SEMICONDUCTOR PHYSICS

DR. PRASANTA GHOSH

SCHOOLOFPHYSICS,
MIT-WPUPUNE
The properties and applications of any solid are governed by it’s two characteristics,

its atom and the inter-atomic interaction

Significant difference in the electrical properties of conductors, semiconductors and insulators

A quantum mechanical theory called as band theory of solids which explains reason for this
variation

This theory explains the behavior of electronic devices such as diodes, BJT, FET, LED,
Photodiodes, solar cells, etc
Eleven elements in the periodic table show semiconducting properties.

Germanium and silicon are called elemental semiconductors.

Compound semiconductors such as gallium arsenide, cadmium sulfide, indium phosphide etc.

The semiconducting materials have acquired a unique place in electronics due to their typical
characteristics.

i. The electrical properties can be significantly improved by adding impurities

ii. Ability to conduct due to two kinds of charge carriers, namely electrons (negative) and holes
(positive).
iii. Temperature dependence of resistivity (this makes thermistors possible)
iv. Due to the typical band gaps, semiconductors can convert light in to electricity and electricity in to
light.
The invention of transistor in 1947 was a turning point in the history of technology

Transforming old vacuum tube electronics to solid state electronics.

PN junction diode, NPN or PNP transistor, FET, photodiode, LED, solar cells etc.

Several advantages
compact size, fast action, requirement of less power etc.
 FREE ELECTRON THEORY
Metals and their alloys such as silver, copper, aluminum, lead behave like conductors.

Electrical and even thermal properties are superior as compared to the other materials.

This behavior is explained by classical free electron theory, which was developed by Paul Drude in 1900,
just three years after the discovery of the electron by J.J. Thomson.

This theory was later refined by Hendrik Lorentz in 1905 - Drude-Lorentz theory.

Some limitations of theory were overcome by quantum free electron theory, developed by Summerfield in
1928.

Further, the limitations of even quantum free electron theory were overcome by band theory of solids
developed by Felix Bloch
 FREE ELECTRON THEORY

Salient features of the classical free electron theory:

1. In metals, the valance electrons are free to move and hence they are the conduction electrons.
2. The electrostatic forces of attraction between the electrons and positive ion cores and the electrostatic
forces of repulsion between the electrons are negligible.
3. According to classical theory the energies of the electrons in metals are distributed according to the
classical Maxwell-Boltzman statistics (which is mainly applicable to the molecules in the gas).
According to this statistics, the kinetic energy of the electrons in motion is given by

1 3
𝑚𝑣 2 = 𝑘𝑇
2 2

4. Drift motion of electrons in metals: When electric field is applied, the electrons start drifting in a
direction opposite to the electric field. The drift motion results in to drift current.
 FREE ELECTRON THEORY
Drawbacks of classical free electron theory:

1. This theory proposes that the electrical conductivity depends upon the concentration of the free electrons.
This requires that trivalent (aluminum, indium etc.) and divalent metals (cadmium and zinc) should have
better conductivity than the monovalent metals (copper, silver etc.) However, experimentally it is found that
copper and silver have better conductivity than divalent and trivalent metals.
2. This theory proposes that the conductivity of the metals is due to negatively charged electrons. Thus all the
1
metals should have negative Hall coefficient 𝑅𝐻 = 𝑛𝑞 . However some metals such as zinc have positive
Hall coefficient. This cannot be explained by using classical free electron theory.
3. The average distance covered by the electron between two successive collisions is called as mean free
path. Classical free electron theory predicts that the mean free path should be of the order of 3Ao.
However, it is experientially found that typical mean free paths of the electrons in the metals are of the
order of 50 Ao.
4. Classical free electron theory predicts that the resistivity of the metals should be proportional to the square
root of the absolute temperature. However experimentally it is found that the resistivity of the metals is
directly proportional to the absolute temperature.
 FREE ELECTRON THEORY

5. Classical free electron theory predicts that the molar specific heat at constant volume is given by

3
𝐶𝑉 = 𝑅
2
However, experimentally it is found that,
𝐶𝑉 = 10−4 𝑅𝑇
Thus experimentally CV is much smaller and also temperature dependent. This is against classical theory

6. The classical free electron theory predicts that the magnetic susceptibility of metals to be inversely
proportional to the temperature, however experiments indicate that it is independent of the temperature.
7. This theory fails to explain the classification of the solids in to conductors, semiconductors and insulators.

To overcome these drawbacks, a quantum free electron theory was developed by Somerfield in 1928.
 QUANTUM FREE ELECTRON THEORY
This theory retains the following vital features of the classical free electron theory

1. The metal consists of the free electron gas in which the positive ion cores are symmetrically distributed
2. Both the electrostatic force of attraction between the electrons and positive ion cores and the electrostatic
force of the repulsion between the electrons are negligible.
3. The electrons in the metals are trapped in the metallic potential well having constant potential. The well is
surrounded by the high potential barrier at the boundaries.

The fundamental difference between the classical and quantum free electron theory is related to the energy
distribution of the electrons.

The classical free electron theory suggests that the energies of the free electrons in the metals are distributed
according to Maxwell Boltzman statistics.

However, quantum theory proposes that the electrons are distributed in the various energy levels according to
Fermi Dirac statistics which is mainly based on Pauli’s exclusion principle.
 QUANTUM FREE ELECTRON THEORY
Limitations of Quantum free electron theory:

1. It cannot explain the classification of the solids in to conductors, semiconductors and insulators.
2. The quantum theory also fails to explain the positive Hall coefficients of some elements like zinc.

The fundamental difference between the classical and quantum free electron theory is related to the energy
distribution of the electrons.

The classical free electron theory suggests that the energies of the free electrons in the metals are distributed
according to Maxwell Boltzman statistics.

However, quantum theory proposes that the electrons are distributed in the various energy levels according to
Fermi Dirac statistics which is mainly based on Pauli’s exclusion principle.

This drawback was overcome by band theory of solids which was developed by
Felix Bloch around 1928.
 BAND THEORY OF SOLIDS

●
The energy levels of atoms become energy bands
when a solid is formed
●
Explained on two approaches
– Felix Bloch : Restrictions on the allowed energy levels of
the electrons due to the periodicity of the crystal
– Walter Heitler and Fritz London: Splitting of the energy
levels of the electrons due to their interaction
● EX: COPPER – 29 ELECTRONS
According to Pauli’s exclusion principle, these
electrons are distributed in to 29 exclusively
different quantum states characterized by a

▪ group of four quantum numbers (n, l, ml, ms)

▪ Electronic configuration of copper

1s2 2s2 2p6 3s2 3p6 3d10 4s1

Discrete energy levels

N copper atoms at infinite distance, then such atoms do not interact with each other

At equilibrium, the inter-atomic distance of the copper atoms in the solid form is 2.6Å.

If the atoms come close to each other at such distance, the valance electrons will start strongly
interacting with each other.
In such case the electrons cannot be accommodated in the same energy levels. For ex. the two
4s electrons cannot be accommodated in the same 4s level.
The 4s level splits in to two closely spaced 4s sublevels.
The electrons undergo a strong electrostatic interaction hence their energies change

According to Quantum Mechanics, the electrons behave like waves. Thus when electrons
come in the proximity, their wavefunctions overlap. As a result, symmetric and anti-
symmetric wavefunctions, which correspond to two distinct energy levels, are formed.

when an Avagadro number of atoms come close and form bonds becomes a single system
and thus Pauli’s exclusion principle applies to the pair.

Accordingly, there have to be N different quantum states with different energies for the N 4s
electrons.

Thus when N atoms come together, there will be N number of energy levels, all belonging
to 4s category.

The wavefunctions of these electrons also overlap and the levels split and form bands

The number of quantum states in the band is the product of occupancy of the corresponding level
and the number of atoms.

s band consists of 2N states, p band contains 6N states, d band contains 10N states
 Inbetween any two allowed energy
bands, there are forbidden bands,
which consist of forbidden energy
levels

The energy band diagram is

characterized by the total number of
bands which it involves, their
occupancy, size and the forbidden
gap between them

4s band has an occupancy of 2N

states, however, it contains only N
number of electrons occupying the
lower levels. 4s band is thus partially
(half) filled

the higher band have greater width as compared to the lower bands
the upper bands are due to stronger interaction of valance electrons. As we move down, the
lower level electrons interact weakly due to greater distance and screening effect of the
upper level electrons.
Due to larger distance, the wavefunctions of these electrons overlap to a lesser extent.
As a result the splitting effect is weak and therefore the bands are narrower
 VALANCE BAND, CONDUCTION BAND AND FORBIDDEN GAP

The electrons in the outermost orbit of any element are called as valance electrons

The group of energy levels of the valance electrons constitutes a valance band

topmost or highest occupied band

All the bands above the valance bands are empty

In the group of unoccupied bands, the lowest unoccupied band is called as conduction band

When the electrons in the valance band are bound to their parent atoms, but when these
electrons are excited to conduction band, they become free and thus take part in conduction

Between the valance band & conduction band, there is a group of forbidden energy levels.
Quantum mechanically these levels are not allowed to the electrons.

This group is called as a forbidden gap or the band gap

 Lithium (Z =3)
1s2 2s1
Energy level diagram consists of 1s level having two electrons and 2s level having 1 electron.
Lithium atoms come together and form bonds, the energy levels split and form energy bands

There are two bands, namely 1s band and 2s band.

1s band has 2N states filled by 2N electrons. Thus 1s band is completely filled.

2s band has 2N available states, but as the 2s level for lithium atom contains only 1 electron, the
2s band contains only N number of electrons

According to auf-bau rule, these electrons fill up the states in the increasing order of
energy. Thus the lower 2N states are filled with 2N electrons and upper 2N states remain
unoccupied.

Thus 2s band of lithium is partially (half) filled.

The lower filled part itself is the highest occupied band that it valance band, while in the same
2s band, the upper unfilled band is the lowest unoccupied band that is conduction band
the gap between these two bands is almost 0 eV (actually 10-28 eV)
According to kinetic theory, the thermal energy of the electrons is given by (3/2)kT.
At 0K, there is no energy available for the electrons to rise in the conduction band,

at 300 K, the thermal energy is 0.038 eV, which is sufficient for the electrons to go in the
conduction band.
It may be noted that there are no forbidden levels in between the valance and the conduction
band.

The allowed but empty energy levels in the conduction band are in the vicinity of the filled level
in the valance band.
When the electrons are excited to the conduction band, they become free.

As the valance electrons can be made free very easily, lithium, at room temperature, has plenty
of electrons in conduction band, and consequently, lithium behaves as conductor.

Lithium belongs to group I in the periodic table.

The other elements in group I, such as sodium (Z= 11), potassium etc. also have partially filled
valance bands and thus are conductors.

The other examples of the elements with partially filled valance band are, aluminium (Z =13),
copper(Z = 29), silver (Z = 47) and gold (Z = 79). As we know, all these elements are good
conductors of electricity.

Any element having partially filled valance band should behave like a conductor
 Beryllium (Z = 4)
1s2 2s2 2p...

we expect that the 2s band of the beryllium should be completely filled and should behave like
the valance band.
The next band should be 2p, it should be empty and should behave like a conduction band.
We expect that because of the existence of a finite gap (containing forbidden levels) between the
valance and conduction band, beryllium should behave like an insulator.
But behaves like a conductor.
Beryllium atoms come close, the 2s and 2p bands split.
As the distance approaches equilibrium separation, the 2s and 2p band expand and at certain
stage they merge (overlap) in to each other.
Thus instead of separate 2s and 2p bands, we get 2s-2p hybrid band having 2N + 6N = 8N
available states. Out of these 8N states only lower 2N states are filled while the upper 6N states
are unfilled.
Thus 2s-2p hybrid band is partially filled and thus beryllium behaves as a conductor.
Beryllium belongs to group II elements in the periodic table.
 INSULATORS
ex: Diamond, band gap of 5.7 eV.

The thermal energy of the valance electrons at 300 K is approximately 0.038 eV

This is quite insufficient for the valance electrons to cross the band gap and enter in the conduction band

The electron cannot be excited to any energy level within the band gap, as these energy levels are forbidden.

Thus the concentration of the free electrons in diamond is extremely poor it is an insulator

Thus thermal energy at room temperature is not sufficient to impart conducting properties
to diamond.

Imparted any energy is immediately lost by the electron during it’s collisions with the crystal
imperfections

. It can be shown that an electric field of 108V/m is required to impart 5.7 eV energy to the electron by considering
the energy losses due to frequent collisions.

The band gap of diamond is 5.7 eV

The energies of the photons in the visible range is 1 to 3 eV. Thus light
cannot be absorbed in diamond
Therefore it is transparent to the light
 CONDUCTORS

Due to either partially filled valance band or overlapping of valance and conduction bands,
there is no forbidden gap in the conductors

At absolute zero, the energy of electrons is not sufficient to enter the conduction band

At room temperature, due to thermal energy, the electrons just below the Fermi level can be
easily excited to the conduction band.

When electric field is applied, the electrons start moving an constitute the current.

The resistivity of the conductors (metals)

n - Number density of the electrons in the conduction band

τ - The relaxation time, that is the average time between the successive collisions.

Zero band gap, all valance electrons are set free at room temperature
When the temperature is raised further, this sea of electrons is thermally agitated offers
resistance to the flow of conduction electron and increases with the temperature
Conductors thus have Positive temperature coefficient of resistance
 SEMICONDUCTORS
Band gap of silicon --- 1.12 eV
Band gap of germanium --- 0.72 eV
Band gap of insulator > Band gap of Semiconductor > Band gap of of conductor

At room temperature, the thermal energy raises a small number of electrons in to conduction band

free electron density

silicon 1016 /m3 < copper (1028 /m3)

semiconductors exhibit a small conductivity better than insulators

the concentration of the charge carriers in the semiconductors can be increased significantly by
adding impurities

Semiconductors have two kinds of charge carriers

When an electron is excited to the conduction band, it leaves a vacancy in the valance band. This
vacancy is called hole (positive charge carrier).
Thus semiconductors have bipolar conduction (electrons and holes).

Thermistors: (Temperature dependent resistors)

When heat is applied, the electrons are thermally excited to the conduction band. Thus resistance
of the semiconductors decreases with temperature.
Semiconductors are thus said to have Negative temperature coefficient of resistance

Semiconductors can convert light in to electricity and vice versa

 Statistical description of electrons in semiconductors

When we consider a system of many particles (such as molecules, photons or electrons), we need a statistical
approach to describe their properties. There are three such approaches, in all of which we use certain laws to
describe a probability P(E) that a particular particle will have an energy E.

Maxwell-Boltzmann Statistics: This statistics is considered as a ‘classical’ statistics. It is mainly applicable to the
molecules in the gas, which are considered as identical but distinguishable particles, whose wavefunctions do not
overlap appreciably. In this statistics, there is no limit to the number of particles that can exist in a state of given
energy.
Maxwell Boltzmann statistics clearly explains raining which occurs due to evaporation of the water molecules on the
sea surface. It also explains the sunshine, which occurs due to fusion of high energy protons in it.

This statistics is applicable only in the situations where the temperature is high enough and density is low enough so
that quantum effects can be ignored.

Bose Einstein statistics: This is a quantum statistics and is applicable to photons; the spin zero particles. It is also
applicable to the particles having integer spin.

Photons are also called bosons (in memory of Satyendranath Bose, an Indian Physicist) and they do not obey Pauli’s
exclusion principle. Photons are identical particles, but they are indistinguishable as their wavefunctions overlap
considerably.

This statistics explain all properties of photons, for example the black body spectrum.
 FERMI DIRAC STATISTICS

It is the statistical description of an electrons in semiconductors It gives the

probabilityof an electron to have an energy E
It is a quantum statistics, which is applicable to all particles having odd half integral spin
i.e. 1/2,3/2, 5/2..
Ex: Electron

In this statistics electrons are treated as wave, which obeys Pauli’s exclusion principle-- Fermions

Fermions are identicle but indistinguishable, whose wave functions overlap considerably

The wave functions of Fermions change the sign when electrons in any pair are exchanged
and thus are called as antisymmetric wave functions

The probability P(E) that fermion occupies a quantum state of energy E, at Temparature T is given by

k- Boltzman Constant EF- Fermi energy

Probability of occupancy of higher energy levels increases with temperature

The above formula clearly indicates that the probability of occupancy of higher energy levels increases with
temperature.
 FERMI ENERGY

The energy of the top most filled level in the ground state of the N-electron
system
EF= (ħ2/2m) (Nπ/2L)2= (ħ 2/2m) (nFπ/L) 2

N=2nF
nF --- Uppermost filled Level

An electron of mass m confined to a length L

Fermi energy plays important role in determining properties of semiconductors

Fermi Level

The highest occupied energy level of the electrons in the valance band at 0o K

At 0oK all energy states upto E are occupied and all states above E are unoccupied
F F
(empty)
Probability P(E) that fermions occupies a quantum state of energy E, at Temparature T is given
by

k- Boltzman Constant
EF- Fermi energy

At T=0K, there are two possibilities

E>EF or E<EF
At T=0K and E>EF we have (E-EF) >0 Positive, Thus

Thus at T = 0 K, the probability of occupancy of all the levels above the Fermi level is zero.
Thus at T = 0 K, all the energy levels above the Fermi level are certainly unoccupied
At T=0K and E<EF we have (E-EF) <0 Negative, Thus

Thus at T = 0 K, the probability of occupancy of all the energy levels below the Fermi level
is one.
Thus at T = 0 K, all the energy levels below the Fermi level are certainly occupied.

At T=0K and E=EF we get

This not possible

At T>0K and E=EF we get

Thus the probability of the occupancy of the Fermi level at any finite (nonzero)
temperature is Always 50%.
This is applicable to metals, semiconductors as well as insulators
Physical significance of Fermi level Depending on temperatures
The significance of Fermi level in conductors (metals) at T = 0 K
The significance of Fermi level in conductors (metals) at T > 0 K
Graphical representation of Fermi Dirac statistics. Physical significance of the Fermi level
Density of states at T = 0 K

Density of states = number of quantum states n(E), per

unit energy interval
 Density of states at T > 0 K

The probability of the occupancy of the levels above the fermi level increases at the cost of
energy levels below EF.
Thus density of the occupied states above EF increases at the cost of density of occupied
states below EF

The graphs at T>0K are almost similar to the graphs of T=0K indicates that even at higher
temperatures, modifications take place only in the closet vicinity of Fermi level.
 Fermi Level in Intrinsic Semiconductors

At 0oK highest occupied level is fermi level

At any non-zero temperature, Fermi level is an average level whose probability of occupancy is
50%

EV and EC are the average energies of the electrons in valance and conduction bands

Intrinsic semiconductors, the Fermi level is situated in the middle of the band gap
 Fermi Level in N-Type Semiconductors

There are donor energy levels just below the conduction band.
At 0oK, these levels are occupied by unshared electrons of pentavalent impurity.

Majority of the electrons in the conduction band come from these donor levels.

At 0 K, all these electrons exist in donor levels only. Therefore the Fermi level of N type of
semiconductor at 0 K is situated in the middle of the donor levels

EC ≥ EFNT=0K ≥ ED

EC --- lowest energy level in the conduction band

ED --- lowest energy level amongst the donor levels

As every phosphorous atom ‘donates’ one free electron, it is called as‘donor impurity’. Unlike the
4 shared electrons, the energy levels of unshared electrons cannot be represented in the valance
band, as it would mean that their energy gap is 1.1 eV. The negligible energy required to detach
the unshared electrons indicates that the energy levels ofthese electrons be shown just below the
conduction band. These energy levels are called donor levels
 ED ≥ EFN T > 0K ≥ EFi

Temperature is raised, the electrons in the donor levels start gaining thermal energy and they
move to conduction band.
Thus the donor levels are gradually depleted.
As a result, the Fermi level starts moving down.

Temperature is further raised, the electrons in the valance band also start going in to the
conduction band.
This results in generation of holes in valance band.
To account for this, the Fermi level further shifts down.
At moderate temperatures, the Fermi level can never reach the intrinsic Fermi level,
in between the lowest donor level and intrinsic Fermi level
At extremely higher temperature, almost all electrons in the valance band are excited to the
conduction band.
These electrons are extremely large in number as compared to the electron donated by the donor
levels.
Thus at very high temperatures, the concentration of the electrons in conduction band and
holes in the valance band is almost equal.

This indicates that at very high temperatures, N type semiconductor is converted to intrinsic
semiconductor.

In this situation the Fermi level of the N type of semiconductor coincides with the intrinsic Fermi
level.
 Effect of dopant concentration on Fermi level of N type semiconductor

At moderate concentration, the distance between the donor atoms is large enough.
Interaction between the donor atoms is also weak.
Therefore the donor levels do not split and they remain discrete.
If the concentration is further increased the distance between the donor atoms decreases
and they start interacting.

----> the donor levels interact and they become a donor band.
If the concentration is still increased, donor levels split to a greater extent and the donor
band expands and invades in the conduction band.

As a result the Fermi level also enters in the conduction band

 Fermi Level in P-Type Semiconductors

EF in P-type is opposite to N-Type semiconductor

There are acceptor levels near the valance band.
The holes in P type semiconductor are only due to transitions of electrons from valance band to
the acceptor levels.
As the valance band is the source of electrons and acceptor levels are recipients of them,
the Fermi level is situated in the middle of the group of acceptor levels

EA ≥ EFPT=0K ≥ EV
EA is the highest acceptor level
EV is the highest level in the valance band
Temperature is raised, the electrons in the valance band gain thermal energy to raise in
the acceptor levels.

As a result, concentration of electrons in the acceptor levels increases and the Fermi
level moves upward.

As the temperature is further increased, at moderate temperatures, the electrons in the

valance band are thermally excited to the conduction band.
Thus, concentration of the electrons in the conduction band increases.
As a result, the Fermi level further shifts up

EFi ≥ EFP T > 0K ≥ EA

at moderate temperatures, it cannot touch the intrinsic

Fermi level
if the temperature is extremely high, then the almost all the
electrons in the valance band are excited to the conduction
band.

concentration of intrinsic charge carriers overcomes the

concentration of holes due to acceptor impurity.

concentration of electrons in the conduction band and

holes in the valance band becomes almost equal.

Thus at extremely high temperatures, the P type

semiconductor is converted in to intrinsic
Effect of dopant concentration on Fermi level of P type semiconductor

At very low doping level, the distance between

the acceptor impurity atoms is comparatively
large.
Interaction between the acceptor impurity atoms
is weak
acceptor levels are discrete.

Doping level is increased, the atom-to-atom

distance of the acceptor impurity decreases.
They interact and thus the acceptor levels split.

Doping level is further increased, the acceptor

levels become acceptor band and then it
penetrates in the valance band
The current which occurs due to a gradual drift of electrons towards the positive terminal (or
holes towards the negative terminal) is called as drift current, The corresponding velocity is
called as drift velocity (vd )

Drift velocity depends upon the applied electric field. A quantity which is more fundamental and
independent of the electric field is mobility. It is defined as drift velocity acquired per unit electric
field.

Mobility

Current= I = nqvd A

n-- carrier concentration/charge particle density/number of charge carriers per unit volume
J - current density
Fundamental expression which expresses conductivity in terms of carrier concentration, and mobility.

Conductivity = σ = nqμ
As metals contain only one kind of charge carrier (electron), we can use these Eqn. as it is for metals.
However, semiconductors contain two kinds of charge carriers; electrons and holes. Thus, for semiconductors,
we must write
σi = nie(μe + μh )
Where e represents the charge on electron (as well as hole), 𝑛𝑒 and 𝑛ℎ represents carrier concentration of
electrons and holes and 𝜇𝑒 and 𝜇ℎ represent mobility of electrons and holes respectively.

Now recalling that the carrier concentrations of electrons and holes in intrinsic semiconductors are equal,
we have𝑛𝑒 = 𝑛ℎ = 𝑛𝑖. Thus
𝜎𝑖 = 𝑛𝑖𝑒(𝜇𝑒 + 𝜇ℎ)
We know that for N type semiconductors 𝑛𝑒 ≫ 𝑛ℎ and for P type semiconductors 𝑛ℎ ≫ 𝑛𝑒. Thus we have

𝜎N = 𝑛e𝑒𝜇𝑒
𝜎P = 𝑛h𝑒𝜇ℎ