Ø—#
The most compact and versatile method of showing a band structure is the reduced zone scheme.
 Construction of Fermi surfaces and Brillouin zones

The equation of the zone boundaries is 2𝑘. 𝐺⃗ + 𝐺⃗ = 0 and is satisfied if 𝑘 terminates on the plane normal to 𝐺⃗ at the
midpoint of 𝐺⃗ .
This construction divides the k-space into distinct regions called 1st, 2nd, 3rd, … Brillouin zone.
Each Brillouin zone encloses equal area (volume in 3D).
 Fermi surface and electrons in reduced zone scheme

Free electron model Fermi surface In reduced zone scheme

In extended zone scheme

𝜓‡• 𝜓j• 𝜓f• 𝜓Z•

 Reduced zone scheme

In some cases, periodic zone scheme let us visualize symmetry of the Fermi surface in a better manner.