8/20/2024

Semiconductor Devices
PH 205

Prof. Dr. P. K. Giri

Professor of Physics
Indian Institute of Technology Guwahati
Guwahati, India
[Link]

Syllabus
Physics: Energy bands in solids and Charge carriers. Semiconductors: Elemental and
compound semiconductors, intrinsic and extrinsic materials, Direct and indirect band-gap
semiconductors, Heavily doped semiconductors. Charge carriers in semiconductors:
mobility, impurity band conduction, nonlinear conductivity, excess carriers in
semiconductors. Semiconductor Bloch equation, transport properties.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
P-N junctions: fabrication, static and dynamic behavior of p-n junction diodes, Junction
breakdown in p-n junctions, tunnel diode, Schottky diode.
Metal Semiconductor junctions: Schottky effect, rectifying and Ohmic contacts.
Integrated circuits, fabrication methods
Optoelectronic Devices: photodiodes, light emitting diodes, semiconductor lasers,
photovoltaic cells
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Bipolar Junction Transistor: fundamentals of BJT operation, BJT fabrication, carrier
distribution and terminal current, generalized biasing, switches, frequency limitations of
transistors. Field Effect Transistors: JFET, MOSFET.
Texts:
1. S. M. Sze, Physics of Semiconductor devices, 2nd Ed., John Wiley, 1982.
2. J. Singh, Semiconductor Devices - Basic Principles, John Wiley, 2001.
References:
1. M. S. Tyagi, Introduction to Semiconductor Materials and Devices, John Wiley, 2008.
2. B. G. Streetman, Solid State Electronic Devices, 5th Ed., PHI, 2001.
2
Some knowledge of Quantum Mechanics and Statistical Physics will be needed to follow the course.

1
 8/20/2024

Evaluation Scheme

• Quizzes: 2 nos (10 marks each)

• Mid-sem: 30 marks

• End-sem: 50 marks

• Attendance (min 50%) & class response

• Relative grading
4

2
 8/20/2024

Semiconductor Components in Electronic Devices

• Transistors: These are fundamental semiconductor devices
used for amplification, switching, and signal processing.
• Diodes: Semiconductor components that allow current to
flow in one direction only.
• Integrated Circuits (ICs): Complex assemblies of
semiconductor components that perform specific functions.
• Light-Emitting Diodes (LEDs): Semiconductor devices that
emit light when current passes through them.
• Solar Cells: Semiconductor devices that convert light energy
into electrical energy.
• Rectifiers: Semiconductor components that convert AC to
DC. 5

Computer Motherboard

Integrated circuit
6

3
 8/20/2024

Properties that gives Semiconductors

unique abilities to produce intelligent
devices

Band structure:
Valence band, conduction band and region of
forbidden bandgap

Controlled introduction of dopants/impurities

to alter material conductivity by up to 10
orders of magnitude

Isolation of regions of a device and electrical

contacts

Small dimensional structures for high speed/high

density
7

Device Building Blocks

• Semiconductor devices have been studied for over 130 years.
• About 60 major devices with over 100 device variations related to them.
• All these devices can be constructed from a small number of device
building block.

• Fig.1: Basic device building blocks. (a) Metal-semiconductor interface; (b)

p-n junction; (c) heterojunction interface; (d) metal-oxide-semiconductor
structure(MOS).

4
 8/20/2024

• Figure 1a is a metal-semiconductor interface,

which is an intimate contact between a metal
and semiconductor.
• This building block was the first
semiconductor device ever studied in 1874.

• The second building block is the p-n junction (Fig1b),

formed between a p-type (with positively charged carriers)
and an n-type (with negatively charged carriers) semiconductors.
• The p-n junction is a key building block for most
semiconductor devices.

5
 8/20/2024

• The third building block (Fig1c), is the

heterojunction interface, an interface formed
between two dissimilar semiconductors.
• the key components for high-speed and
photonic devices.

• Figure 1d shows the metal oxide

semiconductor (MOS) structure.
• The structure can be considered a
combination of a metal-oxide interface and an
oxide-semiconductor interface.

6
 8/20/2024

Major Semiconductor Devices: History

• 1874: metal-semiconductor contacts studied by Braun,
who discovered that the resistance of contacts between
metals and metal sulfides (e.g., copper pyrite, CuS)
depended on the magnitude and polarity of the applied
voltage.
• 1907: The electroluminescence phenomenon (for the
light-emitting diode) was discovered by Round. He
observed the generation of yellowish light from a crystal
of carborundom (SiC) when he applied a potential of 10 V
between two points on the crystals.
• 1947: The point-contact transistor was invented by
Bardeen and Brattain.

• 1949: This was followed by Shockley’s classic paper on p-n

junction and bipolar transistor which is a key
semiconductor device (Fig. 2).

1956 Nobel Prize

Fig.2 The first transistor. (Photograph courtesy of Bell

Laboratories, USA)

7
 8/20/2024

• 1952: Ebers developed the basic model for the thyristor,

which is an extremely versatile switching device.
• 1954: The solar cell was developed by Chapin, et al. using a
silicon p-n junction..
• 1957: Kroemer proposed the heterojunction bipolar
transistor to improve the transistor performance. Nobel Prize
2000
• 1958: Esaki observed negative resistance charecteristics in a
heavily doped p-n junction, which led to the discovery of the
tunnel diode (Nobel Prize 1973)
• 1960: The most important device for advanced integrated
circuits is the MOSFET which was reported by Kahng and
Atalla.

The MOSFET and its related

integrated circuits now
constitute about 90% of the
semiconductor device market.
An ultrasmall MOSFET with a
channel length of 20 nm had
been demonstrated in 2001
which can serve as the basis
for the most advanced
integrated chips containing
over one trillion (>1012)
devices

Fig.3 The first metal-oxide semiconductor field-effect transistor

MOSFET (Photograph courtesy of Bell Laboratories, USA.)

8
 8/20/2024

• 1962: Hall et al. first achieved lasing in semiconductors.

• 1963: Kroemer, Alferov and Kazarinov proposed the

heterostructure laser. These proposals laid the foundation for
modern laser diodes, which can be operated continuosly at room
temperature. Laser diodes are the key components for a wide range
of applications, including digital video disk, optical fiber
comunication, laser printing, and atmospheric-pollution monitoring.
During this year, Gunn invented transferred-electron diode which is
used in such millimeter-wave applications as detections systems,
remote controls, and microwave test instrument.

• 1965: IMPATT diode’s operation was first observed by Johnston et

al. They can generate the highest continuous wave (CW) power at
millimeter-wave frequencies of all semiconductor devices. They are
used in radar and alarm systems.

• 1966: Mead invented MESFET

which is a key device for
monolithic microwave
integrated circuits (MMIC).
• 1967: Kahng and Sze invented
the nonvolatile semiconductor
memory (NVSM) which can
retain its stored information
when the power supply is
switched off. A schematic
diagram of the first NVSM is
shown in Fig.4a. Fig.4

• 1994: The operation of a SEMC ( a limiting case of the floating-

gate NVSM, single-electron memory cell (Fig.4b)) at room
temperature was first demonstrated by Yano et al.

9
 8/20/2024

• 1970: Boyle and Smith invented the charge-coupled

device (CCD) which is used extensively in video cameras
and in optical sensing applications.
• 1974: The resonant tunneling diode (RTD) was first
studied by Chang et al. RTD is the basis for most
quantum-effect devices.
• 1980: Minura et al. developed the MODFET (modulation-
doped field-efffect transistor). With the proper selection
of heterojunction materials, the MODFET is expected to
be the fastest field-effect transistor.

Key Semiconductor Technologies

• 1798: The lithography process was invented. In this first
process, the pattern, or image, was transferred from a
stone plate (litho).
• 1918: Czochralski developed a liquid-solid
monocomponent growth technique. The Czochralski
crystal growth is the process used to grow most of the
crystals from which silicon wafers are produced.
• 1925: Another growth technique was developed by
Bridgman. The Bridgman crystal growth has been used
extensively for the growth of gallium arsenide and related
compound semiconductor crystals.
• 1952: Welker noted that gallium arsenide and its related
III-V compounds were semiconductors.

10
 8/20/2024

• 1952: The diffusion of impurity atoms in semiconductors is important for

device processing. The basic diffusion theory was considered by Flick in
1855. The idea of using diffusion techniques to alter the type of
conductivity in silicon was disclosed in a patent by Pfann.
• 1957: The ancient lithography process was applied to semiconductor-
device fabrication by Andrus. He used photosensitive etch-resistant
polymers (photoresist) for pattern transfer.
Lithography is a key technology for the semiconductor industry. The
continued growth of the industry has been the direct result of
lithographic photoresist technology. Lithography is also a significant
economic factor., Currrently representing over 35% of the integrated-
circuit manufacturing cost.
In the same year the epitaxial growth process based on chemical vapor
deposition technique was developed by Sheftal et al.
Epitaxy, derived from the Greek word epi, meaning on, and taxis, meaning
arrangement, describes a technique of crystal growth to form a thin layer
of semiconductor materials on the surfce of a crystal that has a lattice
structure identical to that of the crystal. This method is important for the
improvement of device performance and the creation of novel device
structures.

• 1959: An integrated circuit (IC)

was made by Kilby. (Nobel Prize
2000)
• Also, Noyce proposed the
monolithic (single stone) IC by
fabricating all devices in a single
semiconductor substrate and
connecting the devices by
alluminum metallization. The
alluminum interconnection lines
were obtained by etching
evaporated aluminum layer over
the entire oxide surface using the
lithographic technique. Fig.5

• 1967: The dynamic random access memory (DRAM) was

invented by Dennard. The memory cell contains one MOSFET and
one charge-storage capacitor. The MOSFET serves as a switch to
charge or discharge the capacitor.

11
 8/20/2024

Technology Trends

24

12
 8/20/2024

25

Intel Goal- Internal Report

26

13
 8/20/2024

Important dates in the history of Semiconductor Physics:

Year Development
1879 Discovery of Hall effect, +ve Hall coefficient
1926 Bloch’s theorem
1931 Wilson lays the foundation of the modern theory of semiconductors as
insulators with narrow forbidden bands and introduced the idea of hole

1948 Discovery of the transistor by Bardeen & Brattain

1960 Appearance of planar technology
1982 World production of 3x1013 binary units of active memory, 64 kilo bytes
1990 Manufacture of DRAM of 4 MB per chip
1991 High definition TV camera with 2 MP CCD sensor
1992 Semiconductor component sales worldwide $60 billion
1993 World production of transistors: 2x1017
1995 Manufacture of 64-MB DRAM

1985 Nobel Prize: Integer Quantum Hall Effect, 1998 Nobel Prize: Fractional Quantum Hall Effect
27

Elemental Semiconductors: Si, Ge

28

14
 8/20/2024

Q.M. of the free electron, moving in the absence of any potential.

Electrons inside the semiconductor can be regarded as “free” under

proper conditions.

Schrodinger equation (time independent)

Ψ(r) = , factor comes because we have one

electron per volume V, i.e.

It is assumed that V is a cube of side L.

According to classical mechanics, , p is a continuous

variable.
p=ħk in quantum mechanics due to wave nature of electrons, k is
discrete, not continuous.

Physical conditions

Two kinds of boundary conditions are imposed on wave functions

Ψ 0 at the boundary.

So, sin (kxx) = 0 for x = L.

L
kx=π/L, 2π/L, 3π/L,…… , k → positive values.
Stationary electron confined in a finite region, like quantum well: Standing wave solution

15
 8/20/2024

Periodic boundary condition for moving electrons,

Ψ(x, y, z+L)= Ψ(x, y, z) for a volume V.

The boundary condition imposes certain

restrictions in k of the wave function.

Energy

Momentum

Velocity
Because of the boundary conditions, allowed values of k are ( nx,
positive and negative both)

kx = , ky = , ………

If L is large, k spacing is small.

Volume in k-space that each electronic state occupies

No. of electronic states in

volume Ω of k-space

16
 8/20/2024

Density of states (DOS) for a 3-dimensional system:

The concept of density of states is extremely powerful, and
important physical properties, such as optical absorption, transport, etc.,
are intimately dependent upon this concept. The density of states is the
number of available electronic states per unit volume per unit
energy around an energy E. If we denote the density of states by
N(E), the number of states in a unit volume in an energy interval dE
around an energy E is N(E)dE.
To calculate the density of states, we need to know the dimensionality
of the system and the energy vs. k relation that the electrons obey.

If density of states is N(E), No. of states in dE interval= N(E) dE.

For the calculation of DOS, we need to know the dimensionality
and E-k relation. For free electron

In the energy range E and E+dE, no. of electrons:

Accounting for spin and effective mass, the DOS in 3-D is

ye root of 2 uppar kaise chala gaya myu = m^2/2m

17
 8/20/2024

Density of states in 2-D and 1-D:

Let us now consider a 2-D system, a
concept that has become a reality with the
use of quantum wells. Similar arguments
tell that the density of states for a parabolic
band is (including spin)

Finally, in a 1-D sysetm or a ‘quantum wire’

the density of states is (including spin)

We note that as the dimensionality

changes, the energy dependence of the Fig. 2: Geometry used to calculate
density of states also changes. As shown density of states in 3-, 2-, and 1-
in Fig. 3, for a 3-D system, we have a E1/2 dimensions. By finding the &-space
dependence, for a 2-D system we have no volume in an energy interval
energy dependence; and for a 1-D system between E and E + dE, one can find
we have an E-1/2 dependence. out how many allowed states there
are.

Fig.3: Variation in the energy dependence

of the density of states in (a) 3-dimensional,
(b) 2-dimensional, and (c) 1-dimensional
systems. The energy dependence of the
density of states is determined by the
dimensionality of the system. This feature is
exploited in advanced
semiconductor devices. .

18
 8/20/2024

Filling of electronic states

• How do the electrons distribute themselves among
various allowed electronic states?
• Thermodynamics tells us that a system will be in
equilibrium when its free energy is minimum.
Free energy F = U – T S , where U is the internal energy and
T S is the entropy energy (disorder related).
• In classical physics, all particles are distinguishable
and can occupy the same state. Occupation
distribution is called Boltzmann distribution:
𝐸−𝐸𝐹
f(E) = exp(− )
𝑘𝑇
37

Distribution function: Fermions

• In QM, spin of the quantum particle in important.
• Particles with half-integral spin (electrons, protons,
neutrons, etc..) are called Fermions.
• Particles with zero/integral spin (photons, mesons
etc) are called Bosons.
• A particular electronic state can only be occupied by
a single Fermion. But, since electrons can have +h/2
and –h/2 spins, two electrons can sit in a particular
energy level.
38

19
 8/20/2024

Fermi Distribution Function

𝟏
• f(E) = 𝑬−𝑬𝑭
𝟏+𝐞𝐱𝐩( )
𝒌𝑻
where EF is called Chemical potential or Fermi energy.
• The distribution function f(E) gives us the probability
that the state with energy E is occupied.
• EF can be determined once the density of states and
the electron density are known.
∞ 𝑁 𝐸 𝑑𝐸
• Electron density n = ‫׬ = 𝐸𝑑 𝐸 𝑓 𝐸 𝑁 ׬‬0 𝐸−𝐸𝐹
exp +1
𝑘𝑇
39

Fermi distribution
function at various
temperatures.
• At T=0, it’s a step
function.
• As the temperature
increases, the function
smears out.
• At E=EF, f(E) = 0.5

𝟏
At T=0, f(E) = 𝑬−𝑬𝑭 =𝟏 𝒊𝒇 𝑬 𝑬𝑭
𝟏+𝐞𝐱𝐩( )
𝒌𝑻

= 0 otherwise.

40

20
 8/20/2024

ℏ2
𝐸𝐹 = (3𝜋 2 𝑛)2/3
2𝑚0
This expression is applicable to metals, such as Cu, Au,
Ag, etc. It’s not used for semiconductors.

At finite temperatures, it's not easy to determine n in

terms of EF.
If the electron density is small so that f(E) is small, then
we can neglect 1 in the denominator and then f(E) is like
Boltzmann function.

Where Nc is called the effective

density of states. 41

Joyce-Dixon Approximation
• If the electron density is high so that f(E) approaches 1
(degenerate condition), then using Joyce-Dixon
approximation, we get:
𝒏 𝟏 𝒏
𝑬𝑭 = 𝒌𝑻[𝒍𝒏 + ]
𝑵𝒄 𝟖 𝑵𝒄
• In case of 2D semiconductor, where the DOS N(E) is
constant, EF can be evaluated exactly when carrier
concentration n is known.

42

21
 8/20/2024

Periodicity of a crystal:
Crystal structure → lattice + basis
The various kinds of lattice structures possible in nature are
described by the symmetry group that describes their
properties. Rotation is one of the important symmetry
groups. Lattices can be found which have a rotation
symmetry. The rotation symmetries are denoted by 1, 2, 3, 4,
and 6.
There are 14 types of lattices in 3D. These lattice classes are
defined by the relationships between the primitive vectors a1,
a2, a3 and the angles , ,  between them. Table 1 provides
the basic properties of these three dimensional lattices.
We will focus on the cubic lattice, which is the structure
taken by all semiconductors.

Table 1: The 14
Bravais lattices
in 3-dimensional
systems and
their properties.

22
 8/20/2024

There are 3 kinds of cubic lattices: simple cubic (sc), body centered
cubic (bcc), and face centered cubic (fcc).

The fcc and bcc lattices are of great importance, since an enormous variety of
solids crystallize in these forms with an atom (or ion) at each lattice site.
Essentially all semiconductors of interest for electronics and
optoelectronics have fcc structure.

The Diamond and Zinc Blende Structures

Most semiconductors of interest for electronics and optoelectronics have an

underlying fcc lattice. However, they have two atoms per basis. The
coordinates of the two basis atoms are

Since each atom lies on its own fcc lattice, such a two-atom basis structure
may be thought of as two inter-penetrating fcc lattices, one displaced from
the other by a translation along a body diagonal direction

Fig. 4: (a) Diamond (b) zinc blende crystal structure. The structure
consists of the interpenetrating fcc lattices, one displaced from the other by a
distance (a/4,a/4,a/4) along the body diagonal. The underlying Bravais lattice
is fcc with a two atom basis. The positions of the two atoms is (000) and
(a/4,a/4,a/4).

23
 8/20/2024

Figure 4 gives details of this important structure. If the two atoms of the basis
are identical, the structure is called diamond. Semiconductors such as Si,
Ge, C etc., fall in this category.
If the two atoms are different, the structure is called the Zinc Blende
structure. Semiconductors such as GaAs, AlAs, CdS, etc., fall in this category.
No. of atoms per unit cell is 8.

Semiconductors with diamond structure are often called elemental

semiconductors, while the Zinc Blende semiconductors are called compound
semiconductors.
The compound semiconductors are also denoted by the position of the atoms
in the periodic chart, e.g., GaAs, AlAs, InP are called III-V (three-five)
semiconductors while CdS, HgTe, CdTe, etc., are called II-VI (two-six)
semiconductors.
Some semiconductors, such as ZnO, crystallizes in the wurtzite structure or
hexagonal structure (Fig. 5).
In a periodic structure, electrons do not scatter at all from ions. The
effect of it is to alter the apparent E-k relation of the electron,
change the mass, introduce bandgap, and even make electrons
behave like positively charged particles, i.e., holes.

Two intertwined hexagonal close-packed Zn and O lattices; each Zn2+ ion is coordinated by
four O2- ions in a tetrahedral arrangement.

ZnO wurtzite cell: Zn in yellow, O in grey, the tetrahedral coordination is

shown for both types of atoms.

Fig. 5

24
 8/20/2024

In a periodic structure , electrons do not scatter at all from ions. The effect
of it is to alter apparent E-k relation of the electron , change the mass,
introduce bandgap and even make electrons behave like positively
charged particles.
Notation to Denote Planes and points in a Lattice: Miller Indices

Miller indices:
(i) Define x , y , z axis, (ii) Take the intercepts of planes along the axes in unit of
lattice constant, (iii) take the reciprocal of the intercept and reduce them to
smallest integers.

The notation {h, k ,l} - family of

parallel planes/ equivalent
planes.

{100}, {010}, {001} are

equivalent planes.

49

In a cubic system, Miller indices of a plane are the same as the direction
perpendicular to the plane.

[ ] is for a set of parallel directions.

< > is for a set of equivalent directions.

Wigner-Seitz cell: The primitive cell of a reciprocal lattice can be

represented by a Wigner-Seitz cell.

Defined in the reciprocal lattice-

(i) Draw lines to connect a given lattice point to all neighboring points.
(ii) Draw bisecting lines or planes to the previous lines.
The volume enclosed within the bisection is called Wigner Seitz cell.

Reciprocal Lattice: Real space to k-space

50

25
 8/20/2024

The Wigner-Seitz cell is constructed by drawing perpendicular bisector planes in the

reciprocal lattice from the chosen center to the nearest equivalent reciprocal lattice sites. This
technique can also be applied to a direct lattice. The Wigner-Seitz cell in
the reciprocal lattice is called the first Brillouin zone.

Fig.6: Brillouin zones for (a) fcc, diamond, and zincblende lattices, (b) bcc lattice, and (c)
wurtzite lattice.

51

Brillouin zone → Wigner seize cell in reciprocal lattice

Effective mass:

for free electron, k2 determines the curvature of E-k curve.

1/m determines the curvature.
In real solids, the E-k curvature determines the effective mass m*,
which is usually less than the free electron mass.
Electrons in periodic potential:
Properties of electrons in a perfectly periodic crystal

U(r) has same periodic as crystal U(r)=U(r+R), For U(r)=0,

p=ħk

Bloch’s theorem states that for a periodic potential, uk has the same periodicity as the periodic
potential.

52

26
 8/20/2024

Kronig-Penny model for Band structure

U(x) Fig. 7: The periodic
potential to study the
band structure in the
Kronig-Penny model.
The potential varies
between o and U0 as
shown and has a
U(x) = 0, 0≤x≤a periodicity of d=a+b.
= u0 -b < x ≤ 0
where , Φ=kxd

Ψ(x)={

Where,
From continuity condition of wave function
derivative at x=0 and x=a

We obtain, A+B=D+F
β(A-B)= α (D-F)
53

Non-trivial solution for the variables A, B, D, F are obtained only if the determinant
of the co-efficient vanishes, which gives

if 0<E<u0

if E>u0

where,

Energy is physically allowed only if -1cosΦ ≤ +1

If E<u0 , R.H.S. is denoted by f(E)

The function must lie between -1 and +1, since it is equal to cos (=cos kxd) . We wish
to find relationship between E and  or E and kx. Graphically one can see (Fig.
8(a)) that f(E) remains between ±1 bounds only for certain region of energies.
These allowed energies for the allowed bands and are seperated by bandgaps. One
can obtain the banstrcture as shown in Fig. 8(b).
54

27
 8/20/2024

Fig. 8: (a) The

graphical solution to
obtain the allowed
energy levels. (b) The
allowed and forbidden
bands are plotted in
the E vs. k relation
using the results from
fig. a. This E-k
description is called
band structure.

55

• Crystal momentum

We know p=ħk , crystal momentum for free space.

crystal momentum inside the crystal.
Effect of the crystal is to modify the E-k relationship.
wavelength is related to lattice spacing.
K vectors are confined to Brillouin zone.
• Important high symmetry points
Γ point kx=ky=kz=0, origin of the Brillouin zone.
X point , ky=kz=0
L point
The value of k varies from 0 to some finite values, determined by the lattice spacing. In
the fcc lattice (for semiconductors), the k-vector values are thus confined to a volume
called the Brillouin zone. This Brillouin zone for fcc lattice is shown in Fig. 6.

The Brillouin zone looks like a truncated octahedron with the cube in the fcc
structure.
From the center, join eight corners and draw perpendicular bisectors to get the
octahedron. 56

28
 8/20/2024

METALS, SEMICONDUCTORS AND INSULATORS

When an allowed band is completely filled with electrons, the electrons in the band cannot
conduct any current. This important concept is central to the special properties of
semiconductors.

When we have a material in which a band is completely filled, while the next allowed band is
separated in energy and empty, the material has, in principle, infinite resistivity and is called
an insulator or a semiconductor. The material in which a band
is only half full with electrons has a very low resistivity and is called a metal.

The band that is normally filled with electrons at 0 K in semiconductors is called the valence
band while the upper unfilled band is called the conduction band.

We have already discussed that the solution of the Schrodinger equation leads to the
band structure of the semiconductor. The top of the valence band of most
semiconductors occurs at k=0, i.e., at effective momentum equal to zero. A typical
bandstructure of a semiconductor near the top of the valence band is shown in Fig. 9

57

Fig. 9: Schematic of the valence band, direct bandgap and indirect bandgap conduction
bands. The conduction band of the direct gap semiconductor is shown in the solid line
while the conduction band of the indirect gap semiconductor is shown in the dashed
line. The valence band states are called holes.

58

29
 8/20/2024

Ye mp kisi kaam ka nahi h

Near the conduction bandedges (occurring at k = k0 ) it is usually possible to represent
the bandstructure by a simple relation of the form

where the index i represents the x,y,z components of k or k0. For direct bandgap materials k0 =
(0,0,0).
If the bandstructure is isotropic as is the case for direct gap semiconductors, the relation
becomes
Ye direct bandgap ke liye lagado

where Ec is the conduction bandedge, and the bandstructure is a simple parabola with
equal energy surfaces being the surfaces of a sphere.
For indirect materials like Si, the bottom of the conduction band occurs at six equivalent
points: 2(0.85,0,0)/a, 2(0,0.85,0)/a, 2(0,0,0.85)/a and their inverses. The energy
momentum relation has the form

YE indirect band gap k liye lagado

where kl is the longitudinal part of k (i.e., parallel to the k-value at the bandedge) and kt is the
transverse part measured from the conduction bandedge k-value. The constant energy
surface of the bandstructure is an ellipsoid. 59

where, for most semiconductors

where mo is the free electron mass.

The conduction band effective mass thus decreases rapidly with decreasing bandgap.

Near the top of the valence band, in Fig. 9, there are two important curves. The heavier mass
band is called the heavy hole band (I in Fig. 9). The second lighter band is called light hole band
(II in Fig. 9).
The masses of the valence band electrons are usually much heavier than those in the
conduction band and are also negative. The reason we call the valence band states "holes"
will be discussed next.
In general, the effective mass is defined by the relation

60

30
 8/20/2024

Concept of hole

At 0K, valence band is full , conduction band is empty.

When all valence band states are filled, Σ ki=0= Σ ki+ ke where ki ≠ ke

Missing states are called holes and wave vector of the system –ke
is attributed to it.

If electron moves opposite to the field direction, unoccupied states or holes

move in the field direction as if it has a mass mh and positive charge

Fig. 10: Diagram illustrating

the wavevector of the missing
electron ke. The wavevector
of the system with the missing
electron is — ke, which is
associated with the hole.
61

BAND STRUCTURES OF SOME IMPORTANT SEMICONDUCTORS

Silicon
Silicon is the unchallenged material of choice for electronic products. The bandstructure of Si
is shown in Fig. 11a. The bandgap is 1.1 eV with the bottom of the conduction
band occurring at k = (0.85*2/a, 0, 0) and the five other equivalent points, where a is the
lattice constant (5.43 A). The bandstructure near the conduction band minima is {k is
measured from the bandedge value)

where ml* = 0.98 mo and mt* = 0.19 mo, and gives ellipsoid constant energy surfaces (Fig.
11b.

Fig. 11 (a) Fig. 11(b) 62

31
 8/20/2024

Being an indirect material, Si has very poor optical properties and cannot be used to
make lasers. The reason for this is the need for "vertical k" transitions for optical processes
due to momentum conservations. The hole transport properties of Si are alsc quite poor,
since the hole masses are quite large. However, for electronic devices, silicon is the material
of choice because of its highly reliable processing technology.

GaAs
The bandstructure of GaAs is shown in Fig.
12. The bandgap is direct which is the chief
attraction of GaAs. The direct bandgap
ensures excellent optical properties of
GaAs as well as superior electron transport
in the conduction band. The bandedge E vs k
relation is quite isotropic leading to spherical
equal energy surfaces. The bandstructure can
be represented by the relation

with m* = 0.067 mo. A better relationship is the

non-parabolic Fig. 12

with  = 0.67 eV -1. The values of the hole

masses are m*hh = 0.45 mo , m*lh = 0.1 mo. 63

Fig. 13

The band
structure of
several other
semiconductor
s: Ge, AlAs,
InAs, InP

It is also important to note that the bandgap of

semiconductors, in general, decreases as the
temperature increases. 64

32
 8/20/2024

Band structure modification :

(i) Alloying of two or more semiconductors
Vegard’s law : (i) aalloy=x aA+ (1-x) aB

Band structure of alloy: Egalloy = xEgA + (1-x) EgB

Due to disorder, there is a bowing effect:
Egallow= a+ bx+ cx2 C is the bowing parameter.
(ii) Hetero structures
Epitaxial technique of crystal growth. Using epitaxial technique to grow sequence of
semiconductor layer with different band gap. An important example is the GaAs and the
AlxGa1-xAs system, which have a very good match of their lattice constants.

Fig. 14

(b) 65

An extremely important parameter in the quantum well problem is the band-edge

discontinuity produced when two semiconductors are brought together. As shown in Fig.
14a, a part of the bandgap discontinuity (EgA — EgB) of two semiconductors A and B would
appear in the conduction band and a part would appear in the valence
band.

Once the band discontinuity is known, the bandstructure of a quantum well structure can
be calculated. The simplest way to do this to use the effective mass theory in which the
electron in each region of the structure is represented by its effective mass.
The effect of the quantum well is to impose a background confining potential.

Due to confinement of electrons, there is sub-band formation (see Fig. 14b).The effective
band-gap changes. These sub-bands have important consequences for optical and
transport properties of heterostructures.

(iii) Band structure modification by Strain:

Fig. 15: The consequence of
pseudomorphic strain on the
bandedges of a direct bandgap
semiconductor. The valence
band degeneracy at k = 0 is
lifted as the HH and LH states
are split as shown.
66

33
 8/20/2024

INTRINSIC CARRIER CONCENTRATION

The intrinsic carrier concentration depends upon the bandgap and temperature and the
details of the band edge masses. We will assume that the band edge density of states for
electrons and holes originate from parabolic E-k relationships. The conduction and valence
band density of states are shown in Fig. 16, along with the position of a Fermi level.

Fig.16: A schematic of the density of states of the conduction and valence band. N e and
Nh are the election and hole density of states. Also shown is the Fermi function giving the
occupation probability for the electrons.

The concentration of electrons

in the conduction band is

67

𝟏 𝟏Τ𝟑
𝒎∗𝒅𝒐𝒔 = 𝒎𝟏 𝒎𝟐 𝒎𝟑 𝟑 = 𝒎∗𝒍 𝒎∗𝟐
𝒕

𝟐Τ𝟑 𝟏Τ𝟑
𝒎∗𝒆 = 𝟔 𝒎∗𝒍 𝒎∗𝟐
𝒕

Nc is known as the effective

density of states at the
conduction band edge.

The carrier concentration is known when EF is calculated. To find the intrinsic

carrier concentration, this requires finding the hole concentration p as well.

fh=1-fe
where Nv is the effective density of states for the valence bandedge.

34
 8/20/2024

In intrinsic semiconductors, the electron concentration is equal to

the hole concentration since each electron in the conduction band
leaves a hole in the valence band. If we multiply the electron and
hole concentrations, we get

69

and since for the intrinsic case n = ni=p=pi, we have from the square
root of the equation above

69

If we set n = p, we also obtain the Fermi level position measured from the valence band
edge

Thus the Fermi level of an intrinsic material lies close to the midgap.
• Note that the carrier concentration increases exponentially as the bandgap
decreases and has a strong temperature dependence.
• We note that the product np calculated above is independent of the Fermi level EF.
This is an expression of the law of mass action. This result is valid not only for the
intrinsic case but also when we have dopants.
• At room temperature, the np product is 2.25 x 1020 cm-6, 5.76 x 1026 cm-6 and 3.24 x
1012 cm-6 for Si, Ge, and GaAs, respectively. The corresponding intrinsic carrier
densities are 1.5 x 1010 cm-3, 2.4 x 1013 cm-3 and 1.8 x 106 cm-3, respectively.
Temperature dependence: Fig. 17
• It is quite clear that pure semiconductors have a very low concentration of carriers that can
conduct current. One must compare the room temperature concentrations of ~ 1011 cm-3 to the
carrier concentrations of ~ 1021 cm-3 in metals. Indeed, pure semiconductors would have little
use by themselves. 70

35
 8/20/2024

• Doping of impurities is essential to increase the carrier concentration

in semiconductors. The dopants are chosen from the periodic table so
that they either have an extra electron in their outer shell compared to the
host semiconductor, or have one less electron. The resulting dopant is
called a donor or acceptor.

Fig. 17: Intrinsic carrier

concentrations of Si and GaAs
as a function of reciprocal
temperature.

71

Fig. 18: Schematic for

understanding donors
in semiconductors

Bohr radius
in Si ~ 100 Å

The donor atom is treated as a shallow defect and the electron- donor interaction is
represented by a Coulombic potential. The problem is solved by a simple analogy to the
hydrogen atom problem. The donor produces a shallow level whose energy is simply
given by the hydrogen atom problem in quantum mechanics, except that the mass of the
electron is replaced by the effective mass at the conduction bandedge:
4
0.045 eV for P in Si
0.057 eV for Al in Si
0.006 eV for Si in GaAs
donor energy levels depend only upon the host crystal (through  and m*) and not on the
nature of the dopant.
Conductivity effective mass (m*): response of electrons to external potentials. This mass
is used for donor energies as well as for charge transport in an electric field. For direct
bandgap materials like GaAs, this is simply the effective mass. For materials like Si the
conductivity mass is
72

36
 8/20/2024

Donor level

Acceptor level

Fig.19: Schematic band diagram, density of states, Fermi-Dirac distribution, and carrier
concentrations for (a) intrinsic, (b) n-type, and (c) p-type semiconductors at thermal
equilibrium.
Note that pn = ni2 for all three cases.
73

Carriers in Doped Semiconductors

• Because of the doping, n – p  0 = n ,
but np = constant= ni2

Solving for n (eliminate p) , we get ∆𝑛 1

𝑛= + (∆𝑛2 + 4𝑛𝑖2 )1/2
2 2

Using Boltzman approximation for Fermi function f(E),

𝑛
= 𝑒 (𝐸𝐹 −𝐸𝐹𝑖 )/𝑘𝑇
𝑛𝑖
𝑝
Similarly, = 𝑒 −(𝐸𝐹 −𝐸𝐹𝑖 )/𝑘𝑇
𝑛𝑖

When the semiconductor is doped n-type (p-type), the Fermi level moves closer to
Conduction (valence) band edge.
74

37
 8/20/2024

Carriers in Doped Semiconductors

• In this case, Boltzman approximation is not valid for

f(E).
• Then we can use the Joyace-Dixon approximation for
the integration in n.
𝒏 𝟏 𝒏
𝑬𝑭 = 𝑬𝒄 + 𝒌𝑻[𝒍𝒏 + ]
𝑵𝒄 √𝟖 𝑵𝒄

75

Doped Semiconductor

n = ni + nd

Fig. 21: Electron density as a function of temperature for a Si sample with

donor impurity concentration of 1015 cm-3.
76

38
 8/20/2024

Degenerate Semiconductors: High doping density, probability of band-

edge electronic levels are occupied by electrons/ holes is close to 1.

Modulation Doping
Impurity electron scattering. Controllable free carrier density Without
scattering.
Heterostructure with high band gap region doped.
(GaAs/AlGaAs). Fermi level is measured from bottom of subband E1.
DOS in QW

77

Fig. 22: Modulation doping approach 78

39
 8/20/2024

Doping Techniques
1. Epitaxial techniques : Where dopant is introduced during growth.
2. Ion-implantation: High-energy dopant ions are implanted into semiconductors (after growth).
3. Diffusion: When a lump of the dopant material is placed on the semiconductor & is annealed,
causing a diffusion of the dopant atoms into the semiconductor.

Lattice vibration : Phonon, resulting force due to binding force

Phonon-dispersion - Acoustic Phonon

- Optical Phonon Diatomic Bases
Measurements: Neutron scattering.

Transport processes: No scattering of Bloch electrons in ideal periodic potential.

Real semiconductor crystals: many sources of scattering:
Sources of Scattering Imperfections
Ionized Impurity Due to dopant in semiconductors
Phonon Due to lattice vibrations
Alloy Random potential function
Interface roughness Impurity in heterostructure devices
Chemical Impurity Unintentional impurities with short range potential
Photons Electrons

Relaxation time (r)- Average time it takes to loose coherence or memory of initial state properties.
79

Velocity-Electric Field Relations in Semiconductors

When an electron distribution is subjected to an electric field, the electrons tend to move in
the field direction (opposite to the field F) and gain velocity from the field. However, because
of imperfections, they scatter in random directions. A steady state is established in which the
electrons have some net drift velocity in the field direction. The response of the electrons to
the field can be represented by a velocity-field relation.

Low Field Response: Mobility

Three assumptions:
i) The electrons in the semiconductor do not interact with each other. This approximation is
called the independent electron approximation.
ii) Electrons suffer collisions from various scattering sources and the time sc describes the
mean time between successive collisions.
iii) The electrons move according to the free electron equation

This average velocity gain is then that of an electron with mass m* traveling in a field F
for a time sc :

where Vd is the drift velocity. The current density is now

Comparing this with the Ohm's law result for conductivity 

we get
80

40
 8/20/2024

The resistivity of the semiconductor is simply the inverse of the conductivity. From the
definition of mobility , for electrons
we have

Notice that the mobility has an explicit 1/m* dependence in it. Additionally sc decreases with
m* also as we had discussed in the previous section due to the dependence of scattering
rates on the density of states. Thus, the mobility has a strong dependence on the carrier mass.
The effective mass to be used in these transport parameters is the conductivity mass.

 for Si ~1500 cm2/V-s

 GaAs~ 8500 ‘’
AlGaAs ~11000 ‘’

81
Fig. 23: Electron Mobility as a function of temperature for low doping and high doping cases.

For ionized impurity scattering: I  T 3/ 2

Acoustic phonon scattering:  AP  T −3 / 2
Alloy scattering:  A  T −1/ 2 Effect of modulation doping.
‘’

vd
Vd Saturates at
~10000 V/cm

F (V/cm)

(c)

Fig. 24:: Mobility as a function

of impurity concentration,

82

41
 8/20/2024

The response of the free carriers to an external electric field is given v & F relationship.

(i) For low field, v=μ F , μ = mobility = velocity/ unit field

(ii) Higher field (F ≥ 1 KV/cm), v-F relationship non-linear negative resistance in direct gap materials.
(iii) Externally high field , ( F > 105 V/cm). Breakdown of semiconductor due to impact ionization or
tunneling. Avalanche detectors exploit this phenomena.

For ionized impurity scattering

mobility increases with temperature.

For acoustic phonon scattering

Mobility decreases with temperature

83

High field transport-Breakdown phenomena

For F ≥ 100 KV/cm, breakdown due to carrier multiplication.
(A) Impact Ionization (Avalanche Breakdown)
(i) For low field , electron remains in the same band.
(ii) High field , hot electron scatters with VB electron via
Coulombic interaction.
(iii) Initial electron must have higher energy than
Bandgap.
(iv) Carrier multiplication- avalanching.

Energy conservation:

Because of simultaneous condition, minimum value of Ek1:

Current depends on bandgap. 𝑑𝐼

= 𝛼𝑡 𝐼 or
In narrow gap material, it happens for low
𝑑𝑡
Rate of ionization per unit time.
electric field. 𝛼𝑡 strongly depends on bandgap.
No. of electrons suffering impact ionization after traveling a distance x,

84

42
 8/20/2024

Ionization rate of
electrons/holes vs
1/F.
Fig. 25

Fig. 26: The impact ionization process where a high energy electron scatters from a valence band electron
producing two conduction band electrons and a hole. Hot holes can undergo a similar process. Both energy
and momentum conservation are required in such a process. Thus the initiating electron must have an
energy above a minimum threshold. 85

Critical field is defined as αimp= 104 V/cm. One impact ionization for travel of 1 m.

Disadvantage : Limitation of power output of a device. (control of device due to Breakdown).

Advantage: Avalanche photodetector - High gain detection.
Also applied in Microwave devices.

Band-to-Band Tunneling: Zener Tunneling

From quantum mechanics, Tunneling probability through
rectangular barrier:
m*- e-h reduced mass.
In narrow gap materials, Zener tunneling is very important.
It is the basis of Zener diode, where current is 0 until b-b tunneling starts & current
increases sharply. T ≥ 10-6 necessary for breakdown to start.

Fig.27: E-x and E-k diagram of p-n junction 86

43
 8/20/2024

Carrier transport by Diffusion:

Fig.28: Concentration profile of electrons as a function of space

Diffusion due to thermodynamic reasons. From high concentration

region to low concentration region- due to concentration gradient.
Disorder/entropy decreases.
Flux to the right through x0

diffusion coefficient Dn depends on temperature 87

Transport by drift and diffusion (combined): Einstein’s relation

μF is const.= vd
Force on electron in Field F= - e F
As Force = - ∇u(x),
Pot. Energy U=e F.x for the uniform electric field.
As electron charge is negative
Ec(x)=Ec(F=0)+ eF. X
At equilibrium, Jn =0 =Jp
88

44
 8/20/2024

at equilibrium
𝒅𝒏 𝒏 𝒙 𝒅𝑬𝑭𝒊 =0
=
𝒅𝒙 𝒌𝑻 𝒅𝒙

at equilibrium

Einstein Relation
Charge injection and Quasi-Fermi Level:
For non-equilibrium electron and hole concentrations, the occupation of the carriers is given
by an electron and hole Fermi function independent of each other.
• Note that in equilibrium, the same Fermi function describes both electrons and holes.

with Boltzman approx.

EFn and EFp are quasi-Fermi levels for e, h
With Joyace-Dixon approx.

89

Hall effect
Force in hole F= q( E + v x B)
Fy= q(Ey – vxBz)
Ey = vxBz
ve=eE=e F/e, Fy=0

Fig. 29: Hall effect set up

90

45
 8/20/2024

If n and p are comparable, then the following treatment applies:

when n and p are

Hall coefficient: comparable

91

OPTICAL PROPERTIES OF SEMICONDUCTORS

Scattering caused by light or photon
Electro-magnetic wave eqn.

Light as electromagnetic wave travelling through semiconductor

Electric field vector

α → absorption coefficient
z → propagation direction
For nonzero α , photon flux

Absorption → By impurities , multiband absorption (C-B Absorption:

electron) Emission:
Most important in semiconductors is Band to band transition.
Momentum conservation ensures vertical transition. Photon:

92

46
 8/20/2024

Scattering involves the following issues :

Conservation of energy : Absorption
Emission
Conservation of momentum :

So, k conservation ensures that initial and final electron should have same k-
value.i.e., only vertical k-transitions are allowed.

Fig. 32: Band to band absorption and emission in semiconductors. In (a) an electron in the
valence band absorbs a photon and moves into the conduction band; (b) In the reverse process
an electron in the conduction band emits a photon and moves "vertically" down into the valence
band. 93

Direct bandgap optical transitions are strong.

Indirect bandgap optical transitions are weak.
Indirect gap optical transitions require the help of lattice vibrations to satisfy k-conservation.
The k-values for the electron & holes are same in vertical transition.

mr* is the reduced mass.

So, relevant density of states here in the scattering process is the reduced mass.
This DOS is called Joint Density of States. (J-DOS). Dipole matrix
For light polarized along arbitrary direction , absorption coefficient is element
between VB
and CB

for most semiconductors.

94

47
 8/20/2024

95

2.25x

5.6x
96

48
 8/20/2024

1/2
For GaAs,

So, a few μm of GaAs is enough to absorb a significant fraction of height above bandgap.
 (Si) is low near bandedge. k-conservation rule makes transition weak.
Spontaneous emission (no photon beforehand)
Emission rate

Stimulated emission rate

Radiative recombination rate of an electron having momentum ħk with a hole of moment

ħk or ħ

In indirect gap material τ0 is large (~μs)

In direct gap material τ0 is small. (~ns)

97

For a given gap of energy 𝐸 to be

considered "indirect", the momentum
offset 𝑝 is larger than the momentum
of a photon of energy 𝐸. In other
words, 𝑝≫𝐸/𝑐

Fig. 33: Absorption coefficient for several direct and indirect semiconductors. For the direct
gap material, the absorption coefficient is higher.

Excess carrier Recombination Mechanisms:

(i) Direct → band to band
(ii) Indirect → via deep energy level
(iii) Auger recombination (3 particle process) : boun to bound (trap mediated) 98

49