Quantum Mechanics for

Scientists and Engineers

David Miller
Band structures
Band structures

Band structure diagrams

Band structure

If we knew the periodic potential VP  r 

we could solve the resulting one-electron Schrödinger
equation
2 2
    r   VP  r   r   E  r 
2me
using the Bloch function form   r   u  r  exp  ik  r 
calculating the energies E of all the possible states
These calculations give a “band structure”
There are various band structure calculation methods
Many methods “guess” VP  r  , adjusting it to fit data
Band structure diagrams

To construct a band structure

presuming we know VP  r 
we choose one of the
allowed values of k E
For simplicity we restrict to one
dimension for the moment
Solving the equation
gives energy eigensolutions
We continue with the other -p/a 0 p/a
allowed values of k k
Band structure diagrams
A larger crystal gives more allowed
values of k
For a large crystal
the sets of “dots” effectively E
become like lines
We refer to the group of dots on a
line as a “band”
The number of k-states in a band
is the number of unit cells in
the crystal -p/a 0 p/a
k
In practice, we just show the lines
Band structure diagrams

There are multiple bands in a

band structure
in fact an infinite number
but usually only a few are E
important for the
properties of a material
In each band, we only have to
plot k-values from –/a to /a
This range is known as -p/a 0 p/a
the (first) Brillouin zone k
 Extended Brillouin zone scheme

If we continue to
larger k
the band structure
just repeats E
in multiple  
Brillouin zones
an “extended
zone scheme” 
3

2

 0  2 3
so we only need a a a a a a
k
to plot one
Brillouin zone
Band structure diagrams

Each band loosely corresponds

to a different atomic state in
the constituent atoms
or at least orthogonal E
combinations of atomic
states
The bands are formed from the
atomic states
as the atoms are “pushed
-p/a 0 p/a
together” to make the crystal k
Kramers degeneracy

The band structure is drawn to

be symmetric about k = 0
This common symmetry is
easily proved E
Suppose that the Bloch function
  r   uk  r  exp  ik  r 
satisfies the Schrödinger
equation for a specific k
Note the unit cell function uk  r  -p/a 0 p/a
may be different for different k k
Kramers degeneracy

Hence we have
H  k , r   Ek  k , r 
where Ek is the eigenenergy
associated with this specific k E
in this specific band
and
H     2 / 2me   2  VP  r 

-p/a 0 p/a
k
Kramers degeneracy

Taking the complex conjugate of

both sides of
H  k , r   Ek  k , r 
noting that H  H  E
and that Ek is necessarily real
H   k , r   Ek   k , r 
But
   k , r   uk  r  exp  ik  r 
which is also a wavevector in
Bloch form -p/a 0 p/a
k
but for -k
Kramers degeneracy

Hence we are saying that

for every solution with
wavevector k and energy Ek
there is one with wavevector E
-k with the same energy
Hence the band structure is
symmetric about k = 0
We can choose to write
   k , r   uk  r  exp  ik  r  -p/a 0 p/a
k
 u k  r  exp  ik  r     k , r 
Kramers degeneracy

This equivalence of the energies

for k and -k is known as
Kramers degeneracy
Note that, once we include spin E
these two states will have
opposite spin
but often the spin makes no
difference to the energy
Hence bands often have minima -p/a 0 p/a
or maxima at k = 0 k
 Band structures

Semiconductors, insulators and metals

Semiconductors and insulators
conduction empty
Semiconductors and insulators band states
have an (almost) completely
full band
the valence band E
EG
separated by a “bandgap” valence
energy EG band

from an (almost) completely

full states
empty band
the conduction band -p/a 0 p/a
k
Semiconductors and insulators
conduction empty
Note that band states
an empty band does not
conduct electricity
there are no mobile electrons E
EG
Also valence
a full band does not conduct band

electricity
full states
the electrons cannot change
states within the band -p/a 0 p/a
because all the states are full k
Semiconductors and insulators
conduction empty
The difference between band states
semiconductors and
insulators is primarily that
insulators have such a large E
bandgap energy EG
valence
that there is negligible band
thermal excitation of
full states
electrons
from the valence band
-p/a 0 p/a
to the conduction band k
Semiconductors
conduction empty
At finite temperatures in a band states
semiconductor
a small number of electrons
are excited E
from the valence band EG
valence
to the conduction band band

full states

-p/a 0 p/a
k
Semiconductors
conduction empty
These electrons in the band states
conduction band
and
absences of electrons or “holes” E
EG
in the valence band valence
can conduct electricity within band

their bands
full states
So semiconductor materials
conduct electricity weakly -p/a 0 p/a
hence the name k
Doping semiconductors
conduction empty
Substituting a few atoms with band states
more electrons
e.g., a Group V element like
phosphorus in a Group IV E
semiconductor like silicon EG
valence
known as n-type doping band
makes the material conduct
full states
more
using these additional -p/a 0 p/a
electrons k
Doping semiconductors
conduction empty
Substituting a few atoms with band states
fewer electrons
e.g., a Group III element like
boron in a Group IV E
semiconductor like silicon EG
valence
known as p-type doping band
makes the material conduct
full states
more
using these additional -p/a 0 p/a
“holes” k
Direct gap semiconductor
conduction
If the lowest minimum in the band
conduction band
lies directly above
the highest maximum in the E
valence band EG
valence
the semiconductor is said to band
have a
“direct gap”
-p/a 0 p/a
k
Direct gap semiconductor
conduction
Direct gaps are important for band
light emitters
Electrons “pumped” into the
conduction band gather in E
the lowest minimum valence
“Holes” pumped into the band
valence band gather in the
highest maximum
An electron can fall “vertically”
-p/a 0 p/a
to fill in a hole beneath it k
Direct gap semiconductor
conduction
Direct gaps are important for band
light emitters
Electrons “pumped” into the
conduction band gather in E photon
the lowest minimum valence
“Holes” pumped into the band
valence band gather in the
highest maximum
An electron can fall “vertically”
-p/a 0 p/a
to fill in a hole beneath it k
emitting light
Indirect gap semiconductor
conduction
In an indirect gap semiconductor band
E.g., silicon, germanium
the lowest conduction band
minimum (or minima) E
is not directly above the valence
highest valence band band
maximum
Light emission is weak
“non-vertical” transitions by -p/a 0 p/a
emission of photons are weak k
Metals
conduction empty
Because of the number of band states
electrons in the metal atoms
the highest band is partially
full of electrons E
e.g., half-full EG
valence
even at zero temperature band

Hence metals conduct

full states
electricity well
-p/a 0 p/a
k
Band structures

Band structures in 3D
Three-dimensional band structures

We need a four dimensional diagram to draw

a complete band structure for a three
dimensional crystal
three directions plus another axis for
energy
but we can nonetheless still draw
simplified and useful diagrams
To see this
we look at the Brillouin zone for the
diamond or zinc-blende lattice
Brillouin zone in 3D
ky
This Brillouin zone for the
diamond or zinc-blende lattice
is itself a 3D object L

in k-space (or reciprocal space)  X kx

Two important directions are kz
X – along one of the x, y, or z
coordinate directions
L – along one of the cube space
diagonals
The center is called the  point
Band structures for 3D crystals
ky
At least as a first useful
representation of band structure
typically it is calculated only L
along a few directions  X kx
such the along the lines from kz
the  point (at the center of
the Brillouin zone)
to the X point and the L
point
Si band structure 6
conduction
bands
3
Sketch of major valence and
conduction bands 0 “ point”
with the conduction band

Energy (eV)
minimum at the  point -3
By Kramers degeneracy valence
-6
we need only show one half bands
of the band structure -9
so we can use the other
half of the figure -12

for the band structure in L  X

after K. S. Sieh and P. V. Smith, Phys.
another direction Status Solidi (b) 129, 259 (1985)
GaAs band structure 10
conduction
bands
Sketch of major valence and 5
conduction bands
with the conduction band

Energy (eV)
0
minimum at the  point
Note that GaAs is a direct gap -5 valence
bands
semiconductor
-10

L  X
after M. Rohlfing, P. Krüger and J.
Pollmann, Phys. Rev. B 48, 17791 (1993)
Extended zones in 3D

In 3D
additional Brillouin zones
repeat the same band
structure
These zones form “unit
cells” in k-space
filling all k-space
(reciprocal space)
Extended zones in 3D

Marking the same “reciprocal

lattice point” in each cell
Extended zones in 3D

Marking the same “reciprocal

lattice point” in each cell
and erasing the “unit cells”
(Brillouin zone shapes)
themselves for clarity
Extended zones in 3D

Marking the same “reciprocal

lattice point” in each cell
and erasing the “unit cells”
(Brillouin zone shapes)
themselves for clarity
and adding guide lines
shows these extended
Brillouin zones correspond
to a body centered cubic
“reciprocal lattice”