INTERBAND TRANSITIONS

In a semiconductor at low frequencies, the principal electronic

conduction mechanism is associated with free carriers
Band Theory
The theory that explains the band formation in a solid is called energy band theory.
For an isolated atom, the electrons in it occupy distinct energy levels. In solids atoms are closely packed so that
each energy level of the isolated atom splits up into sub-levels, called states. It is because of the forces exerted by
nearby atoms on an electron. These states are discrete but so closely spaced that they appear to form a continuous
energy band. This band is occupied by electrons and it is called an allowed energy band. We discuss only two
following allowed energy bands:
1. Valence Band
2. Conduction Band
VALANCE BAND
The band occupied by the valance electrons is called valance band.
i. It is highest occupied band.
ii. It may either be completely filled or partially filled by electrons but can never be empty.

CONDUCTION BAND
The band above the valance band is called conduction band and it is occupied by free electrons.
i. It plays main role in conduction.
ii. It may be either empty or partially filled.

FORBIDDEN ENERGY GAP

The gap between valance band and conduction band is called forbidden energy gap. There are no electrons
are occupied in this gap.
INSULATOR SEMICONDUCTOR CONDUCTOR
OPTICAL TRANSITION BETWEEN THE BANDS

Interband optical transitions will be possible between these bands if E ≥ 𝐸𝑔 .

During the transition an electron jumps from the band at lower energy to the one
above it by absorbing a photon.
This can only happen if
• There is an electron in the initial state in the lower band.
• The Pauli exclusion principle demands that the final state in the upper band
must be empty.
• The transitions across the fundamental band gap of a semiconductor or
insulator.

• The law of conservation of energy demands

𝐸𝑓 = 𝐸𝑖 + ℏ 𝜔 ⇒ 𝐸𝑓 − 𝐸𝑖 = ℏ 𝜔
• Since there is a continuous range of energy states within the upper and lower bands, the interband
transitions will be possible over a continuous range of frequencies. The range of frequencies is determined
by the upper and lower energy limits of the bands.

• The minimum value of (𝐸𝑓 − 𝐸𝑖 ) is 𝐸𝑔 . This implies that the absorption shows a threshold behavior:
interband transitions will not be possible unless ℏ𝜔 > 𝐸𝑔 .

• The excitation of the electron leaves the initial state at energy 𝐸𝑖 in the lower band unoccupied. This is
equivalent to the creation of a hole in the initial state. The interband absorption process therefore creates a
hole in the initial state and an electron in the final state, and may be considered as the creation of an
electron-hole pair.
 TYPES OF BAND GAP MATERIALS

There are two types of band gap materials:

1. Direct band gap materials: If the relative position of the conduction band minimum and the valence band
maximum both lie at the same k, it is called direct band gap material. 𝑘𝐶𝐵 = 𝑘𝑉𝐵

2. Indirect band gap materials: If the relative position of the conduction band minimum and the valence band
maximum both do not lie at the same k, it is called indirect band gap material. 𝑘𝐶𝐵 = 𝑘𝑉𝐵 ∓ 𝑞
TRANSITION RATE OF DIRECT BANDGAP MATERIALS

To find the transition rate that is the transition probability per unit time for
direct band gap materials, we use Fermi-Golden rule, which is defined as:

2𝜋
𝑊𝑖−𝑓 = 𝑀 2 𝑔(ℏ𝜔)…………(1)
ℏ

Where, 𝑀 =< 𝑓 𝐻′ 𝑖 > is the matrix element that couples initial and final
states of an electron. It describes the size of the given perturbation, and
𝑔(ℏ𝜔) is the density of states (DOS).
It is used for the transitions between discrete levels such as those found in
individual atoms to continuous bands of levels. This situation is the usual
one in solids, and is shown in Fig. 4. Where 𝑔(ℏ𝜔)𝑑𝐸 is the number of
states that fall in the energy range E to E+dE. Transitions are possible to any
of the states that fall within this energy range. The density of states factor in
Eq. (1) is very important in solid state problems. In the discussion below, we
consider the matrix element first, and we consider DOS afterwards.
Matrix Element

The matrix element describes the effect of the external perturbation caused by the light wave on the electrons which are
initially present in the valence band. It is given by

𝑀 =< 𝑓 𝐻′ 𝑖 >

= ‫𝐻 𝒓 ∗𝑓𝜓 ׬‬′ (𝒓)𝜓𝑖 (𝒓)𝑑3 𝒓…………………..(2)

Where, 𝐻′ is the perturbation associated with the light wave i.e. it represents interaction between the atom and the
radiation field. r is the position vector of the electron, 𝜓𝑖 (𝒓) and 𝜓𝑓 (𝒓) are the wave functions of initial | 𝑖 > and final
|𝑓 > states respectively. In order to find this matrix element, we shall use the semi-classical approach in which the
electrons are treated quantum mechanically while the photons are described by electromagnetic waves (classically ).
To evaluate Matrix element in Eq. (2), we need to know the wave functions of the states, and also the form of the
perturbation due to the light wave. The perturbation due to the light can be evaluated by calculating the effect of the
electromagnetic field on the electron in the atom. From classical electromagnetism we know that the field changes the
momentum of a charged particle from p to (p - qA), where q is the charge and A is the vector potential defined by

𝑩= 𝛁×𝑨

The Hamiltonian for an electron with q = -e in an electromagnetic field is therefore:

1
𝐻= (𝐩 + e𝐀)2 + 𝑉 (𝒓)
2𝑚0

𝐩𝟐 𝑒 𝑒 2 𝐴2
= + 𝑉 ( 𝒓) + (𝑷 ∙ 𝑨 + 𝑨 ∙ 𝑷) + = 𝑯𝟎 + 𝑯′
2𝑚0 2𝑚0 2𝑚0

𝐩𝟐 𝑒 𝑒 2 𝐴2
Where, 𝑯𝟎 = + 𝑉 (𝒓) and 𝑯′ = ( 𝑷 ∙ 𝑨 + 𝑨 ∙ 𝑷) + are the Hamiltonians of an
2𝑚 0 2𝑚 0 2𝑚 0
electron when there is no field and when the field is applied.
Let we try to simplify the perturbed Hamiltonian as,

′ 𝑒 𝑒 2 𝐴2
𝑯 = ( 𝑷 ∙ 𝑨 + 𝑨 ∙ 𝑷) + …………….(3)
2𝑚 0 2𝑚 0
We make the following approximations to simplify it,
1. For low intensity light, A will be small; therefore, the term involving 𝐴2 can be neglected.
2. A and P commute i.e., 𝑷 ∙ 𝑨 − 𝑨 ∙ 𝑷 = 𝟎.

Using the above approximations, Eq. (3) can be written as:

𝑒
𝑯′ = 𝑚 𝑷 ∙ 𝑨………………………..(4)
0

Since E and B both vary in time and space of an electromagnetic wave, so A must also do so,
Therefore,
𝑨 = 𝑨𝟎 𝑒 𝑖 𝒌.𝒓−𝜔𝑡

= 𝑨𝟎 𝑒 𝑖 𝒌.𝒓 𝑒 −𝑖 𝜔𝑡
By applying the Taylor’s series, we get

𝒊𝒌.𝒓 2
= 𝑨𝟎 [𝟏 + 𝒊𝒌. 𝒓 + + ⋯ ]𝑒 −𝑖 𝜔𝑡
2

At optical frequencies,
𝜆~1 𝜇𝑚

𝑟~10−10 𝑚 (size of an atom)

2𝜋|𝒓|
⇒ 𝒌. 𝒓 = ~10−3
𝜆
Therefore, 𝒌. 𝒓 and its higher powers can be neglected in the expression of A and it can be written as
𝑨 ≅ 𝑨𝟎 𝑒 −𝑖 𝜔𝑡
For 𝑡 → 0, we get
𝑨~𝑨𝟎

Therefore, Eq. (4) can be written as

𝑒
𝑯′ = 𝑷 ∙ 𝑨𝟎
𝑚0

So the matrix element in Eq. (2) will becomes,

𝑒
𝑀 = 𝑚 < 𝑓 𝑷 ∙ 𝑨𝟎 𝑖 >…………………………(5)
0
The equation of motion for time dependent operator O is given by,

𝑑𝐎 𝑖
= 𝑯 ,𝑶
𝑑𝑡 ℏ 𝟎

𝑖
= ℏ (𝑯𝟎 𝑶 − 𝑶𝑯𝟎 )

Similarly, we can write

𝑑𝐫 𝑖
= 𝑯 ,𝒓
𝑑𝑡 ℏ 𝟎

𝑖
= ℏ (𝑯𝟎 𝒓 − 𝒓𝑯𝟎 )

𝑑𝐫
∴ < 𝑓 𝑷 𝑖 >= 𝑚0 < 𝑓 𝑖>
𝑑𝑡
𝑖
= ℏ 𝑚0 < 𝑓|(𝑯𝟎 𝒓 − 𝒓𝑯𝟎 )|𝑖 >
 𝑖
= ℏ 𝑚0 ( 𝐸𝑓 − 𝐸𝑖 ) < 𝑓|𝒓|𝑖 >

(We use 𝐻𝜓 = 𝐸𝜓)

𝐴𝑠 𝐸𝑓 − 𝐸𝑖 = ℏ𝜔

∴ < 𝑓 𝑷 𝑖 >= 𝑖𝑚0 𝜔 < 𝑓|𝒓|𝑖 >………………………..(6)

Therefore, Eq. (5) becomes

𝑀 = 𝑖𝑒𝜔 < 𝑓 𝒓 ∙ 𝑨𝟎 𝑖 >

=< 𝑓 𝑒𝒓 ∙ 𝑖𝜔𝑨𝟎 𝑖 >

𝛿𝑨
As 𝑨~𝑨𝟎 and 𝑬 = − 𝛿𝑡 = 𝑖𝜔𝑨𝟎 , also 𝒑 = −𝑒𝒓, is the dipole moment.

∴ 𝑀 =< 𝑓 −𝒑 ∙ 𝑬 𝑖 >……………………(7)
By coparing (2) and (7), we get

𝐻′ = −𝒑 ∙ 𝑬………………………………..(8)

Where, 𝒑 = −𝑒𝒓, is the electric dipole moment. This is in fact exactly equal to the interaction energy
experienced by a dipole in an electric field. That’s why the interband transitions are also known as
electric dipole transitions.
The wave functions of electron in a crystalline solid are described by Bloch functions.
“The wave functions as a product of a plane wave and an envelope function that has the periodicity of the crystal
lattice.”

Where 𝑢𝑖 and 𝑢𝑓 are the appropriate envelope functions for the initial and final bands respectively, and V is the
normalization volume. 𝑘𝑖 and 𝑘𝑓 are the wave vectors of the initial and final electron states.

M= ‫𝐻 𝒓 ∗𝑓𝜓 ׬‬′ (𝒓)𝜓𝑖 (𝒓)𝑑 3 𝒓

The integral can be simplified by following two considerations:

1. The change in crystal momentum of the electron must equal the momentum of the photon,
that is 𝑘𝑓 − 𝑘𝑖 = 𝑘

𝑒
M= 𝑉 ‫𝐸( 𝒓 ∗𝑓𝑢 ׬‬0 ∙ 𝒓) 𝑢𝑖 (𝒓)𝑑 3 𝒓

When the light is polarized along x-axis, then

𝑒𝐸0 ∗
M= ‫׬‬ 𝑢𝑓 𝒓 𝑥𝑢𝑖 (𝒓)𝑑 3 𝒓
𝑉

2. Bloch’s theorem requires that 𝑢𝑖 and 𝑢𝑓 are periodic functions with the same periodicity as the lattice. This
implies that we can separate the integral over the whole crystal into a sum over identical unit cells, because the unit
cells are equivalent
Joint Density of States
The density of states function describes the distribution of the states within the bands. The joint density of states
accounts for the fact that both the initial and final electron states lie within continuous bands.

where g(k) is the density of states in momentum space. The extra factor of 2 here compared to the fact that there
are two electron spin states for each k state. This gives:
3.1: Apply Born–von Karman periodic boundary conditions (i.e. 𝑒 𝑖𝑘𝑥 = 𝑒 𝑖𝑘(𝑥+𝐿) etc., where L is a
1 3
macroscopic length) to show that the density of states per unit volume in k space is .
2𝜋

𝑒 𝑖𝑘𝑥 𝑥 = 𝑒 𝑖𝑘𝑥 (𝑥+𝐿)

⇒ 𝑘𝑥 𝐿 = 2𝜋𝑛𝑥

Similarly
𝑘𝑦 𝐿 = 2𝜋𝑛𝑦

𝑘𝑧 𝐿 = 2𝜋𝑧

The density of states per unit volume in k space

3
1
= 2𝜋 3
2𝜋 Volume in k-space =𝑘𝑥 × 𝑘𝑦 × 𝑘𝑧 = 𝐿
𝑛𝑥 𝑛𝑦 𝑛𝑧
2𝜋 3
Volume of an allowed state in k-space =
𝐿
2𝜋 3 2𝜋 3
= 𝐿3
= 𝑉
 ℏ2 𝑘 2
3.2: Show that the density of states for an electron with 𝐸 𝑘 = is given by
2𝑚∗
3/2
1 2𝑚∗
𝑔 𝐸 = 2 𝐸1/2
2𝜋 ℏ2

ℏ2 𝑘 2
𝐸 𝑘 =
2𝑚∗

𝑑𝐸 ℏ2 𝑘
⇒ = ∗
𝑑𝑘 𝑚

k 2ൗ
2𝜋 2 𝑚∗ 𝑘
𝑔 𝐸 =2 2 = 2 2
ℏ 𝑘 𝜋 ℏ
𝑚 ∗

3/2
1 2𝑚∗
𝑔 𝐸 = 2 𝐸1/2
2𝜋 ℏ2
The joint density of states factor is finally obtained by evaluating 𝑔(𝐸) at 𝐸𝑖 and 𝐸𝑓 when they are related to
ℏ𝜔 through the band energies. To proceed further, we therefore need detailed knowledge of the band
structure. We use the following four-band model to find joint density of states:

ℏ2 𝑘 2
𝐸𝑒 𝑘 = 𝐸𝑔 +
2𝑚𝑒∗
ℏ2 𝑘 2
𝐸ℎℎ 𝑘 =− ∗
2𝑚ℎℎ

ℏ2 𝑘 2
𝐸𝑙ℎ 𝑘 =− ∗
2𝑚𝑙ℎ

ℏ2 𝑘 2
𝐸𝑠𝑜 𝑘 = −Δ − ∗
2𝑚𝑠𝑜
For transition 1

ℏ𝜔 = 𝐸𝑒 𝑘 − 𝐸ℎℎ (𝑘)

ℏ2 𝑘 2 ℏ2 𝑘 2
= 𝐸𝑔 + +
2𝑚𝑒∗ 2𝑚ℎℎ ∗

ℏ2 𝑘 2 1 1 1 1 1
= 𝐸𝑔 + + ∵ + ∗ =
2 𝑚𝑒∗ 𝑚ℎℎ
∗ 𝑚𝑒∗ 𝑚ℎℎ 𝜇

ℏ2 𝑘 2
𝐸 = 𝐸𝑔 +
2𝜇
k 2ൗ
2𝜋 2 𝑚∗ 𝑘
𝑔 𝐸 =2 2 = 2 2
𝑑𝐸 ℏ2 𝑘 ℏ 𝑘 𝜋 ℏ
⇒ = 𝑚 ∗
𝑑𝑘 𝜇

3
1 2𝜇 2 1/2
𝑔 𝐸 = 𝐸 − 𝐸𝑔
2𝜋 2 ℏ2
 3
2
1 2𝜇 2 1/2
𝑊𝑖−𝑓 ∝ 𝑀 𝐸 − 𝐸𝑔
2𝜋 2 ℏ2

1/2
𝑊𝑖−𝑓 ∝ 𝐸 − 𝐸𝑔

𝛼 ℏ𝜔 ∝ 𝑊𝑖−𝑓

1/2
𝛼 ℏ𝜔 = 𝐶 𝐸 − 𝐸𝑔

𝛼 2 ℏ𝜔 = 𝐶1 𝐸 − 𝐸𝑔

It shows results for the absorption coefficient of the direct gap III–V semiconductor indium arsenide at room
temperature. The graph plots 𝛼 2 against the photon energy in the spectral region close to the band gap. The straight-
line relationship between 𝛼 2 and (ℏ𝜔 − 𝐸𝑔 ) indicates that the model we have developed is a good one. The band
gap can be read from the data as the point at which the absorption goes to zero. This gives a value of 0.35 eV, which
is in good agreement with values deduced from electrical measurements.

Note that the values of the absorption coefficient are very large. This is a consequence of the
very large density of states in the solid phase.
In many III–V semiconductors, including GaAs itself, it is found that the frequency dependence predicted by
equation is only approximately obeyed. There are a number of reasons for this.

• We have neglected the Coulomb attraction between the electron and hole, which can enhance the absorption
rate and cause exciton formation. These effects become stronger as the band gap gets
larger and the temperature is lowered. This is why we have presented room temperature data for a
semiconductor with a band gap.

• The semiconductor may contain impurity or defect states with energies within the band gap, and these
may allow absorption for photon energies less than the band gap.

• The parabolic band approximations embodied in the dispersion relations are only valid near k = 0.