Summary solid states physics

Crystal Structures
Solids are often crystalline, with the atoms arranged on a regular lattice. We will often imagine a
solid as one crystal, a perfect lattice of atoms without any defects

General description of crystal structures

A lattice is a set of regularly spaced points with positions defines as multiples of generating vectors.

Bravais lattice = all the points that can be reached by the vectors R, created from 3 vectors:

Number of possible Bravais lattices is limited to 14 in 3D (5 in 2D).

Length of the vectors a1, a2, a3 = lattice constant.

Primitive unit cell = any volume of space that, when translated through all the vectors of the Bravais
lattice, fills space without overlap and without leaving voids. Contains only one lattice point.
If there is a lattice point at the edge of a cell and thus shared with another cell, it is only counted half.
Accordingly, a point located on the corner of a cube is shared by 8 cubes and would count with 1818.
Special choice of primitive unit cell: Wigner-Seitz cell -> region of space that is closer to one given
lattice point than to any other.
Nonprimitive unit cells contain several lattice points. These fill space without leaving voids when
translated through a subset of the Bravais lattice vectors.

Basis = building block for the real crystal, can consist of one or several atoms.

Some important crystal structures

Cubic structures
The FCC has the highest possible
packing density for spheres.

BCC and FCC are very common

Only for the simple cubic

structure, the cube is identical
with the Bravais lattice. For bcc
and fcc lattices, the cube is also a
unit cell, but not the primitive one.

Cubic structure with a more complex basis than a single atom: CsCl and NaCl (basis containing two
atoms).

Close-packed structures
Many metals prefer structural arrangements where the
atoms are packed as closely s possible.
closest possible packing of ions in 2D: hexagonal structure
Adding a third layer-> 2 possibilities:
- put the ions in the holes just on top of the first layer
ABABAB: hexagonal close-packed structure (hcp)
- put them into other type of holes ABCABCABC: fcc
Both have exactly the same packing density (spheres fill 74% of the total volume) and are very
common for elemental metals. These structures also maximize the number of nearest neighbors for a
given atom (coordination number-> 12)

Crystal structure determination

X-ray diffraction
Incoming X-rays get scattered not more than once (most are not scattered at all because X-rays
interact rather weakly with matter).
Assume that the X-ray source and detector are very far away from the sample so that the incoming
and outgoing waves can be treated as plane waves.

Bragg Theory (special case of complex Laue description of X-ray diffraction)

Reflection of the incoming X-rays at flat crystal planes (e.g. close-packed planes making up the fcc
and hcp crystals or alternating Cs and Cl planes making up the CsCl structure).

The condition for constructive

interference is that the path length
difference between the X-rays reflected
from one layer and the next layer is an
integer multiple of the wavelength:
2AB = nλ
with sinϴ=AB/d we arrive at the Bragg
condition:

fulfilled for any number of layers with identical spacing only for λ < 2d

Works for all possible parallel planes in a structure.

useful for extracting the distances between lattice planes in a crystal but it has its limitations:
- does not give any information on what the lattice actually consists of (basis)
- crystal planes are certainly not flat for X-rays that have a wavelength similar to the atomic spacing

Lattice planes and Miller indices

Lattice plane = plane containing at least three noncollinear points of a given Bravais lattice.
Can be characterized by a set of three integers = Miller indices

1) We find the intercepts of the plane with the crystallographic axes.

2) We take the reciprocal value
3) By multiplying with some factor, we reduce the numbers to the smallest set of integers having the
same ration

General diffraction theory (M. von Laue)

The electromagnetic filed of the X-
rays forces the electrons in the
material to oscillate with the same
frequency as that of the field. The
oscillating electrons then emit new
X-rays that give rise to an
interference pattern (Something
scatters the X-rays).
 Derivation p. 10/11

relates the measured intensity to

the electron concentration in the
sample.

K = k’ – k = scattering vector
length of k’ = length of k (elastic scattering)

The Reciprocal Lattice

It permits us to exploit the crystal symmetry for the analysis of many problems.
For a given Bravais lattice R we define the reciprocal lattice as the set of vectors G for which:

b1, b2, b3 spanning the reciprocal lattice.

X-Ray Diffraction from Periodic structures

The reciprocal lattice facilitates the description of functions that have the periodicity of the lattice.
The electron concentration is lattice-periodic.

insert this expression in I(K) we get:

Exponential function of I(K): plane wave with a wave vector G-K.

Laue condition: K = k’-k = G. Difference between incoming and scattered wave vector = reciprocal
lattice vector. Value of the Integral = volume of the crystal.

Edwald Construction
To visualize the Laue condition and to determine the directions k’ for which constructive interference
is to be expected.

1) Draw the wave vector k of the incoming X-rays such that it

ends in the origin of the reciprocal lattice.
2) We construct a circle of radius k around the starting point
of k
3) Wherever the circle touches a reciprocal lattice point, the
Laue condition k’-k = G is fulfilled.

It is not likely for the sphere to hit a second reciprocal lattice

point, co that constructive interference is only expected for
very few directions. As in the Bragg description, we see that
the wavelength of the X-rays has to be short enough (k has to
be long enough) for any constructive interference to occur.
Bonding in solids
Only the valence electrons, the electrons in the outermost shell(s), participate in the bonding

Attractive and Repulsive Forces

An attractive force is necessary for any bonding.
A repulsive force is required to keep the atoms from getting too close to each other.

The reason for the strong repulsion at short distances is the Pauli
exclusion principle (forbidden to have more than two electrons in
the same quantum state).

Ionic Bonding
Involves the transfer of electrons from an electropositive atom to an electronegative atom.
Bonding force: Coulomb attraction.

turning the atoms into ions costs some energy:

ex. (NaCl): Ionization energy of Na: 5.1 eV, electron affinity of Cl: 3.6 eV-> net energy cost: 1.5 eV
energy gain is given by the Coulomb interaction. For just one Na and ond Cl ion separated by the
distance a = 0.28 mm this is -e2/4πε0a = 5.1 eV.
Knowing the crystal structure for NaCl we can also calculate the electrostatic energy gain (= lattice
energy) for forming an entire crystal:

Md = Madelung constant (specific for a given structure)-> can calculate it.

Cohesive energy = total energy difference between any solid and the isolated atoms it is made of.

Covalent Bonding
Based on the true sharing of electrons between different atoms.
often found for elements with a roughly half-filled outer shell.
It is very stable to make linear combinations of orbitals: sp3, sp2

Metallic Bonding
In metals, the valence electrons are removed from the ion cores, but in contrast to ionic solids, there
are no electronegative ions to bind them. Therefore, they are free to migrate between the ion cores.
These delocalized valence electrons (conduction electrons) are involved in the conduction of
electricity.
Metals prefer close-packed structures. The metallic bonding does not have any directional
preference. Closed-packed structures secure the highest possible overlap between the valence
orbitals of the atoms, maximizing the delocalization of the electrons and thereby the kinetic energy
gain. The structures also maximize the number of nearest neighbors for any given atom.
Metallic bonding is not as strong as covalent or ionic bonding.
Hydrogen Bonding
Hydrogen atoms: only one electron-> can form one covalent bond
If the bond is formed with a very electronegative atom (like F or O), the electron is mostly located
close to that atom and the hydrogen nucleus represents an isolated positive (partial) charge which
can attract negative (partial) charges in other molecules to form an electrostatic bond.
e.g. bonding of the double helix in DNA

Van der Waals Bonding

Refers to a weak and purely quantum mechanical effect.
A simple atom with a closed shell can thus be viewed as a fluctuating dipole: the field of this dipole
can polarize other atoms nearby, and the interaction of the two neighboring dipoles reduces the
total energy-> lead to bonding.
- much weaker than ionic, covalent, or metallic bonding.
- Van der Waals bonding is only observable for solids that do not show other bonding behavior (e.g.,
noble gases). Pure van der Waals crystals can only exist at very low temperatures.

Thermal Properties of the Lattice

Lattice Vibrations
The atoms in a crystal can vibrate around their equilibrium position. Assume a linear restoring force
leads to a description of the lattice vibrations as harmonic oscillators.

A simple harmonic oscillator

Describe the vibrations of a solid with N atoms simply as 3N (3 directions) independent harmonic
oscillators (reality: the oscillators would couple to each other).

Equation of motion where M = mass of vibrating atom

𝛾
frequency w = √
𝑀

Total energy: equipartition theorem: kbT/2 = <E>

1
=> 2 𝛾𝑥𝑚𝑎𝑥 2 = 𝑘𝑏 𝑇

An infinite Chain of Atoms (lead to infinite heat capacities)

The oscillators are coupled to each other (longitudinal vibrations)

One Atom per Unit Cell

Un = displacement of the nth atom in the chain.

The equation of motion can be solved by:

Substituting into the equation of motion, we get the solution:

 Dispersion relation (connecting a

frequency/energy to a wave vector).
 Particular solution with w(k) = normal mode
 𝑤 𝜕𝑤
Phase velocity: Group velocity:
𝑘 𝜕𝑘
2𝜋
How do the atoms move? k =
𝜆
𝝅
k << 𝒂

The wavelength of the mode must be much longer than the lattice constant-> The atoms that are
close to each other must very nearly move in phase.
The particular atomic structure is thus not important, it would be sufficient to view the chain as a
macroscopic elastic medium.

We see that phase velocity = group velocity = ν

The propagation speed of the waves does not depend on their frequency.
The long-wavelength limit corresponds to the propagation of sound waves with ν being the speed of
sound.
𝝅
k=𝒂

The shortest possible wavelength λ = 2a. If an atom moves to the right, the atom two lattice spacings
away must perform the same motion. The atom on the neighboring lattice site, is half a wavelength
𝜋
away and must therefore move in the opposite direction. The group velocity for k = is zero .
𝑎
2𝜋
The solutions of the equation of motion are standing waves. w(k) is periodic in k with a period of 𝑎
.
This periodicity corresponds precisely to one reciprocal lattice vector (actual motion is unaffected if
we add a reciprocal lattice vector to k).
𝜋 𝜋
The First Brillouin Zone = region between k=− 𝑎 and k= 𝑎
The dispersion relation w(k) and the motion of the atoms is unaffected if we change the wave vector
2𝜋
by multiples of the reciprocal lattice vector .
𝑎

Two atoms per unit cell

The solutions have again the periodicity of the

reciprocal lattice.
We have now two branches of solutions.

Acoustic branch: the solution that goes to zero for

small k-> propagation of sound waves through the
crystal.

Optical branch: The solution that has a finite w at k = 0, possibility to couple these vibrations to the
oscillating electric field of an electromagnetic wave.
A finite chain of atoms
 For a real solid we want a finite but long chain of atoms: limiting the length and introducing
boundary conditions (e.g. hold the atoms at the ends fixed-> leads to standing waves in the
chain because we have fixed the nodes).

Cyclic/periodic boundary condition: end is tied to the beginning

The dispersion relations w(k) are not affected by the chain being finite but the periodic boundary
conditions restrict the possible k values for the waves in the crystal.
The longest possible wavelength for a chain of N atoms with spacing of a is: Na

𝑒 𝑖𝑘𝑎𝑛 = 𝑒 𝑖𝑘𝑎(𝑁+𝑛) 𝑠𝑜 𝑡ℎ𝑎𝑡 𝑒 𝑖𝑘𝑎𝑁 = 1

2𝜋
For this to be fulfilled-> periodic boundary condition: 𝑘 = 𝑚 where m is an integer
𝑎𝑁
2𝜋
The vibrations are unaffected by adding multiples of to k-> we only get N possible different values
𝑎
for k, N different vibrational frequencies w per dispersion branch.

Quantized Vibrations, Phonons

 The lattice vibrations have a quantized character

Quantized energy levels (derived from the harmonic oscillator).

k takes the role of a quantum number, used to label different normal
modes.

In atoms, we have spherical symmetry-> quantum numbers l and m

In solids, we have translational symmetry -> quantum number k

The normal modes can thus be excited in discrete energy quanta-> excitation = phonons
(analogy: photons = excitations of the electromagnetic field)
As photons, phonons can have wave character as well as particle character.
When we want to describe properties like thermal conductivity, we need to think of phonons as
particles that are generated at the hot end and conducted to the cold end.
Phonons and Photons are Bosons and therefore not subject to the Pauli exclusion principle.
Only phonons with k = 0 can be excited by light.

Three dimensional solids

- 3D solid with 2 atoms per unit cell:
6 vibrational frequencies w for every value of k
3 acoustic branches; 1 with longitudinal polarization, 2 with transverse polarization
3 optical branches
- 3D solid with 1 atom per unit cell:
3 acoustic branches

Generalization of periodic boundary

condition:

Where L = macroscopic side length of the crystal

Brillouin zone = Wigner-Seitz cell of the reciprocal lattice.

- Points of high symmetry are abbreviated by certain letters:

e.g. Γ= center of the Brillouin zone
Heat capacity of the lattice
Classical Theory and Experimental Results
Two different approaches to calculate the heat capacity of a solid giving not the same answer

- Classical thermodynamics: does not tell us anything about the value of the heat capacity at
finite temperature but it can be shown that it should vanish for zero temperature
- Classical statistical mechanics: gives us the probability to calculate a numerical value for the
heat capacity of a solid via the equipartition theorem (3D: <E> = 3kBT)
𝜕<𝐸>
-> heat capacity: 𝜕𝑇
= 3𝑘𝐵
for 1 mol of atoms: 3𝑘𝐵 𝑁𝐴 = 3𝑅 = 29.9 𝐽𝐾 −1
= rule of Dulong-Petit: heat capacity is independent of the temperature and the material ->
conflict with thermodynamics (vanishing heat capacity at zero temperature)
Looking at the experiment we see that the
Dulong-Petit value is rather good at room
temperature but less good at the boiling point of
nitrogen (heat capacity is smaller). A vanishing
heat capacity at zero temperature is also
supported by the experimental results.

Looking at the picture on the right, at high

temperatures the heat capacity of diamond
approaches the Dulong-Petit value. In the low-
temperature limit 𝐶 ∝ 𝑇 3

Einstein model
Using quantum mechanics to describe the oscillators in the solid.

<n> can be found using the Bose-Einstein distribution:

Heat capacity:

In the high-temperature limit (at least temperature corresponding to the vibrational frequency of the
ħ𝑤𝐸
oscillators Θ𝐸 = 𝑘𝐵
: reproduce the Dulong-Petit value
The head capacity also drops to zero at lower temperatures but too quickly (shows an exponential
behavior instead of 𝐶 ∝ 𝑇 3.-> All the Einstein oscillators get frozen out below Θ𝐸 when there is not
enough thermal energy available to supply the ħ𝑤𝐸 required to excite them out of their ground
state.

Debye Model
Problem of Einstein can be cured by using a more realistic model for the lattice vibrations:
quantum mechanical oscillators with very small energy level spacings near k = 0-> can be excited at
very low temperatures. These excitations are the sound waves for which the dispersion of the
acoustic branch (p.5) is approximated by the linear dispersion (p.6).
Assumption: dispersion hold for all values of k.
ignores the existence of optical branches-> not excited at low temperatures anyway
Mean energy for all oscillators in 3D:
ωD = highest phonon frequency in the material.
3 possibilities for the wave polarization

Assuming all 3 waves follow the dispersion 𝜔 = 𝜈𝑘

There may be more oscillators in the frequency interval 𝜔1 + 𝑑𝜔 than in 𝜔2 + 𝑑𝜔
-> weighting factor g(𝜔) density of states = number of states in a small frequency interval d𝜔

Calculate density of states:

1) figure out N(𝜔); the number of states below a certain frequency 𝜔
4 4 𝐿𝑘𝑚𝑎𝑥 3 4 𝐿𝜔 3 𝑉
within a sphere of radius nmax: 𝑁 = 3 𝜋𝑛𝑚𝑎𝑥 3 = 3 𝜋 ( 2𝜋
) = 3 𝜋 (2𝜋𝜈) = 6𝜋2 𝜈3 𝜔3
𝑑𝑁 𝑉
2) 𝑔(𝜔) = 𝑑𝜔 = 2𝜋2 𝜈3 𝜔2

Find the upper limit 𝜔𝐷 of the integral:

we should recover the correct number of normal modes, for N atoms in the solid:

And we find the Debye frequency

(highest vibrational frequency):
ħ𝜔𝐷
Θ𝐷 = 𝑘𝐵
: Debye temperature: only for temperatures above, the Dulong-Petit law is obeyed.

Plug everything in the integral:

For high T: x small-> 𝑒 𝑥 = 1+x

<E> = 2NkBT (Dulong Petit)

For low T: x large-> xD -> ∞

Thermal Conductivity
 We need vibrational excitations that can be generated at the hot end of the solid and which
then propagate to the cold end (phonons = particles)
 Achieved by the superposition of normal modes to generate wave packets that then travel
through the crystal with a certain group velocity.
 Phonons are a gas of particles bouncing through the solid

Total thermal conductivity = sum of the lattice and electronic thermal conductivity:

c = heat capacity, vp = speed of sound

mean free path λp:
at low temperatures: scattering from crystal imperfection, heat capacity decreases-> Kp decreases
at high temperatures: number of phonons increases-> phonons can be scatted from other phonons
=> λp decreases => Kp decreases
Kp reaches a maximum at an intermediate temperature (about 10% of the Debye temperature).

Not only metals show high thermal conductivity (lots of free electrons) but also insulators, especially
diamond which has a nearly perfect crystal structure such that defect scattering of phonons in
unimportant.
Electronic Properties of Metals: Classical Approach
Basic Assumptions of the Drude Model
 Simple model to explain many of the observed properties of metals
 He combines the existence of electrons as charge carriers with the highly successful kinetic
gas theory.
 The model is based on the following assumptions
- Independent electron approximation: electrons in a solid behave like a classical ideal gas.
They do not interact with each other at all: no Coulomb interaction, no collisions between
each other
- Free electron approximation: positive charge is located on immobile ion cores. The electrons
can collide with ion cores which change their velocity. In between collisions, the electrons do
not interact with the ions.
- The electrons reach thermal equilibrium with the lattice through the collisions with the ions.
mean kinetic energy:
- In between collisions, the electrons move freely. The mean length of this free movement =
𝜆
mean free path λ-> mean time between collisions: 𝜏 =
𝑣
- conducting electron density n = number of conduction electrons per volume
𝑁 𝑁𝜌
𝑛= =
𝑉 𝑀
Results of the Drude Model
 several properties of metals can be explained by the Drude model

DC Electrical Conductivity
 calculate the drift velocity induced by the electric field of the electrons from the equation of
motion
 drift velocity is very slow compared to the thermal movement of the electrons-> don’t have a
significant effect on the relaxation time
 with the drift velocity we can calculate the current density and we get the conductivity σ and
the resistivity ρ
e2: get the same result for
positive charge carriers

This result is the familiar Ohm’s law.

mobility characterizes how quickly an electron can move through a metal or

semiconductor when pulled by an electric field.

Hall Effect
 influence of a magnetic field on the current in a conductor
 An electric field (Hall field EH) is built up, which is perpendicular to both the magnetic field
and the current density
RH = Hall coefficient
For the electrons to pass through the sample, the Hall field must exactly compensate the Lorentz
force in the opposite direction:

with the current density we get:

Measuring the hall coefficient provides direct experimental access to the conduction electron density
The measured RH is even worse for Be, Mg and Al -> RH is positive-> the current in Be, Mg and Al is
carried by positive charges.

The Wiedemann-Franz Law

 The ratio of thermal to electrical conductivity is constant for all metals at a given
temperature.
where L is the Lorenz number

Electronic Properties of Solids: Quantum Mechanical Approach

- Going beyond the assumption of free electrons-> also possible to explain nonmetallic solids
- Retain the approximation that the electrons move independently from each other
- Ignore the motion of the ions-> freezing them into their equilibrium position (but there are
thermal vibrations)
➔ Born-Oppenheimer approximation: separate the electronic and ionic motions
- One-electron approximation: Do not consider the correlated motion of the electrons

The Idea of Energy Bands

Example:
We build a molecule from 2 Na atoms that have only
one valence electron.
a) The atoms approach each other, and a bonding
molecular orbital is formed. To fulfill the Pauli principle,
the electrons have opposite spins.
b) The energy gain is maximized for the separation a,
and zero for a very large separation of the atoms.
N Na atoms: N atomic energy levels split up into N
nondegenerate molecular levels. N/2 of these levels are then
occupied by two electrons each. For very large N, there is a
quasi-continuum = energy band between the lowest and highest
energy levels.
The energy band of the valence electrons in Na is exactly half-
filled. When an electric field is applied, the electrons have to go into a state with slightly higher
energy-> there are plenty of unoccupied states available-> Na show metallic behavior

Consider the formation of energy bands in Si:

valence electrons involved in the bonding: two 3s and two 3p electrons
As they approach, the orbitals hybridize and form two bands of states at equilibrium distance a.
The lower band consists of the sp3 orbitals, and these are fully occupied by the four valence electrons
of each Si atom. The upper band is completely unoccupied, and between the two bands, there is an
energy region without any states-> band gap
When a voltage is applied, the electrons in the filled sp3 band cannot increase their kinetic energy
because there are no vacant states with energies slightly above the band. --> insulating behavior
Free Electron Model
 Quantum mechanical analogue to the Drude model

The Quantum Mechanical Eigenstates

We want to solve: With periodic boundary conditions (cubic box with V = L3):

Normalized solutions: plane waves allowed k-values:

Energy levels:

separation between the energy levels is of order: small-> L macroscopic distance

 This model already gives rise to a quasi-continuum (or a band) of energy levels

What is the highest occupied energy = Fermi energy EF that we get when filling up the states?

- N electrons in our enclosed volume-> electron density n = N/V = N/L3

- These can be accommodated on the N/2 states with the lowest energy since we can have 2
electrons per state
- We have to fill all states that lie within a sphere of radius nmax/ kmax such that:
𝑁 4 2𝜋 𝑁 2𝜋 3 4
= 𝜋𝑛𝑚𝑎𝑥 3 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒 𝐹𝑒𝑟𝑚𝑖 𝑤𝑎𝑣𝑒 𝑣𝑒𝑐𝑡𝑜𝑟 𝑘𝑚𝑎𝑥 = 𝑛𝑚𝑎𝑥 => ( ) = 𝜋𝑘𝑚𝑎𝑥 3
2 3 𝐿 2 𝐿 3

Fermi velocity vF = 2 EF/ me = order of 106 m/s-> very high (calculated for 0K)

Density of states:

At a finite temperature, electrons will be thermally excited from their ground state.
Occupation probability for fermions: Fermi-Dirac distribution f(E, T):

µ = chemical potential (for metals µ = EF)

At zero temperature:
E < µ: f(E,T) = 1 all states below µ are occupied
E > µ: f(E,T) = 0 all states above µ are empty

At finite temperature:
The Fermi-Dirac distribution develops a soft zone around µ in
which the occupation probability is no longer 1 or 0 but something in between.

Density of occupied electron states at a given energy and temperature: multiplying the desity of states g€
with the Fermi-Dirac distribution f(E,T). Total number of electrons N:

-> calculate the chemical potential at any temperature

Electronic heat Capacity
When the temperature of the solid is raised, only a very small fraction of electrons can be thermally
excited. There are already other electrons occupying the states at slightly higher energies. A
contribution to the heat capacity is, in fact, only possible for the electrons near the Fermi energy.

Heat capacity is proportional to the density of states at the Fermi energy: only
the electrons close to the Fermi energy can participate in thermal excitations.

Heat Capacity is linear in T, whereas the (low temperature) heat capacity of the lattice is proportional
to T3.-> At low temperature the lattice contribution vanishes faster than the electronic contribution.

The general form of the electronic states

Shortcomings of the free electron model:

- We got a continuum of states which does not agree for insulators with the energy gap
- It does not cure the problem of the positive hall coefficient of Aluminum which is a simple
metal.
- Not able to resolve some of the most basic questions of electron motion in solid
example: the mean free path of the electrons can reach macroscopic distances at low
temperatures. We do not understand how the electrons can move through the lattice of ions
without scattering.

Bloch’s Theorem where uk(r) is a function with the periodicity of the

Bravais lattice uk(r) = uk(r + R)

Solution: plane wave modulated by a lattice-periodic function-> electronic states are spread out over
the whole crystal
The probability density for finding an electron may vary within one unit cell, but it is identical for the
corresponding positions within every unit cell in solid. -> electrons travel through the crystal without
bouncing into the lattice ions at all.

New Schrödinger equation: choose k to lie in the first Brillouin

zone. -> need to find the coefficients to determine the wave
function

Nearly free electron model

1) Instructive solution: Solve SE for a one dimensional solid with lattice constant a, U=0 (treating free
electrons but with the symmetry of the lattice)
2𝜋
Reciprocal lattice is spanned by vectors of length: 𝑔 = 𝑎

Require that only ck, ck-g and ck+g are

different from zero
=> Linear System of three solutions for
every value of k.

The three parabolas are identical to the free electron result

The periodicity has the effect that the parabolas appear to be backfolded at the Brillouin zone
boundary such that the second lowest band in the 1st Brillouin zone is essentially the same band as
the first, but originating in the neighboring zone.
2) Turn on the lattice potential U

opening of gaps between the parabolas at the Brillouin zone boundary

solid has no longer a continuum of states from the lowest energy to infinity.
We obtain a band gap, a range of energies for which there are no states at
all.
En(k) = Dispersion relations = electronic band structure of the solid.
n: band index (here we see the two lowest bands and a part of the third
band).

a free electron traveling perpendicular to a set

of lattice planes separated by a distance a has
the wave function:
Such a wave fulfills the Bragg condition for a
𝑛𝜋
value of 𝑘 = 𝑎 -> lattice will reflect the wave
back to some degree

back reflected wave and forward moving wave:

Both represent standing electron waves with

probability densities |ψ(+)|2 and |ψ(-)|2 which are
shifted with respect to the positive ion potential.

ψ(+) shows an accumulation of probability near the ion cores, whereas ψ(-) shows a depletion.
Therefore ψ(+) has a lower energy than ψ(-).
ψ(+) and ψ(-) correspond to the solution just below and above the energy gap at the Brillouin zone
boundary.

The group velocity is given by the slope of the bands.

Energy bands in real solids (only discuss fcc Bravais lattice)

Lattice periodic potential has 2 main effects:
1) allows us to consider the dispersion in the first Brillouin zone only-> back-folding of the bands at
the Brillouin zone boundary
2) gap openings between bands

band structure of solids can be determined experimentally by angle resolved photoemission

spectroscopy: sample is exposed to monochromatic ultraviolet photons and electrons are emitted
(photoelectric effect), the emitted electrons are sorted according to their k-vector and energy-> work
back to k and the energy inside the sample -> band structure

Bands of Al (3 electrons per unit cell) in

different high-symmetry directions:
The bands are filled up to the Fermi energy.
Bands crossing the Fermi energy: electrons in
the occupied states just below the Fermi
energy can be excited to states within the
same band just above the Fermi energy:
contribute to the transport of current
Chooses Fermi energy as E = 0. If Al only had 1 electron per unit cell, the Fermi energy would lie
much lower (dashed line). In this case, the situation would be even more free-electron-like. The
bands would cross the Fermi energy at the same k distance from Γ in all directions. The electronis
states at the Fermi energy would therefore constitute a sphere in k-space.

Metal = solid where bands cross the Fermi energy, such that the energy of the electrons in these
bands can be increased by a very small amount.

Band structures of Si (insulator):

For the very lowest energies, the bands still look like parabolas.
Band gap openings at the Brillouin zone boundaries and back-
folded bands can also be identified.
Difference to Al:
- existence of an absolute band gap (gap in the entire BZ-> there
are no states at any k in the gray regions)
- When the states are occupied, the bands below the gray band
gap are exactly filled, there are no electrons left for the bands
above the gap-> Fermi energy somewhere inside the gap region

There are no bands crossing the Fermi energy and no electrons that could increase their energy.

Highest energy
we fill up to
lies at a finite
density of
states -> metal
(band is only
partially filled
with electrons)

There are as many states in the band as there are unit cells (atoms) and each state can accommodate
2 electrons:

Material with an odd number of electrons per unit cell = metal

example: Al has 3 electrons per unit cell (fcc: 1 atom / unit cell, Al: 3 valence electrons /atom). Two of
these can completely fill one band, leaving another band half-filled.

An even number of valence electrons per unit cell does not imply that the material is a
semiconductor/insulator. The electrons could be distributed into different bands.
example: two electrons per unit cell could be either placed in one completely filled band or in two
different bands that overlap in energy.

Special case: graphene (boundary between a metal and an insulator)

Carbon has 4 valence electrons. Graphene has 2 atoms per unit cell-> 4 completely occupied bands.
But there is no band gap between the occupied and unoccupied band.

Transport properties
A Bloch electron does not scatter off the lattice ions at all-> but the conductivity of metals if finite!
The lattice is not perfect-> lattice vibrations at higher temperature-> scattering of Bloch electrons

An electric field causes the Bloch electrons to change their k.

 for a partially filled band:
When an electric field is turned on, the electrons change
their k by dk-> asymmetric distribution.
There will be an inelastic scattering mechanism with a
relaxation time, which brings the accelerated electrons close
to kf to unoccupied states at lower energy, close to – kf.
= stationary state: most electrons have ended up in states
that had been occupied by others before.
The only points where the change is important are around
the Fermi energy crossings kf and - kf -> current is flowing

for electrons in a full band:

As the field is applied, electrons are also move by dk but this
does not change the situation at all because all electrons
move into states that were also occupied before the field
was turned on.
We increase ħk for all electrons, but the average momentum
is still zero -> ħk cannot be interpreted as the momentum of
the Bloch electrons.

Consider the acceleration of an electron initially traveling

with the group velocity vg:

The particles have an effective mass (Quasiparticle):

m* can be much smaller or much bigger than the free electron mass (it also can be negative). For
free electrons m* = me.

 The electrons in the periodic solid can be treated quite similarly to free electrons by a quasi-
classical theory if we replace the free electron mass by an effective mass, which contains the
information about the solid’s band structure.

Semiconductors
 Idea: band gap between the highest occupied states and the lowest unoccupied states is
sufficiently small to get thermally excited electrons at reasonable temperatures (band gaps
below 3 eV). -> insulators if the gap is larger.
 Most common semiconductors are the group IV elements Si and Ge with their characteristic
tetrahedral sp3 bonding.

Intrinsic Semiconductors
Def.: a solid for which the chemical potential at zero temperature is placed at an energy where the
density of states is zero. There is a certain energy band that is completely filled (highest occupied
band = valence band VB), and the next band at higher energy is completely empty (lowest
unoccupied band = conduction band CB).
At finite temperature-> electron distributed according to Fermi-Dirac distribution-> leads to some
missing electrons in the VB and some extra electrons in the CB. -> must have as many extra electrons
in the CB as missing electrons in the VB-> determines the position of the chemical potential
The chemical potential must lie close to the middle of the gap between the VBM (valence band
maximum) and CBM (conduction band minimum), otherwise the solid would be charged.

Temperature Dependence of the Carrier Density

Charge carriers: electrons and holes

Electron density n in the CB: Density of holes/missing electrons in the VB:

Since the Fermi-Dirac distribution falls to zero very rapidly away from the
chemical potential, most occupied electrons can be found very close to
the CBM and most holes very close to the VBM-> band dispersion close to
these points is nearly parabolic.
EV = 0, EC = Eg (gap size)

Band structure of GaAs:

Direct band gap: both, the VBM and

the CBM are found at Γ , that is, at k = (0,0,0)

Indirect band gap: Compare to Si (p. 15) only the VBM is found at
Γ and the CBM is somewhere between Γ and X

For us know, the position in k is not important, we are only

interested in the carrier densities.

(−𝑒)𝐸
The electrons near the CBM are accelerated in an external electric field according to 𝑎 =
𝑚𝑒 ∗
Since me* is positive, this is the normal motion of electrons in an electric field. In the VBM the
(−𝑒)𝐸 (𝑒)𝐸
electrons are accelerated the same way but now the effective mass is negative: 𝑎 = −|𝑚 ∗ = 𝑚ℎ ∗
ℎ |
We can interpret it as the motion of holes (positive carriers with a positive effective mass).

Dispersion Density of states (p.12)

CB

VB

If the chemical potential is close to the middle of the gap, and the gap size is at least a few hundred
meV and if we are interested in the properties of the material around room temperature: (E-µ)>>kbT
=> approximate the Fermi-Dirac Districution (replace with a shifted classical Boltzmann distribution).

CB: VB:

Now we can solve the integrals:

Whole band character appears to be lost (still hidden in the effective masses) and it is quite sufficient
to think of the VB and CB as two discrete energy levels.
Useful relationship: law of mass action

 The product of electron and hole concentrations is constant at any given temperature,
independent of chemical potential’s position.

Position of the chemical potential:

For zero temperature the chemical potential does lie in the middle of the gap. Even at finite
temperature, it remains in the middle of the gap, as long as the effective masses for holes and
electrons are equal. To calculate the position when 𝑚ℎ ∗ ≠ 𝑚𝑒 ∗ , we need the gap size Eg and the
effective masses.
gap size: a semiconductor cannot absorb light if the photon energy hν is smaller than the gap size Eg.
Only if hν exceeds Eg, electrons from the VB can be excited into the CB.
effective masses: cyclotron resonance

Doped Semiconductors
The carrier concentrations in most intrinsic semiconductors are too low to give any appreciable
conductivity at room temperature-> adding a very small amount of electrically active impurities
The dopant atoms change the conductivity by either donating (giving) electrons to the CB or
accepting (taking) electrons from the VB (generating holes). They are called donors and acceptors.

N and p Doping (two types of doping)

n doping: putting donor atoms into the lattice that give rise to free electrons in the CB.
p doping: putting acceptor atoms into the lattice that give rise to free holes in the VB.

n doping in Silicon
donor atoms: P, As, Sb with valence
configuration s2p3-> 4 of these 5 electrons
form the sp3 hybrid. The remaining electron
remains loosely bound to the dopant ion,
attracted by one positive net charge. This extra
electron has a very small binding energy and
can therefore easily be removed by thermal
excitations.

Doping reduce the binding

energy to the order of Ed = 40
meV

p doping in Silicon
acceptor atoms: B, Al, Ga
Carrier Density
n-doped semiconductor:

- Low temperature << Ed/kb: none of the donor atoms is ionized-> µ situated between CBM
and the donor level Ed
- raise temperature: donors are ionized, their electrons are moved to the CB
- extrinsic region: some intermediate temperature (room temperature): all donors are ionized
and the carrier concentration in the CB corresponds to the concentration of the donors-> µ
moves toward the middle of the gap
- intrinsic region: At much high temperature: excitation of carriers across the band gap
becomes important, leading again to a strong increase of the carrier concentration. -> µ is
close to the middle of the gap

np = const.: electrons and holes annihilate each other when they meet, since an electron in the
CB can gain energy by taking the place of a hole in the VB-> increasing the number of electrons
by n doping, the number of holes in the VB has to be lower

majority carriers: the carriers which are enhanced by doping

minority carriers: the other types of carriers that are still there, but in a small concentration
(important for transistors)

The density and the sign of the carriers can be measured using the Hall effect:
positive RH-> being caused by the holes in the VB
metals: sign of RH is determined by the details of the band structure. If the Fermi energy of a
metal is placed just below a band maximum similar to the VBM in a semiconductor, the electrons
in that metal will behave like holes in a Hall measurement-> show a positive RH

Conductivity of Semiconductors
Metals: -> replace the free electron mass with the effective mass
conductivity decreases as T is raised: more scattering by phonons

semiconductor:

At higher temperature, the carrier densities (n or p or both) increase stronger as the decrease of the
mobility. -> increased conductivity at higher temperature.
(For most semiconductors, the electron effective mass is smaller than the hole effective mass and
therefore, a higher conductivity can be achieved by n doping than by p doping).

The pn junction
At not too high temperatures:
p-doped: µ is close to the VB, most acceptors are negatively charged, majority carriers: holes
n-doped semiconductor: µ is close to the CB, most donors are ionized, majority carriers are electrons
joining: Electrons from the n side diffuse
into the p side and holes from the p side
diffuse into the n side. When mobile
electrons and holes meet, they recombine.
What is left is a region of immobile ionized
donors and acceptors without the
compensating charge of the generated
carriers: depletion layer
They give rise to an E-field between the two sides which represents an obstacle for the p holes to
move into the n part and for the n electrons to move into the p part. The filed increases as the
depletion layer widens, and the process goes on until equilibrium is reached.

The chemical potential has to be constant in the whole

system-> a macroscopic potential shifts the energy levels
and bends the bands.

Magnetism
Macroscopic Description
- Gauss law for magnetostatics is obeyed:
- There are no magnetic monopoles
- The sources of the magnetic induction B are magnetic dipoles
- In vacuum, the magnetic induction is related to the magnetic field H by:
- In matter we have:

M = macroscopic magnetization
number N of magnetic dipole moments µ per volume V
B0 = external field
- Linear relation between the external field and the magnetization:
(does not always hold ->ferromagnetism)
𝜒𝑚 = magnetic susceptibility relative permeability µ = 1 + 𝜒𝑚
- negative: diamagnetic
- positive: paramagnetic
- The potential energy U of a magnetic dipole µ in a field B0 : U = - µ B0 = - V M B0
When M depend on the field: energy change for a small increase dB0-> integral
For a paramagnetic solid: U is negative (energy decrease for higher field), experience a force
toward locations of higher magnetic fields, they are attracted to either pole of a permanent
magnet.
For diamagnetic solids: they are expelled from regions of high magnetic fields.
Diamagnetic is always present, paramagnetism can only be observed when the atoms in the
solid show a net magnetic moment, already without an external field. Such magnetic
moments can align with the external field, leading to an energy gain.

Quantum Mechanical Description of Magnetism

 How does the energy of an electron changes when a weak magnetic field is included?
 How does the Schrödinger equation change in the presence of a electromagnetic field?
scalar potential: E(r) = -grad 𝜙(𝑟)-> only acts as an addition to the potential that it already present
Vector potential: B = curlA-> only affects the kinetic energy term of the electron
p -> p – q A A = -1/2 r x B p=-iℏ∇

=>

The second term represents the diamagnetism (term is always positive)

The first term contains the angular momentum of the electron projected onto the direction of the
field (z). This is the paramagnetic term that gives an energy-lowering when the electron’s magnetic
moment aligns with the field (when the z-component of the angular momentum is negative->
pointing in the opposite direction from the field). E.g. electron

Paramagnetism and Diamagnetism in Atoms

The magnetic moment of an atom is caused by the orbital
and spin angular momenta.

For atoms with filled shell, the total magnetic moment is zero (components in field direction and
opposite to it are equally strong, there are equally many electrons with spin +1/2 and spin -1/2.

For nonfilled shells, we have to find the total angular momentum J and calculate the magnetic
moment associated with J.
Atoms or ions show paramagnetic behavior only when they have open shells. We need atoms with
open shells, that do not participate in the bonding and therefore do not change their properties
much upon the formation of a solid
Diamagnetism is always there because it is caused by all the electrons (valence & core electrons) in
the atoms and their reaction to the magnetic field.

Weak Magnetism in Solids

Pauli Paramagnetism
 difference between electrons with their magnetic moment
parallel and antiparallel to the field

net magnetization:

The field required to align all the spins have to be such that: 𝜇𝐵 𝐵0 > 𝐸𝐹
exceed the Fermi energy!

Curie Paramagnetism
For many paramagnetic materials, the magnetization of the material is directly proportional to an
applied magnetic field, for sufficiently high temperatures and small fields. However, if the material
is heated, this proportionality is reduced. For a fixed value of the field, the magnetic susceptibility is
inversely proportional to temperature, that is
𝐶
𝜒𝑚 =
𝑇
In the regime of low temperatures or high fields, M tends to a maximum value of nµ, corresponding
to all the particles being completely aligned with the field.

Temperature-dependent magnetization below the Curie

temperature:
Above the Curie temperature, the spontaneous magnetization
is lost.

Magnetic Ordering
 We look at long-range magnetic ordering without any applied field.

Ferromagnetic ordering: originates

from the parallel alignment of the
magnetic moments in the crystal
Antiferromagnetic ordering:
orientation of the moments on
neighboring sites is opposite
Ferrimagnetic ordering:
antiferromagnetic ordering
between moments of different
sizes in one unit cell, ferromagnetic
ordering between the unit cells.
Where does the magnetic ordering come from?

Exchange interaction: type of energy that stems from a combination of Coulomb interaction and the
Pauli principle (need to have antisymmetric wave functions for fermions)
X = exchange energy: separation between the singlet and triplet state
The magnetic dipole moments order because of the exchange interaction that favors either a parallel
(positive X) or an antiparallel (negative X) alignment.
Hysteresis
Beginning: M = 0
For a certain field strength, the sample is completely
magnetized in one direction (magnetization MS is
reached.)
When the field is lowered again, the magnetization
decreases but for B0 = 0 is has not reached zero but
remanent magnetization MR. A magnetization of M = 0
is first reached at the coercive field BC in the opposite
direction.

Superconductivity
 Current can flow in materials with zero resistance-> a quantum phenomenon

Zero Resistivity
Resistivity is vanishing below a certain critical temperature TC.

Metal: resistivity decreases as the temperature is lowered (eventually levels off to a constant value)
It’s caused by imperfections in the lattice (lattice defects, thermal vibrations). For temperatures
much lower than the Debye temperature, the vibrations are frozen in but even at zero temperature,
impurities and defects are present and hence the resistivity is expected to remain finite.

Superconductor: Above the critical temperature TC, which is quite low for all elements, the same
behavior as for a normal metal is observed. At TC however, the resistivity drops to an unmeasurably
small value.

There isn’t any certainty that the other

elements are not superconducting. It is
possible that the necessary low
temperature has just not been reached
yet.
As we see, metals that are good
conductors (like Ag and Cu) are not
necessarily also superconductors.
Ferromagnetic elements are not
superconducting

High pressure could possibly change the structure of a crystal to turn it into a superconductor. But
it’s the electronic structure and the vibrations that count in the superconducting phase transition. If
the elements do crystallize in different structures, these structures also have different electronic
structures and different vibrational properties.
Another possibility to achieve superconductivity is to grow them in ultrathin films or as small clusters
of atoms.

Apart from the temperature, a magnetic field and a current through the sample can also destroy the
superconducting state:
For too high current density, the superconductivity breaks down-> critical current density jC
For too high magnetic field or too high temperature, the superconductivity is lost. For a given
temperature T > TC, the critical magnetic field BC(T) is:
The Meissner Effect
A superconductor shows perfect diamagnetism in its superconducting state 𝜒𝑚 = -1
𝐵
(superconductors are strongly expelled from magnetic fields) and therefore a magnetization M = − 𝜇0
0
which totally cancels the external field B0 inside the superconductor. The origin of M is very different
from that in normal diamagnetic materials where diamagnetism is caused by microscopic magnetic
moments that are induced by the external field throughout the entire solid. In a superconductor M is
caused by macroscopic supercurrents that flow close to the surface of the specimen and keep the
inside field-free.

The Meissner effect is not simply a consequence of the vanishing resistance. There is a difference
between a material that becomes merely a perfect conductor with ρ = 0 below TC and a true
superconductor that also displays the Meissner effect.

Perfect conductor:
below TC: ρ-> 0
magnetic field also turned on for the
conductor that was in the field-free region
before-> Faraday’s law: it expels the field
entirely and the inner part remains field-
free. For the other nothing changes because
the filed had been penetrating already
before the transition below TC.
B-> 0: A current is induced, which keeps the
filed in the specimen as it was before.

 We can have any field we like below TC. It only depends on the history and the situation.

Superconductor:
below TC: The magnetic field is expelled in both cases by a supercurrent at the surface of the
superconductor-> the interior of the superconductor is field-free, independent of the history of the
sample. Once the field is turned off, the interior remains field-free.

Isotope Effect
TC depends at all on M!
Born-Oppenheimer approximation: electronic structure of a solid
should not depend on the mass of the ions
The vibrational properties do depent on the mass of the ions.
=> The lattice vibrations of the solid play some role in superconductivity

Microscopic BCS Theory

Two of the most fundamental assumptions for the treatment of solids are not valid in the case of
superconductivity:
1) Born-Oppenheimer approximation that allowed us to treat the electronic and the vibrational
properties separately
2) The independent electron approximation (considering the properties of one electron in the
average potential of all others).
A macroscopic wave function is realized by very many particles occupying the same quantum sate->
particles cannot be the free electrons (fermions-> Pauli principle)
=> Formation of electron pairs->new ground state of the electron gas
A possible mechanism providing a weak attraction between electrons is the interaction between the
electrons and lattice vibrations (electron-phonon interaction). When an electron Is placed
somewhere in the ionic lattice, it slightly deforms the bonds around it due to the electrostatic
interaction with the ions. The deformation is such that the lattice ions are attracted toward the
position of the electron. This leads to a local polarization in the lattice, which in turn is attractive for
other electrons.
What’s about repulsion from the first electron?
The electron moves through the lattice with a high speed (at
the Fermi energy) and since the ions move much slower, the
polarization is not established instantaneously. By the time
the maximum polarization is reached, the electron has already
traveled around 100 nm. Now its possible for a second
electron to lower its energy by moving in the trail of the first.
The energetically most favorable situation arises when two electrons have exactly opposite k vectors
and opposite spins (total spin of zero) -> Cooper pairs = bosons

The interaction between two electrons via phonons can be viewed as the constant emission and
absorption of “virtual” phonons of energies up to ℏωD. The only electrons that can participate in the
exchange of such phonons and the formation of Cooper pairs are those close to the Fermi energy.
None of the other electrons can emit or absorb phonons because they are trapped by the Fermi-
Dirac distribution (all reachable states around them are already occupied).

For the superconductor there are no

single-particle states close to the
Fermi energy anymore. The electron
states just below EF have been
removed to form the Cooper pairs.
Above EF the lowest possible energy
for a single electron is ∆.
Now the minimal excitation energy
for an unpaired electron is not zero
like in a normal metal but 2∆
The gap is strongly temperature
dependent.

The BCS theory predicts that ∆ at zero temperature is related to the critical temperature TC as
∆= 3.53𝑘𝐵 𝑇𝐶 and

ħ𝜔𝐷
Where Θ𝐷 = 𝑘𝐵
= Debye temperature-> ∆ is proportional to ħ𝜔𝐷

The transition temperature is proportional to the Debye temperature

-> explains the isotope effect: Θ𝐷 ∝ 𝜔𝐷 ∝ 𝑀1/2
-> sets the temperature scale of the phenomenon: the exponential cannot become larger than 1

At finite temperatures, not all the electrons close to EF are bound in Cooper pairs. The formation
energy 2∆ for a Cooper pair depends on the number of pairs already in the ground state and thus
∆ becomes a function of temperature (compare
magnetic ordering: energy gain for the
orientation of a spin depends on the number of
spins already present in orientation)
At the transition temperature TC, the gap is
closed. Upon lowering the temperature further,
the number of Cooper pairs (also the magnetic
alignment) are increased.
At zero temperature, all the electrons close to
EF are bound in Cooper pairs.

When an electric field is applied, the whole condensate of pairs is accelerated. When a current
passes through the sample, all the pairs are still in the same state but with a different k’ ≠ 0 (before
all were in the same state with k = 0). In a normal metal, the current decays because the electrons
are scattered back (individual electrons) to lower lying states by defects or thermal vibrations. For
Cooper pairs its not possible to change k for one pair without changing it for all the others at the
same time. Split a Cooper pair into 2 individual electrons is only possible if the kinetic energy of the
pairs is high enough to provide the necessary 2∆ for the destruction. When the critical current
density jC is reached, they break up altogether (the same for the critical magnetic field BC). jC and BC
decrease as TC is approached because of the shrinking gap size and the lower density of Cooper pairs.

Experimental Detection of the Gap

- Single electron tunneling

small positive tunneling voltage:

no current can flow on the superconductor because the electrons
from the metal do not find empty states
tunneling between to metals: already possible at very small voltage
(dashed line)
As the tunneling voltage reaches ∆/e the electrons close to the
Fermi energy find empty states in the superconductor and tunneling
becomes possible.

- Optical reflectivity: measure the absorption of e-m-radiation as it passes through a

superconductor. As the photon energy reaches 2∆ , absorption sets in an the transmitted
intensity is strongly reduced-> existence of a gap
- Low-temperature heat capacity

There is a qualitatively different behavior at very low temperatures,

where the superconductor shows a exponential decrease
comparing to the linear temperature dependence of a metal.
(compare with the exponential behavior of the Einstein model for
the phonon heat capacity where it is also caused by an energy gap
between the ground state and the first excited state of the Einstein
oscillators)
Type I and Type II Superconductors
Type I:
The magnetic field cannot enter a material in its superconducting state. The superconductivity breaks
down above a certain critical field BC.
Type II:
The magnetic field can enter the bulk sample while the material still remains in its superconducting
state. There are two critical fields BC1 and BC2. Below BC1 the material behaves as Type I and above
BC2 the superconductivity is destroyed. Between these two fields, the magnetic field partly enters
the material. There is a finite magnetization but not large enough to compensate the external field.
The resistivity is still zero between these two critical fields! Type II superconductors can tolerate a
higher field-> makes it important for technical applications.

Connection between superconductivity and magnetic ordering

They are exclusive?
But they have in common that the electrons with the highest energy, those at the Fermi energy, have
a possibility to slightly lower their energy, either by establishing magnetic order or by condensing
into Cooper pairs.