Density of States Derivation

The density of states gives the number of allowed electron (or hole) states per volume at a
given energy. It can be derived from basic quantum mechanics.

Electron Wavefunction

The position of an electron is described by a wavefunction  x, y, z  . The probability of

finding the electron at a specific point (x,y,z) is given by  x, y, z  , where total
2

  x, y, z  dxdydz is normalized to one.

2
probability
all
space

Particle in a Box
The electrons at the bottom of a conduction band (and holes at the top of the valence
band) behave approximately like free particles (with an effective mass) trapped in a box.
We will consider here conduction band electrons, but the result for holes is similar. For
our parabolic conduction band:
2 2
k
 E  Ec   *
2m

For electrons in a rectangular volume Lx by Ly by Lz with an infinite confining potential

((U(x,y,z)=0 inside the box and ∞ outside), the electron wavefunction  must go to zero
on the boundaries, and takes the form of a harmonic function within the region. The
wavefunction solution is:

  x, y, z   sin  k x x  sin  k y y  sin  k z z  (1)

and k x , k y , and k z are the wavevectors for an electron in the x, y, and z directions. The
real wavefunction in a solid is more complex and periodic (with the crystal lattice), but
this is a good approximation for the parabolic regions near the band edges.

First 4 particle in a box wavefunctions across the x direction.

Orthogonal directions are analogous.

Enforcing the boundary conditions: At x, y, or z = 0, the sine functions go to zero. At the

opposite boundaries of the rectangular region, sin k x Lx   0 , sin k y Ly   0 , and
sink z Lz   0 for the x, y, and z directions. The allowed wavevectors satisfy:

kx Lx   nx , k y Ly   ny , k z Lz   nz , for nx , ny , nz integers (2)

K Space

The allowed states can be plotted as a grid of points in k space, a 3-D visualization of the
directions of electron wavevectors. Allowed states are separated by  / Lx , y , z in the 3
directions in k space.

The k space volume taken up by each allowed state is  3 / Lx Ly Lz . The reciprocal is the
state density in k space (# of states per volume in k space), V /  3 where V is the volume
of the semiconductor (in real space).

The number of states available for a given magnitude of wavevector |k| is found by
constructing a spherical shell of radius |k| and thickness dk. The volume of this spherical
shell in k space is 4k 2 dk .
 Allowed wavevector states in k space form a lattice.
A spherical shell gives the number of allowed states at a specific radius |k|.

The number of k states within the spherical shell, g(k)dk, is (approximately) the k space
volume times the k space state density:
V 
g (k )dk  4 k 2  3  dk (3)
 
Each k state can hold 2 electrons (of opposite spins), so the number of electron states is:
V 
g (k )dk  8 k 2  3  dk (4a)
 
Finally, there is a relatively subtle issue. Wavefunctions that differ only in sign are
indistinguishable. Hence we should count only the positive nx, ny, nz states to avoid
multiply counting the same quantum state. Thus, we divide (4a) by 1/8 to get the result:
V  Vk 2 
g (k )dk   k 2  3  dk   2  dk (4b)
    

This is an expression for the number of unique electron states available at a given |k| over
a range dk. We need an expression in terms of energy rather than wavevector k. We
proceed from the relationships between wavevector, momentum p, and energy E:
 2k 2
p  k , E  p / 2m E 
2 *
(5)
2m*
with m* as the effective mass. Rewriting, and noting that the energy of carriers in the
conduction band is given with respect to the conduction band edge energy Ec:

k2 
 E  Ec  2m* (6)
2

Differentiating:
2m*dE
2kdk  (7a)
2
Combining (6) and (7a):
 2m*dE m*dE m*dE
dk   
2k 2 k 2 2
2m*  E  Ec  2

(7b)
m*dE

2m*  E  Ec 

Plugging (6) and (7) into (4b):

Vk 2 m*dE
g (k )dk 
2 2m*  E  Ec 
V  2m*  E  Ec  /   m*dE 
2

 (8)
 2m*  E  Ec  
1/2
2

Vm*  2m*  E  Ec  
1/2

 dE
2 3

Dividing through by V, the number of electron states in the conduction band per unit
volume over an energy range dE is:
m*  2m*  E  Ec  
1/2

g ( E )dE  dE (9)
2 3

This is equivalent to the density of the states given without derivation in the textbook.

3-D density of states, which are filled in order of increasing energy.

Dimensionality

The derivation above is for a 3 dimensional semiconductor volume. What happens if the
semiconductor region is very thin and effectively 2 dimensional?

Confining the electron in the x-y plane, the wavevector z component kz =0. The allowed
states in k space becomes a 2 dimensional lattice of kx and ky values, spaced  / Lx , y apart.
The 2-D k space area taken up by each state is  2 / Lx Ly . The number of states per area
in k space is A /  2 with A as the real-space area of the thin semiconductor.
The number of states available at a given |k| is found using an annular region of radius |k|
and thickness dk rather than the spherical shell from the 3-D case. There is a factor of ¼
due to the equivalent nature of the +/- states (just as there was 1/8 in the 3D case). The
area is  kdk . The number of k space states is:
 A  A
g (k )dk   2  2 kdk  2 spin states 4 equivalent states= k   dk (10)
   
Converting to energy using (7b):
A  2m*  E  Ec  2   
1/2
 A  m*  dE
g (k )dk  k   dk  (11)
     2m*  E  E  1/2 
  c  
Cleaning up and dividing through by area, the density of states per area at an energy E
over a range dE is:
m*dE
g ( E )dE  (12)
 2
Unlike the 3-D case, this expression is independent of energy E!

In a real structure (which is not perfectly 2-D), there are finite energy ranges over which
the energy independence holds (the derivation holds for each single, well separated
possible value of kz). The resulting density of states for a quantum well is a staircase, as
below in red.

Further restriction of the semiconductor dimensionality to 1-D (quantum wire) and 0-D
(quantum dot) results in more and more confined density of states functions.

Density of states for 0-D through 3-D regions.

Low dimensional and confining nanostructures have lots of applications controlling

carriers in semiconductor devices.