Band Theory1

1
Energy Bands in Solids
Prof.P. Ravindran,
Department of Physics, Central University of Tamil
Nadu, India
http://folk.uio.no/ravi/CMP2013
P.Ravindran, PHY075- Condensed Matter Physics, Spring 2013 : Energy Bands in Solids
Electrons in Solids
 Considering electrons in metals as free particles, electron gas

in a box explains many experimental results which?
 Still, why are some solids metals and others insulators?
– For metals we assumed that electrons are free
– We know, however, that there are no free electrons in
insulators
 QM give us the answer!
– Need a more realistic potential for electrons reflecting
periodic ionic structure of solids
Splitting of Energy levels
 If the atoms are far apart- atoms can be considered as

isolated
 Under the influence of neighbouring atoms- splitting of
energy levels occurs
 The width of K-shell is less as compared to L-shell and so
on
 Effect of distance on splitting of energy levels
E3 (M shell)
E2 (L shell)
Energy
E1 (K shell)
Energy Band
Band Energy
Forbidden Energy
Gap
Energy Band
ENERGY BANDS IN SOLIDS
 Explains and distinguish between conductors, insulators and

semi conductors
 Conduction Band – small energy can remove an electron from
an atom
 Valence Band – Below the conduction band
 Filled Band- Near the nucleus
 Insulators- Conduction band is empty but valence is almost
completely filled+ wide energy gap between the two bands
(Poor conductivity)
 Conductors- Overlapping of both conduction and valence
bands
 Conduction es move almost twice as fast as holes
“Realistic” Potential in Solids
  
U(r )  U(r  T )
   
T  n1a  n2 b  n3 c
– ni are integers
 Example: 2D Lattice
  
T  n1a  n2 b
n1  2; n2  3
 For one dimensional case where atoms (ions)
are separated by distance d, we can write
condition of periodicity as
U( x)  U( x  d  n)
 Multi-electron atomic potentials are complex

 Even for hydrogen atom with a “simple”
Coulomb potential solutions are quite
complex
 So we use a model one-dimensional periodic
potential to get insight into the problem
Bloch Functions
Felix Bloch showed that the actual
solutions to the Schrödinger equation for
electrons in a periodic potential must have
the special form:

  i ( kr )
 k ( r )  uk ( r )e
where u has the period of the lattice, that
is
  
uk ( r )  uk ( r  T)
Bloch’s Theorem
 Bloch’s Theorem states that for a particle

moving in the periodic potential, the
wavefunctions ψ(x) are of the form
 ( x)  uk ( x)e  ikx , where uk ( x) is a pe riodicfunction
u k ( x)  u k ( x  d)
 uk(x) is a periodic function with the periodicity

of the potential
– The exact form depends on the potential
associated with atoms (ions) that form the solid
Bloch’s Theorem
 Bloch’s Theorem imposes very special

conditions on any solution of the
Schrödinger equation, independent of the
form of the periodic potential
 The wave vector k has a two-fold role:
1. It is still a wave vector in the plane wave part of
the solution
2. It is also an index to uk(x) because it contains all
the quantum numbers, which enumerate the
wavefunction
Bloch’s Theorem
 What is probability density of finding

particle at coordinate x?
P( x)   ( x)   ( x) ( x)
2 *
 ikx *  ikx
P( x)  [uk ( x)e ] [uk ( x)e uk (x)
2 ]
 ikx  ikx
P( x)  uk ( x)uk ( x)e  uk ( x)uk ( x)
* *
e
P( x)  uk ( x)
2
• But |uk(x)|2 is periodic, so P(x) is as well

Bloch’s Theorem
P( x)  P( x  d)
The probability of finding an electron at
any atom in the solid is the same!!!
 Each electron in a crystalline solid

“belongs” to each and every atom
forming the solid
Kronig-Penney Model
U(x)
U0
(a+b) b 0 a a+b x
The wave equation can be solved when the potential is

simple... such as a periodic square well.
Kronig-Penney Model
 2 d 2
 2
 U ( x )  
2m dx
Region I - where 0 < x < a and U = 0
The eigenfunction is a linear combination of plane waves
traveling both left and right:
iKx
  Ae iKx
 Be
The energy eignevalue is:
2 K 2

2m
Kronig-Penney Model
 2 d 2
 2
 U ( x )  
2m dx
Region II - where b < x < 0 and U = U0
Within the barrier, the eigenfunction looks like this
Qx
  Ce Qx
 De
and
 2Q 2
U0   
2m
Kronig-Penney Model
U(x)
U0
II I III
(a+b) b 0 a a+b x
To satisfy Mr. Bloch, the solution in region III

must also be related to the solution in region II.
 III   II eik (a b)

Kronig-Penney Model
A,B,C, and D are chosen so that both the wavefunction
and its derivative with respect to x are continuous at
the x = 0 and a.
At x = 0...
A B  C  D
iK ( A  B)  Q(C  D)
At x = a...
 III   II eik (a b)
iKa Qb ik ( a b)
Ae iKa
 Be  (Ce Qb
 De )e

iK AeiKa  Be iKa  Q CeQb  DeQb eik (a b)   
Kronig-Penney Model
Result for E < U0:
Q2  K 2
sinh Qb sin Ka  cosh Qb cos Ka  cos k (a  b)
2QK
To obtain a more convenient form Kronig and Penney considered the case where
the potential barrier becomes a delta function, that is, the case where U0 is
infinitely large, over an infinitesimal distance b, but the product U0b remains
finite and constant.
Q 2  U0
and also goes to infinity as U0. Therefore:
Lim (Q 2  K 2 )  Q 2
U0 
Kronig-Penney Model
What happens to the product Qb as U0 goes to infinity?
• b becomes infinitesimal as U0 becomes infinite.

• However, since Q is only proportional to the square
root of U0, it does not go to infinity as fast as b goes to
zero.
• So, the product Qb goes to zero as U0 becomes
infinite.
• As a results of all of this...
sin Qb  Qb  1
cos Qb  1
Q  K
Kronig-Penney Model
Q 2ba P
P  sin Ka  cos Ka  cos ka
2 Ka
Kronig-Penney Model
Plot of energy versus wavenumber for the Kronig-Penney
Potential, with P = 3/2.
 
ka ka
0  2 3  0 
Covalent Bonding Revisited
 When atoms are covalently bonded 1

 1s  e  r / a0
electrons supplied by atoms are a03
shared by these atoms since pull of
each atom is the same or nearly so
– H2, F2, CO,
 Example: the ground state of the
hydrogen atoms forming a molecule
– If the atoms are far apart there is very
little overlap between their wavefunctions
– If atoms are brought together the
wavefunctions overlap and form the
compound wavefunction, ψ1(r)+ψ2(r),
increasing the probability for electrons to
exist between the atoms
Schrödinger Equation Revisited
 If a wavefunctions ψ1(x) and ψ2(x) are
solutions for the Schrödinger equation for
energy E, then functions
– -ψ1(x), -ψ2(x), and ψ1(x)±ψ2(x) are also solutions
of this equations
– the probability density of -ψ1(x) is the same as for
ψ1(x)
 2 d  1, 2 ( x)
2
 2
 U( x)[ 1,2 ( x)]  E 1,2 ( x)
2m dx
 2 d 2 [ 1 ( x)  2 ( x)]
 2
 U( x)[ 1 ( x)  2 ( x)]  E[ 1 ( x)  2 ( x)]
2m dx
Band Theory of Solids
 Consider an atom with only one electron in s-state

outside of a closed shell
 Both of the wavefunctions below are valid and the
choice of each is equivalent
  Zr / na0
 s (r )   Af (r )e
  Zr / na0
 s (r )   Af (r )e
 If the atoms are far apart, as before, the
wavefunctions are the same as for the isolated
atoms
 Once the atoms are brought together the
wavefunctions begin to overlap
– There are two possibilities
1. Overlapping wavefunctions are the same (e.g., ψs+ (r))
2. Overlapping wavefunctions are different
 The sum of them is shown in the

figure
 These two possible combinations
represent two possible states of
two atoms system with different
energies
Symmetric and Anti-symmetric Combinations of
Ground State Eigenfunctions
Six States for Six Atom Solid
Splitting of 1s State of Six Atoms
Atoms and Band Structure
 Consider multi-electron atoms:
1. The outer electrons (large n and l)
are “closer” to each other than the
inner electrons
 Thus, the overlap of the wave-
functions of the outer electrons is
stronger than overlap of those of
inner electrons
 Therefore, the bands formed from
outer electrons are wider than the
bands formed from inner electrons
 Bands with higher energies
are therefore wider!
Splitting of Atomic Levels in Sodium
Occupation of Bands Sodium
Splitting of Atomic Levels in Carbon
Occupation in Carbon at Large Atomic
Separation
Actual Occupation of Energy bands in
Diamond
Insulators, Semiconductors, Metals
 The last completely filled (at least at T = 0 K)

band is called the Valence Band
 The next band with higher energy is the
Conduction Band
– The Conduction Band can be empty or partially
filed
 The energy difference between the bottom of
the CB and the top of the VB is called the
Band Gap (or Forbidden Gap)
 Consider a solid with the empty

Conduction Band
 If apply electric field
to this solid, the
electrons in the
valence band (VB)
cannot participate in
transport (no current)
 The electrons in the VB do not
participate in the current, since
– Classically, electrons in the
electric field accelerate, so they
acquire [kinetic] energy
– In QM this means they must
acquire slightly higher energy
and jump to another quantum
state
– Such states must be available, This solid would behave
i.e. empty allowed states as an insulator
– But no such state are available in
the VB!
 Consider a solid with the half filled

Conduction Band (T = 0K)
 If an electric field is applied
to this solid, electrons in the
CB do participate in
transport, since there are
plenty of empty allowed
states with energies just
above the Fermi energy
 This solid would behave as
a conductor (metal)
Band Overlap
 Many materials are
conductors (metals) due to the
“band overlap” phenomenon
 Often the higher energy bands
become so wide that they
overlap with the lower bands
– additional electron energy levels
are then available
Band Overlap
 Example: Magnesium (Mg; Z =12): 1s22s22p63s2
– Might expect to be insulator; however, it is a metal
– 3s-band overlaps the 3p-band, so now the
conduction band contains 8N energy levels, while
only have 2N electrons
– Other examples: Zn, Be, Ca, Bi
Band Hybridization
 In some cases the opposite occurs

– Due to the overlap, electrons from different shells
form hybrid bands, which can be separated in
energy
– Depending on the magnitude of the gap, solids
can be insulators (Diamond); semiconductors (Si,
Ge, Sn; metals (Pb)
 There is a qualitative difference between
metals and insulators (semiconductors)
– the highest energy band “containing” electrons is
only partially filled for Metals (sometimes due to
the overlap)
 Thus they are good conductors even at very low
temperatures
 The resistivity arises from the electron scattering from
lattice vibrations and lattice defects
 Vibrations increases with temperature  higher
resistivity
 The concentration of carriers does not change
appreciably with temperature
 The difference between Insulators and
Semiconductors is “quantitative”
– The difference in the magnitude of the band gap
 Semiconductors are “Insulators” with a
relatively small band gap
– At high enough temperatures a fraction of electrons
can be found in the conduction band and therefore
participate in transport. Hence resistivity decrease
with increase of temperature in semiconductor.
Insulators vs Semiconductors
 There is no difference between Insulators and

Semiconductors at very low temperatures
 In neither material are there any electrons in the
conduction band – and so conductivity vanishes in
the low temperature limit
Insulators vs Semiconductors
 Differences arises at high temperatures

– A small fraction of the electrons is thermally
excited into the conduction band. These
electrons carry current just as in metals
– The smaller the gap the more electrons in the
conduction band at a given temperature
– Resistivity decreases with temperature due to 1 q 2
n
higher concentration of electrons in the  
conduction band in semiconductore.  m*
Holes
 Consider an insulator (or semiconductor)

with a few electrons excited from the valence
band into the conduction band
 Apply an electric field
– Now electrons in the valence band have some
energy sates into which they can move
– The movement is complicated since it involves
~ 1023 electrons
Concept of Holes
 Consider a semiconductor with a small number of
electrons excited from the valence band into the
conduction band
 If an electric field is applied,

– the conduction band electrons will participate in the
electrical current
– the valence band electrons can “move into” the empty
states, and thus can also contribute to the current
Holes from the Band Structure Point of View
 If we describe such changes via

“movement” of the “empty” states – the
picture can be significantly simplified
 This “empty space” is a Hole
– “Deficiency” of negative charge – holes are
positively charged
– Holes often have a larger effective mass
(heavier) than electrons since they represent
collective behavior of many electrons in the VB
(i.e. low lying energy band)
Holes
 We can “replace” electrons at the top of eth
band which have “negative” mass (and
travel in opposite to the “normal” direction)
by positively charged particles with a
positive mass, and consider all phenomena
using such particles
 Such particles are called Holes
 Holes are positively charged and move in
the opposite direction as electrons “they
replace”
Hole Conduction
 To understand hole motion, one requires
another view of the holes, which represent
them as electrons with negative effective
mass
 To imagine the movement of the hole think of
a row of chairs occupied by people with one
chair empty
 To move all people rise all together and
move in one direction, so the empty spot
moves in the opposite direction
Concept of Holes
 If we describe such changes via “movement”
of the “empty” states – the picture will be
significantly simplified
 This “empty space” is called a Hole
– “Deficiency” of negative charge can be treated as
a positive charge
– Holes act as charge carriers in the sense that
electrons from nearby sites can “move” into the
hole
– Holes are usually heavier than electrons since
they depict collective behavior of many electrons
Conduction
Electrical current for holes and electrons in the opposite direction

What are quantum numbers?
 Quantum numbers label eigenenergies and eigenfunctions of a
Hamiltonian
Sommerfeld: k-vector ( k is momentum)
Bloch: k-vector ( k is the crystal momentum)

and n (the band index).
•The Crystal Momentum is not the Momentum of a Bloch electron: the rate of change
of an electron momentum is given by the total forces on the electron, but the rate of
change of electronic crystal momentum is:
dr 1  n (k ) d k
 v n (k )   nk v  nk 
ˆ ;  e E(r, t )  v n (k )  B(r, t ) 
dt k dt
where forces are exerted only by the external fields, and not by the periodic field of the
lattice.
Semiclassical dynamics of Bloch electrons
•Bloch states have the property that their expectation values of r and k , follow
classical dynamics. The only change is that now  n (k ) (band structure) must be
used: e
H classical   n  k  eA (r )   e (r )  B  Lk
2m
dp d ( k ) H dr H 
  , v 
dt dt r dt p p
•A perfectly periodic ionic arrangement has zero resistance. Resistivity comes from
imperfections (example: a barrier induces a reflected and transmitted Bloch wave),
which control the mean-free path. This can be much larger than the lattice spacing.
•A fully occupied band does not contribute to the current since the electrons cannot be
promoted to other empty states with higher . The current is induced by
k
rearrangement of states near the Fermi energy in a partially occupied band.
•Limits of validity: n  const (no interband transitions)
 gap
2
(k ) eB  gap
2
(k )
eEa , c  ,  gap (k )   n (k )   n (k )
F m F
  gap ,  a
What is the range of quantum numbers?
Sommerfeld: k
runs through all of k-space consistent with the Born-von
Karman periodic boundary conditions:
 ( x  L, y , z )   ( x , y , z ) 
 2
 ( x  L, y, z )   ( x, y  L, z )   k   nx , n y , nz   nx , n y , nz  0, 1, 2,
 L
 ( x, y , z )   ( x, y , z  L ) 
 ( x  L, y , z )   ( x , y , z )  0 
 
 ( x  L, y, z )   ( x, y  L, z )  0   k   nx , n y , nz   nx , n y , nz  1, 2,
 L
 ( x, y , z )   ( x, y , z  L )  0 
Bloch: For eachn k , runs through all wave vectors in a single

primitive cell of the reciprocal lattice consistent with the Born-
von Karman periodic boundary conditions; n runs through an
infinite set of discrete values.
What are the energy levels?
Sommerfeld: 2 2
k
 (k ) 
2m
Bloch: For a given band index n, n (k ) has no simple
explicit form. The only general property is periodicity in
the reciprocal space:
 n (k  G)   n (k )
What is the velocity of electron?
Sommerfeld: The mean velocity of an electron in a
level with wave vector is: k
k 1 
v 
m k
Bloch: The mean velocity of an electron in a level with
band index n
and wave vector is: k
Conductivity of a perfect crystal: 1  n (k )
  v n (k ) 
k
NOTE: Quantum mechanical definition of a mean velocity

ˆ    dr  (r )
v  v  (r ) *
mi
What is the Wave function
Sommerfeld: The wave function of an electron with

wave vector is: k
1
k  (r ) 
ikr
e
V
Bloch: The wave function of an electron with band index n
and wave vector is: k
k (r)  e nk (r ) ikr
where the function nk (r ) has no simple explicit form. The

only general property is its periodicity in the direct lattice
(i.e., real space): nk (r  R )  nk (r )
Sommerfeld vs. Bloch: Density of
States
Sommerfeld → Bloch
2
D( )   dk     (k )   D( )    dk     (k ) 
2
 2   2 
d d n
n B. Z .
Bloch: van Hove singularities in the DOS
2 2 dS
D( )    dk     (k )   n  dk   (k )
 2   2 
d n d
n B.Z . Sn ( E ) k n
1
D( )d  
 2  d
dS k
d     k  n (k )  Δk   k  n (k ) k
D( )
63
Sommerfeld vs. Bloch: Fermi surface
•Fermi energy  F   (T  0) represents the sharp occupancy

cut-off at T=0 for particles described by the Fermi-Dirac statitics.
•Fermi surface is the locus of points in reciprocal space where  (k )   F
F No Fermi surface for

insulators!
kF kF
Points of Fermi
“Surface” in 1D
Sommerfeld vs. Bloch: Fermi surface in 3D
Sommerfeld: Fermi Sphere
Bloch: Sometimes sphere, but more likely anything else
For each partially filled band there will be a surface reciprocal space separating occupied
from the unoccupied levels → the set of all such surfaces is known as the Fermi surface
and represents the generalization to Bloch electrons of the free electron Fermi sphere.
The parts of the Fermi surface arising from individual partially filled bands are branches
of the Fermi surface: for each n solve the equation in
 n (k )   F
variable.
k
Is there a Fermi energy of intrinsic Semiconductors?
•If  F is defined as the energy separating the

highest occupied from the lowest unoccupied
level, then it is not uniquely specified in a solid
with an energy gap, since any energy in the gap
meets this test.
•People nevertheless speak of “the Fermi energy”
on an intrinsic semiconductor. What they mean is
the chemical potential, which is well defined at any
non-zero temperature. As T  0, the chemical
potential of a solid with an energy gap approaches
the energy of the middle of the gap and one
sometimes finds it asserted that this is the “Fermi
energy”. With either the correct of colloquial
definition,  n (k )   F does not have a solution in a
solid with a gap, which therefore has no Fermi
surface
66
DOS of real materials: Silicon, Aluminum, Silver
Colloquial Semiconductor “Terminology” in Pictures
←PURE
DOPPED→
Measuring DOS: Photoemission spectroscopy
Fermi Golden Rule: Probability per unit

time of an electron being ejected is
proportional to the DOS of occupied
electronic states times the probability
(Fermi function) that the state is occupied:
1
I ( kin )   D( bin ) f ( bin )  D( kin     ) f ( kin     )
 ( kin )
Measuring DOS: Photoemission spectroscopy
Once the background is subtracted

off, the subtracted data is proportional
to electronic density of states
convolved with a Fermi functions.
We can also learn about DOS above the Fermi
surface using Inverse Photoemission where electron
beam is focused on the surface and the outgoing flux
of photons is measured.
In isolated atoms the

electrons are arranged
in energy levels
In solids the outer electron energy levels become
smeared out to form bands
The highest occupied band is called the VALENCE band.

This is full.
For conduction of electrical energy there must be
electrons in the CONDUCTION band. Electrons are free
to move in this band.
Insulators : There is a big energy gap between the
valence and conduction band. Examples are plastics, paper
…..
Conductors : There is an overlap between the valence and

conduction band hence electrons are free to move about.
Examples are copper, lead ….
Semiconductors : There is a small energy gap between

the two bands. Thermal excitation is sufficient to move
electrons from the valence to conduction band. Examples
are silicon ,germanium….
When a conductor is But in a semiconductor
heated its resistance the resistance
increases ; The atoms decreases with an
vibrate more and the increase in
electrons find it more temperature. Electrons
difficult to move can be excited up to
through the conductor. the conduction band.
Conductivity increases
R R
T T
Doping
This is the replacement of a silicon atom in

the crystal structure with a different atom.
Typically 1 atom per 100 million silicon atoms
is replaced. This changes the electrical
properties.
Silicon is in group 4 of the periodic table. It

has 4 valence electrons.
P type silicon. Replace a silicon atom with a group
3 atom such as aluminium.( 3
valence electrons)
This lowers the number of

electrons in the valence band which
creates a ‘hole’ in the structure.
Electrons can move through the

structure.
This is called p type silicon. P is for
positive as there is a deficiency of
electrons.
N type
Replace a silicon atom with a group

5 atom such as phosphorous ( 5
valence electrons )
The extra electron introduced

here goes into the conduction band
as the valence band is full.
This is n type silicon.

N for negative as ‘additional ‘
electrons have been introduced
into the structure.
Energy Band View of Semiconductors
Conductors, semiconductors, insulators: Why is it that

when individual atoms get close together to form a
solid – such as copper, silicon, or quartz – they form
materials that have a high, variable, or low ability to
conduct current?
Understand in terms of allowed, empty, and occupied

electronic energy levels and electronic energy bands.
Fig. 1. Calculated energy levels in the diamond structure as a function of
assumed atomic spacing at T = 0o K. (From “Introduction to Semiconductor
Physics”, Wiley, 1964)
In Fig. 1, at right atoms are essentially isolated; at left
atomic separations are just a few tenths of a
nanometer, characteristic of atoms in a silicon crystal.
• If we start with N atoms of silicon at the right,
which have 14 electrons each, there must be 14N
allowed energy levels for the electrons. (You
learned about this in physics in connection with the
Bohr atom, the Pauli Exclusion principle, etc.)
• If the atoms are pushed together to form a solid

chunk of silicon, the electrons of neighboring atoms
will interact and the allowed energy levels will
broaden into energy bands.
When the “actual spacing” is reached, the quantum-
mechanical calculation results are that:
• At lowest energies very narrow ranges of energy are
allowed for inner electrons (these are core electrons, near
the nuclei);
• A higher band of 4N allowed states exists that, at 0oK, is
filled with 4N electrons;
• Then an energy gap, EG, appears with no allowed states
(no electrons permitted!); and
• At highest energies a band of allowed states appears that
is entirely empty at 0oK.
Can this crystal conduct electricity?
NO, it cannot conductor electricity at 0o K
because that involves moving charges and
therefore an increase of electron energy – but
we have only two bands of states separated
by a forbidden energy gap, EG. The (lower)
valence band is entirely filled, and the (upper)
conduction band states are entirely empty.
To conduct electricity we need to have a

band that has some filled states (some
electrons!) and some empty states that
can be occupied by electrons whose
energies increase.
Fig. 2 shows the situation at 0o K for (left) a metallic solid
such as copper, and (right) a semiconductor such as
silicon.
The metal can conduct at 0o K because the uppermost

band contains some electrons and some empty available
energy states. The semiconductor cannot conduct – it is
an insulator.
If we raise the temperature of the semiconductor, some

electrons in the filled valence band may pick up enough
energy to jump up into an unoccupied state in the
conduction band. Thus, at a finite temperature, a pure
(intrinsic) semiconductor has a finite electrical conductivity.
Fig. 2. Electronic energy bands for (a) metallic conductor at T = 0o K; (b)
insulator or intrinsic semiconductor at 0o K.
How much conductivity can a pure (intrinsic)
semiconductor exhibit?
This depends on how much thermal energy
there is and the size of the energy gap, EG:
• Mean thermal energy is kT, where k =
Boltzmann’s constant = 1.38 x 10-23 J/K and T
is the absolute temperature.
• In electron volts this is kT/qe, or 26
millivolts for room temperature (300o K)
• For silicon, EG = 1.12 eV at 300o K
• This leads in pure (intrinsic) Si to a carrier
concentration ni = 1010 carriers/cm3 at 300o K
Adding Impurities (Doping) to Adjust
Carrier Concentrations
Adjust carrier concentrations locally in semiconductor by adding

easily ionized impurities to produce mobile electrons and/or holes
To make silicon N-type:

• Add valence 5 phosphorous (P) atoms to valence 4 silicon. Fifth
electron is easily freed from the atom by a
little thermal energy (0.045 eV for phosphorous) to create (donate) a
mobile electron. Fig. 3a shows the donor energy level just below
bottom of conduction band.
To make silicon P-type:

• Add valence 3 boron (B) to silicon. An electron at the top of the
valence band can pick up enough thermal energy to release it from
the silicon so it attaches to a boron atom, completing its outer ring of
electrons. In the band picture, Fig. 3b, this is represented by an
acceptor. energy level 0.045 eV above the top of the valence band.
Fig. 3a. Electronic energy band for n-type semiconductor
(Ge) with donors only.
Fig. 3b. Electronic energy band for p-type semiconductor
(Ge) with acceptors only.
Energy Bands
 Bloch Functions
 Nearly Free Electron Model
 Kronig-Penney Model
 Wave Equation of Electron in a Periodic
Potential
 Number of Orbitals in a Band
Some successes of the free electron model:
C, κ, σ, χ, …
Some failures of the free electron model:

 Distinction between metals, semimetals, semiconductors & insulators.
 Positive values of Hall coefficent.
 Relation between conduction & valence electrons.
 Magnetotransport.
Band model
finite T impuritie
New concepts: s
 Effective mass
 Holes
Nearly Free Electron Model
Bragg reflection → no wave-like → energy gap
solutions
1
Bragg k  G   k 2
2
→  2k  G   G  0 k G
condition: 2
 2 cos  x / a 
 
   e i x / a
e i  x / a
 k
2i sin  x / a 
a

Origin of the Energy Gap
U  x   U cos  2 x / a 
 2 cos  x / a 

   
2i sin  x / a 

4 cos  x / a 
 2
   

 4sin 2
 x / a 
E  E
a
 
dx U  x            4U  dx cos2  2 x / a   U  0
a

2 2
Eg
0  PHY075- Condensed MatterPhysics, Spring02013 : Energy Bands in Solids
P.Ravindran,
Bloch Functions
Periodic potential → Translational symmetry → Abelian group T = {T(Rl)}
T Rr  r  R  
T  R  f r   f T  R  r  f r  R 
1
k-representation of T(Rl) is e  i k  Rl Basis = e i k  r
Corresponding basis function for the Schrodinger equation must satisfy

T  Rl  k  r    k T  Rl  r
1
   k  r  Rl   e i k Rl k  r 
or  k  r  Rl   e i k R  k  r 
l
This can be satisfied by the Bloch function
 k  r   e i k r uk  r  where uk  r  Rl   uk  r 
ei G  Rl  1 → representative values of k are contained inside the Brillouin zone.
Kronig-Penney Model
2
d 2
  U  x   
2m dx 2
0 n a  b  x  n a  b  a
U 
U 0 n  a  b   a  x   n  1 a  b 
2
K2
0 xa U 0   Ae iK x
 Be i K x

2m
2
Q2
b  x  0 U  U0   Ce Qx
 De Q x
U0   
2m
Bloch theorem:   a  x  a  b     b  x  0  ei k  a b
ψ(0) continuous: A B  C  D
AeiKa  BeiKa    b  ei k  a b   CeQb  DeQb  e  

i k a b
ψ(a) continuous:
ψ (0) continuous: iKA  iKB  QC  QD
ψ (a) continuous: iKAeiKa  iKBeiKa     b  ei k  a b   QCeQb  QDeQb  e

i k  a b 
 1 1 1 1  A  0
 iKa  Qb  ik  a  b  Qb  i k  a  b     
 e e  iKa e e  B  0
 iK iK Q Q  C   0
    
 iKeiKa iKe iKa Qe
 Qb  i k  a  b  Qb  ik  a  b   D 0
 Qe 
Q2  K 2
→ sinh Qb sin Ka  cosh Qb cos Ka  cos k  a  b 
2QK
1 2
Delta function potential: b  0, U 0   such that Q ba  P  finite
2
Thus Q K , Qb 1 so that
P
sin Ka  cos Ka  cos ka
Ka
    Ka 
2 2
K2 2 2
   in units of 2 
 
 2m 2ma    
P
sin Ka  cos Ka  cos ka
Ka
3
P
2
Matrix Mechanics
H   E Ansatz   r    ci i  r 
i
c
j
j H  j r   E  c j  j r 
j
 A    d 3r  *  r  A   r 

j
i H  j c j  E  c j i  j
j
 Hj
ij  E Si j  c j  0 H i j  i H  j Si j  i  j
Matrix equation  H  E S  c  0
Eigen-problem
Secular HE S 0
equation:
Orthonormal basis: Si j  i  j  i j SI

Fourier Series of the Periodic Potential
3
U  r  Rl   U  r   R l   li ai
i 1
3
→ U  r   U  G m  e i Gm  r
G m   mi bi ai  b j  2   i j
m i 1
1 1
d 3x e 
i G m G n r
d 3x e 
i G m G n r
V 
V

  
  mn V = Volume of crystal
  volume of unit cell
1
→ U Gm    d 3 x e  i G m r U  r 
 
For a lattice with atomic basis at positions ρα in the unit cell
U  r   U  r  ρ  U  r  ρ  Rl   U  r  ρ 

U  r   U  G m  ei Gm  ρ ei Gm  r  U  G m  ei Gm  r
m  m
U  G m   U  G m  ei Gm  ρ is the structural factor


Plane Wave Expansion
ei Gm  r   Cm r k  G m
1 i k r
Bloch function  k r  
V
e C
m
m
m
V = Volume of
1 i k r
r k  e k  Gm k  G n   mn crystal
V
 k  Gm   
2 2
k  Gn H k  Gm  nm k  Gn U k  Gm
2m
k  G n U k  G m  U  G p   n , p  m  U  G n  G m 
p
Matrix form of the Schrodinger

equation:
 2
 k  Gn  
2

  E 
m 
 2m
  nm  U  G n  G m   Cm  0
 
 
 k 
2 2
n = 0:  E   C0  U  G m  Cm  0 (central equation)
 PHY075- Condensed
P.Ravindran, 2 m  Matter
m Physics, Spring 2013 : Energy Bands in Solids
Crystal Momentum of an Electron
Properties of
k:
 k  r   ei k r uk  r  →  k  r  Rl   ei k R ei k ruk  r  Rl   ei k R  k  r 
l l
U=0 →  k  r   ei k  r
Selection rules in collision processes → crystal momentum of electron

is  k.
Eq., phonon k  q  k  G
absorption:
Solution of the Central Equation
  E   
2
k2
 U  G n  G m   Cm  0 k 
m
k G m nm  2m
1-D lattice, U  g   U  g   U  0
only
    
    
 k  2 g  E U 0 0  C 2   0
 U k  g  E U 0  C1   0 
    
 0 U k  E U 0  0    0 
C
 0 U k  g  E U  C1   0 
    
 0 0 U k  2 g  E  C2   0 
    
    
Kronig-Penney Model in Reciprocal Space
a /2
U  x   A    x  sa   2 U n cosGn x
1 1
a  a/2
dx cos Gm x cos Gn x   mn
s n 2
a /2 a /2
1 1
U n   dx U  x  cos Gn x   dx A  A (only s = 0 term contributes)
a  a /2 a  a /2
Eigen-equation:   E  
m
k G m  nm  A Cm  0

A A
A 1  
Cn 
E  k G n
C m → E  k G n 2
 2 n 
2
E 
n n
 k 
m
cot x  
1 2m  a 
n n  x
2  
1  

2mE
1 1 1 K
2mA 2mE  2 n 
2    
2 K n  k  2 n  K k  2 n  K 
2
 
n
 k 
 a 
2
 a a 
a  1 1 
  cot  k  K  a  cot  k  K  a
4 K P.Ravindran, 2 Physics, Spring 2013 : Energy Bands in Solids
2 PHY075- Condensed Matter
 1 
2
a 1
  cot  k  K  a  cot  k  K  a  K
2mE
2
2mA 4K 2 2

a sin Ka sin  a  b 
1  1  cot a  cot b  
sin   k  K  a  sin   k  K  a 
4K sin a sin b
2  2 
a sin Ka
 2sin a sin b  cos  a  b   cos  a  b 
2 K cos ka  cos Ka
P mAa 2
→ sin Ka  cos Ka  cos ka with P (Kronig-Penney
Ka 2 2 model)
Empty Lattice Approximation
2
k2
Free electron in k 
2m
vacuum:
k  G 
2 2
Free electron in empty  k   nk  

lattice: 2m
k  k  G
k 1st BZ
Simple
cubic
Approximate Solution Near a Zone Boundary
 
2
k2

m
 E  k G  nm  U  G n  G m   Cm  0
 m  k 
2m
Weak U, λk2g >> U
 k  g  E U 0   C1   0 
    
 U k  E U  C0    0
    C   0
 0 U k  g E  1   
1  U
k near zone right k
2
g
a → k  g U C1 
E  k  g
C0 C0
boundary:
 k  g  E U  C1   0  for E near λk
   
 U k  E  C0   0 
E 2   k  g  k  E  k  g k  U 2  0
1 
E   k  g  k   k  g  k   4U 2 
2
2 PHY075- Condensed Matter Physics, Spring 2013 : Energy

P.Ravindran,  Bands in Solids
1 
  k 
2 1
E   k  g  k  k g  4U 
2
k gK K << g/2
2  2
2 2
2 g 2 g
 K    K  
 2  2  2 g2 
2
k  g  k    K  
2m 2m m 4 
2 2
2 g 2 g
 K    K   2
Kg
k  g  k   2
  2  2U
2m 2m m
 1  2 Kg 
2

  k   4U
2
k g
2
2U 1  2    
 8U  m  
 g /2 2 K 2 
 2U 1  2 2  
 U 2 m 
2
g
22
K2  g /2 
E    U   1  2 
2m  2  2m  U 
C 1
C0
C  k  g   C1
C  k   C0
  x   C0 ei k x  C1 ei  k  g  x
Number of Orbitals in a Band
Linear crystal of length L composed of of N cells of lattice L  Na
constant a.
Periodic boundary   x  L     x   ei k x u  x 
condition: 2 n 2 n
→ k L  2 n k   → N inequivalent values
L Na
of k
Generalization to 3-D crystals:
Number of k points in 1st BZ = Number of primitive
cells → Each primitive cell contributes one k point to each
band.
Crystals with odd numbers of electrons in primitive cell must be metals,
e.g., alkali & noble metals
metal semi-metal insulator
Braggs Law of Diffraction
 2dsin(0)=nl  Bragg reflection of

 d=spacing of parallel electron waves in
atomic planes crystals cause energy
 0=angle of incidence gap.
 l=wavelength of  At locations of Bragg
incident beam reflection solution to
the Schrödinger
 Requires l<=2d equation do not exist.
Brillouin Zone
 Defined as the primitive

zone in reciprocal
lattice
 Made up of all the wave
vectors that satisfy the
diffraction condition
 Wave vectors outside of
Brillouin zone do not
lead to solution of the
Schrödinger Equation
Origin of Energy Gap
 In one dimension, solution to the

Schrödinger equation at the boundaries of
Brillouin are standing waves.
 Standing waves made up of equal parts of
left and right moving waves
 Waves traveling in different directions
have different energy, corresponding to
energy gap
Band Gap
zone boundary
Origin of the Band Gap
To get a standing wave at the boundaries, you can
take a linear combination of two plane waves:
ix / a ix / a  x 
  1  2  e e  2 cos 
 a 
1  eikx  eix /+
a
2  e ikx  e i
x / a
i x / a  i x / a x 
   1   2  e e  2i sin  
 a 
Electron Density
Summary Band Structures
 What is being plotted? Energy vs. k, where k is the wavevector that
gives the phase of the AO’s as well as the wavelength of the electron
wavefunction (crystal momentum).
 How many lines are there in a band structure diagram? As many
as there are orbitals in the unit cell.
 How is the center of gravity energy level of each band
determined? Usually follows from the MO diagram.
 How do we determine whether a band runs uphill or downhill?
By comparing the orbital overlap at k=0 and k=/a.
 How do we distinguish metals from semiconductors and
insulators? The Fermi level cuts a band in a metal, whereas there is a
gap between the filled and empty states in a semiconductor.
 Why are some bands flat and others steep? This depends on the
degree of orbital overlap between building units.
Wide bands  Large intermolecular overlap  delocalized e-
Narrow bands  Weak intermolecular overlap  localized e-
Energy Bands in Solids:
Bandstructure - Nomenclature
For most purposes, it is sufficient to know the En(k) curves - the dispersion
relations - along the major directions of the reciprocal lattice.
This is exactly what is done when real band diagrams of crystals are
shown. Directions are chosen that lead from the center of the Wigner-
Seitz unit cell - or the Brillouin zones - to special symmetry points. These
points are labeled according to the following rules:
• Points (and lines) inside the Brillouin zone are denoted

with Greek letters.
• Points on the surface of the Brillouin zone with Roman
letters.
• The center of the Wigner-Seitz cell is always denoted by a
G
Bandstructure - Nomenclature
For cubic reciprocal lattices, the points with a high symmetry on the Wigner-
Seitz cell are the intersections of the Wigner Seitz cell with the low-indexed
directions in the cubic elementary cell.
simple
cubic
BZ for fcc - Nomenclature
We use the following nomenclature: (fcc):
The intersection point with the [100] direction is called X
(H)
The line G—X is called .
The intersection point with the [110] direction is called K

(N)
The line G—K is called S.
The intersection point with the [111] direction is called L

(P)
The line G—L is called L.
Brillouin Zone for fcc is wigner seitz cell for bcc

and vice versa.
Nomenclature BZ for bcc lattice
We use the following nomenclature: (red for fcc, blue for bcc):
The intersection point with the [100] direction is called X
(H)
The line G—X is called .
The intersection point with the [110] direction is called K

(N)
The line G—K is called S.
The intersection point with the [111] direction is called L

(P)
The line G—L is called L.
Electron Energy Bands in 3D
Real crystals are three-dimensional and we must consider their
band structure in three dimensions, too.
Of course, we must consider the reciprocal lattice, and, as always

if we look at electronic properties, use the Wigner-Seitz cell
(identical to the 1st Brillouin zone) as the unit cell. There is no
way to express quantities that change as a function of three
coordinates graphically, so we look at a two dimensional crystal
first (which do exist in semiconductor and nanoscale physics).
The qualitative recipe for obtaining the band structure of a

two-dimensional lattice using the slightly adjusted parabolas
of the free electron gas model is simple:
Electron Energy Bands in 3D
The lower part (the "cup") is
contained in the 1st Brillouin zone,
the upper part (the "top") comes
from the second BZ, but it is folded
back into the first one. It thus
would carry a different band index.
This could be continued ad
infinitum; but Brillouin zones with
energies well above the Fermi
energy are of no real interest.
These are tracings along major

directions. Evidently, they contain
most of the relevant information in
condensed form. It is clear that
this structure has no band gap.
Band Structure: KCl
We first depict the band structure of an ionic crystal, KCl. The bands are very
narrow, almost like atomic ones. The band gap is large around 9 eV. For alkali
halides they are generally in the range 7-14 eV.
Band Structure: silver (fcc)
Electron Density of States
The electron density of states is a key parameter in the determination
of the physical phenomena of solids.
Knowing the energy levels, we can count how many energy levels are
contained in an interval E at the energy E. This is best done in k -
space.
In phase space, a surface of constant energy
is a sphere as schematically shown in the
picture.
Any "state", i.e. solution of the Schrodinger
equation with a specific k, occupies the
volume given by one of the little cubes in
phase space.
The number of cubes fitting inside the
sphere at energy E thus is the number of all
energy levels up to E.
Electron Density of States: Free Electrons
Counting the number of cells (each containing one possible state

of ) in an energy interval E, E + E thus correspond to taking
the difference of the numbers of cubes contained in a sphere with
"radius" E + E and of “radius” E. We thus obtain the density of
states D(E) as
1 N ( E , E  E )  N ( E )
D( E ) 
V E
1 dN

V dE
where N(E) is the number of states between E = 0 and E per
volume unit; and V is the volume of the crystal.
The volume of the sphere in k-space is
4
V   k3
3
The volume Vk of one unit cell,
containing two electron states is
 2 
3
Vk   
 L 
The total number of states is then
 3 3 
V  4 k L  k L3 3
N 2 2 
Vk  3 8 3  3  2
   
 2k 2 2mE
E k 
2m 2
3 3 3 3/ 2
 2mE 
k L L
N  2 2 
3 2
3  
3/2
1 dN 1  2m 
D( E )  3  2 2  E
L dE 2  
D(E)
DOS in real Materials
Electron Density of States:
If we know the band structure at every point in the Brillouin zone, then the
DOS is given by the formula
D( )    3  n  k 
dS 1
n 4
where the integral is over the surface Sn() is the surface in k space at
which the nth eigenvalue has the value n.
Obviously we can not evaluate this integral directly, since we don't

know n(k) at all points; and we can only guess at the properties of its
gradient. One common approximation is to use the tetrahedron method,
which divides the Brillouin zone into (surprise) tetrahedra, and then
linearly interpolate within the tetrahedra to determine the gradient. This
method is an approximation, but its accuracy obviously improves as we
increase the number of k-points.
D( )    3  n  k 
dS 1
n 4
When the denominator in the integral is zero, peaks due to van Hove
singularities occur. Flat bands give rise to a high density of states. It is
also higher close to the zone boundaries as illustrated for a two
dimensional lattice below.
Leon van Hove

Electron Density of States: Real Solids
• For the case of metals, the
bands are very free electron-
like (remember we compared
with the empty lattice) and the
conduction bands are partly
filled.
• The figure shows the DOS for

the cases of a metal , Cu, and a
semiconductor Ge. Copper has
a free electron-like s-band,
upon which d-bands are
superimposed. The peaks are
due to the d-bands. For Ge the
valence and conduction bands
are clearly seen.
Electron Density of States
fcc
The basic shape of

the density of states
versus energy is
determined by an
overlap of orbitals. In
this case s and d
orbitals…
bcc
tungsten

Band Theory1

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

Band Theory1

Uploaded by

Copyright:

Available Formats

1

Energy Bands in Solids

 Considering electrons in metals as free particles, electron gas

 If the atoms are far apart- atoms can be considered as

 Explains and distinguish between conductors, insulators and

 Multi-electron atomic potentials are complex

 Bloch’s Theorem states that for a particle

 uk(x) is a periodic function with the periodicity

 Bloch’s Theorem imposes very special

 What is probability density of finding

• But |uk(x)|2 is periodic, so P(x) is as well

 Each electron in a crystalline solid

The wave equation can be solved when the potential is

To satisfy Mr. Bloch, the solution in region III

 III   II eik (a b)

• b becomes infinitesimal as U0 becomes infinite.

 When atoms are covalently bonded 1

 Consider an atom with only one electron in s-state

 The sum of them is shown in the

 The last completely filled (at least at T = 0 K)

 Consider a solid with the empty

 Consider a solid with the half filled

 In some cases the opposite occurs

 There is no difference between Insulators and

 Differences arises at high temperatures

 Consider an insulator (or semiconductor)

 If an electric field is applied,

 If we describe such changes via

Electrical current for holes and electrons in the opposite direction

Sommerfeld: k-vector ( k is momentum)

Bloch: k-vector ( k is the crystal momentum)

Bloch: For eachn k , runs through all wave vectors in a single

Sommerfeld: The wave function of an electron with

where the function nk (r ) has no simple explicit form. The

•Fermi energy  F   (T  0) represents the sharp occupancy

F No Fermi surface for

•If  F is defined as the energy separating the

Fermi Golden Rule: Probability per unit

Once the background is subtracted

In isolated atoms the

The highest occupied band is called the VALENCE band.

Conductors : There is an overlap between the valence and

Semiconductors : There is a small energy gap between

This is the replacement of a silicon atom in

Silicon is in group 4 of the periodic table. It

This lowers the number of

Electrons can move through the

Replace a silicon atom with a group

The extra electron introduced

This is n type silicon.

Conductors, semiconductors, insulators: Why is it that

Understand in terms of allowed, empty, and occupied

• If the atoms are pushed together to form a solid

To conduct electricity we need to have a

The metal can conduct at 0o K because the uppermost

If we raise the temperature of the semiconductor, some

Adjust carrier concentrations locally in semiconductor by adding

To make silicon N-type:

To make silicon P-type:

Some failures of the free electron model:

k-representation of T(Rl) is e  i k  Rl Basis = e i k  r

Corresponding basis function for the Schrodinger equation must satisfy

This can be satisfied by the Bloch function