dr. p. sagayaraj state physics unit … · space of the boundaries of the zones; these positions...
TRANSCRIPT
Solid State Physics Module-II
THEORY OF METALS AND SEMICONDUCTORS
E- Content study material compiled by
Dr. P. Sagayaraj
Associate Professor of Physics, Loyola
College (Autonomous), Chennai - 34
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.1
⁄
BRILLOUIN ZONES
In the propagation of any type of wave motion through a crystal lattice, the
frequency is a periodic function of wave vector k. In order to simplify the treatment of
wave motion in a crystal, a zone in k -space is defined which forms the fundamental
periodic region, such that the frequency or energy for a k outside this region may be
determined from one of those in it. This region is known as the Brillouin zone (sometimes
called the first or the central Brillouin zone). The concept of Brillouin zone provides a
way to understand the origin of allowed and forbidden bands in solids.
The energy discontinuities in the monatomic one-dimensional lattice occur
when the wave number is
= (1)
where n is any positive or negative integer.
The equation determining the position of the energy discontinuities which actually
occur will depend somewhat on the type of the crystal lattice.
Fig. 1Brillouin zones of a linear monatomic lattice with lattice constant a
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.2
ting
/
the
as
/
sho
<
w Fig
/
.
nes
/
; the
<
s t −
/ an /
<
< the
/
first
for
ne
+
d by t he
=
soluti ons
(
of t
+
e equ
)/
ation
tin
= ±
± 1
⁄
, n2
ing
= ± ⁄
In the one-dimensional monatomic lattice a line represen value of is
divided up by the energy discontinuities into segments of length n in
1. The line segments are known as Brillouin zo egmen
is Brillouin zone; the two segments −2 < − d <
2 m the second Brillouin zone, etc. The zone description was introduced by
Brillouin, who pointed out that many important and characteristic features of electron
propagation in periodic structures could be described by considering the positions in k-
space of the boundaries of the zones; these positions are independent of the details of
the electron lattice interaction, being determined instead by the crystal struct ure. The
Brillouin zones of a simple square lattice in two dimensions are shown in Fig. 2 The
zone boundaries are determi
(2)
Here n1, n2 are integers and a is the lattice constant. The equation essentially expresses
the Bragg law for reflection of a wave by a periodic lattice. We find the boundaries of the
first zone by first set g n1 = = 0, obtaining
(3a)
and then setting n1 = 0, n2 = ±1, obtain
(3b)
The four lines of the above two equations determine the boundary of the first zone.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.3
in
± = 2 /
es
Fig. 2 The first three Brillouin zones for a two-dimensional square lattice
The outer boundary of the second zone is determined by setting n1 = ± 1, n2 = ±1,
obtaining the equations of the four l
± (4)
where the signs are independent.
ELECTRONS IN A PERIODIC POTENTIAL
An electron passing through a crystal structure experiences a periodic variation in
potential energy, caused in a metal by the positive cores of the metal ions. In sodium, for
example, the ion cores are singly charged, with 10 electrons in the configuration ls22s22p6,
while the outer electron, which in the free atom is the 3s valence electron, becomes in the
metal a conduction electron. The periodic nature of the potential has far-reaching
consequences for the behavior of the conduction
electrons:
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.4
.
or t e ri dic p
.
Nature of the wave functions: The plane wave solutions for the wave functions of
the free electron model go over f h pe o otential to solutions of the form
= ( )
Where, uk(r) has the periodicity of the lattice. Wave functions of this form are called
Bloch functions and are basic to the theory of metals.
Allowed and forbidden bands: On the free electron model all values of the energy were
allowed; but in a periodic potential there are forbidden ranges of energy (Figs. 1 and 2)
where solutions representing an electron moving through the crystal do not exist.
Fig. 1 Dependence of energy levels upon lattice constant, for a line of 6 hydrogen atoms,
showing the incipient formation of allowed and forbidden energy bands
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.5
lly a qu
= (
adr
⁄2
tic fu
)
nction
fo ion
⁄ = ⁄ ∗. I
Fig.2 Allowed and forbidden energy bands. The energy is plotted as a function of
the wave number k; the dashed line shows the connection with the case of
entirely free electrons
Fig. 2 shows the allowed and forbidden vibrational frequencies of a diatomic crystal in
one dimension. We know also that x-rays do not propagate through crystals at certain
frequencies and orientations, but instead are reflected. In fact, the Bragg equation
determines the occurrence of the forbidden electronic bands, as waves satisfying the
Bragg condition are so strongly reflected that they cannot propagate in the crystal.
EFFECTIVE MASS
Near the top or bottom of a band the energy is genera a of
the wave numbers, so that by analogy with the express r free
electrons we may define an effective mass m* such that t may be
shown that the motion of a wave packet in applied electric or magnetic fields is
characterized by using m* as the mass. Near the top of a band m* is negative, so that here
the motion corresponds to that of a positive charge.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.6
ecti
⁄
ve m
. Due
The effective mass concept does not imply that Newton's second law =
fails for metals. The effective mass applies to the acceleration of electrons relative to the
crystal lattice. Newton's law holds when applied to the entire system of electrons plus
lattice, thereby allowing for momentum transfer between the accelerated electron and the
lattice.
The variation of effective mass with k is shown in Fig.1
Fig.1 The variation of energy, velocity and effective mass with k
The concept of the negative eff ass may be understood in terms of the Bragg’s
reflection when k is close to ± to the Bragg’s reflection, a force applied in
one direction lead to a gain of momentum in the opposite direction which results in the
negative effective mass.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.7
+ ( − ħ
)
sol tions
(
of t
)
he
.
+ 2 + ħ
( − − ) = 0
The effective mass is a new concept and arises because of the interaction of the electron
wave packet with the periodic lattice. It also has a physical significance. It provides a
satisfactory description of the charge carriers in crystals.
KRONIG-PENNEY MODEL
We demonstrate some of the characteristic features of electron propagation in
crystals by considering the periodic square-well structure in one dimension (Fig. 1).
Fig. 1 Kronig and Penney one-dimensional periodic potential
The wave equation of the problem is
(1)
The running wave solutions will be of the form of a plane wave modulated with the
periodicity of the lattice. We obtain u form
(2)
where uk(x) is a periodic function in x with the period (a + b) and is determined by
substituting (2) into (1):
where, Ek = ħ2k2 /2m.
(3)
= 0
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.8
uation
=
has t
(
he
) +
tion
( )
= ⁄ ) ⁄ .
< < + the
= ( ) + ( )
= ( − )⁄ ) ⁄
inear ho
+
mog
=
eneou
+ ;
s eq
( ) − ( ) = ( − ) − ( + ) ; −
( )
+
( ) ( ) ( )
( ) ( ) (
+
) ( ) −
+
) ( )
;
+ ) ( )
si n
ly if the d
sin
eterm
+ co
nt of the c
cos
oeffic
= c
nts v
(
ani
+
shes
)
ina ie
sh os
In the region 0 < x < a the eq
provided that
(4)
(5)
In the region
Provided that
(6)
(7)
The constants A, B, C, D are to be chosen so that u and du/dx are continuous at x = 0 and
x = a, and by the periodicity required of u(x) the values at x = a must equal those at x =
- b. Thus we have the four l uations:
=
− + = ( − (
(8)
These have a solution on , or
(9)
solu
(2
solution is
(2
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.10
→ →
=
mes
+ c o = c
+ c o
ig. 2
, for = ⁄2
2
3
ħ f
.Th
or
rbidden
(
r
=
= /
lim
os
In order to obtain a handier equation we represent the potential by a periodic delta
function, passing to the limit where b = 0 and Vo = ∞ in such a way that β2b stays finite.
We set
(10)
So that the condition (9) beco
(11)
This transcendental equation must have a solution for a in order that wave functions of
the form (1) should exist.
F
Fig. 2 Plot of the function s e allowed values of the
energy W are given by those ranges of α = which the function lies
between +1 and -1.
If P is small, the fo anges disappear. If P —» ∞: the allowed ranges of
αa reduce to the points ±1, ±2, ±3 … . ). The energy spectrum becomes
discrete, and the eigen values
8
are those of an electron in a box of length a.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.11
uation c
8 +
be writt
[
en as
− ( ) ]
( ) m
( ) = ( + )
( ) = ( ) ( )
( ( . )
respe
(
ctiv
) an
y. h ent th
(
e fr
) or
ee le
) = ( )
BLOCH THEOREM
In order to understand the difference between the conductors and insulators, it is
necessary to incorporate the variation of potential inside the crystal due to the presence
of positive ion cores in the free electron model. The potential is minimum at the positive
ion sites and maximum between the two ions. The corresponding one dimensional
Schrödinger eq an
= 0 (1)
Where the periodic potential ay be defined by means of the lattice constant “a”
as
(2)
Employing the periodic potential the one dimensional solution of the Schrödinger
takes the form
(3)
In three dimensions it is given by
(4)
The eqns (3) and (4) are known as Bloch function in one and three dimensions,
el T ey repres e ctron wave modulated by the periodic function d
( ), where ( ) is periodic with the periodicity of the lattice
in one and three dimensions, respectively. Therefore, considering only the one
dimensional case and suppose if we have N (even) number of atoms in a linear chain of
atoms of length L, then we can write,
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.12
( ) = ( + )
an
(
depe
+
nds on th
) =
e e
(
xac
+
t nature o
) [
f the pot
( +
e ntial fie
)]
( ) (x)exp(ik ( ) ( ) (6)
as
∗ + ) = (− ) ∗ ( )
ives
(
us
+ ) ∗ ( + ) = ( ) ∗ )
(6) gives us
( ) = 1
= 2 × = 2
2 =
2 = (9)
, we here
This is frequently referred to as Bloch condition. Similarly, the complex conjugate of
eqn (6) can be written
( (7)
The eqs. (6) and (7) g
( (8)
This indicates that the probability of finding the electron is same everywhere in the
whole chain of atoms, i.e, it is not localized around any particular atom but is shared
by all atoms in the chain. Thus eqn.
This will be true only if
W n = 0,±1, ±2, … and L is the length of chain of atoms. When n =
have k = , which is the edge of the first Brillouin zone. When L is large, the allowed
values of k would come close together and their distribution along k-axis becomes
quasi- continuous. The total number of allowed k values in the first zone is
(5)
ld . From eqn (3) and (5) we have
x) =
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.13
2 ⁄ = =
2 ⁄ =
nction
( )
is zero. i.e
= 0
u
∫
aves,
) = exp( )
( ) = 8
(10)
This is equal to the total number of atoms in the chain of atoms (or in the unit cell in
three dimensions).
THE NEARLY FREE ELECTRON MODEL
In this model, the crystal potential is assumed to be very weak as compared to the
electron kinetic energy so that the electrons behave essentially like free particles. The
weak periodic potential introduces only a small amount of perturbing effect on the free
electrons in the solid. Therefore this model demands the application of very elementary
perturbation theory. It is convenient to choose the zero of energy so that the mean free
value of the potential f .,
(1)
where a is the periodicity of the lattice. The unperturbed wave functions
corresponding to V= 0 are the plane w
∅ ( (2)
where the wave functions are normalized over a microcrystal containing N atoms. The
unperturbed electron energies are
(3)
This is free electron case where k can take any value. Now suppose we take in to
account periodicity of the lattice keeping V = 0 and restricting k to lie in the first
Brillouin zone only. Then we have,
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.14
( ) = exp( ′ )
′ = + = + ,
( ′) = 2
+
≤ = 0 ± 1
iodi
(
c pote
) =
l, the
) +
ave fun
′ (
cti
)∅
′ ( ) ar
are giv
′ ( )
en
=
′| |
) −
>
( ′
<
( )
′| | > ∗ ∅
d to sec
( )
ond
=
orde
( )
r
+
s
′| |
) −
> |
(
|<
( ′)
he lattic
( ) =
rms of
ex
he Fourier series,
) and ∗ =
it ca
′| | ≥
p(
if − ′ =
(4)
Where and
(5)
Where , ±2, … … …
As we introduce the per ntia real w on becomes
∅ ( ∑ ′ ′ (6)
where e constants and are found to be small from perturbation theory. Their
values corrected to the first order by
(7)
Where < = ∫ ∅ (8)
The perturbed energy, correcte i
(9) ′
Because of eqn (2), the first order term becomes zero.
Now let us take in to account the real periodic potential which may have the same
periodicity as that of t e. In te n be expressed as,
∑
Then the integral (8) becomes <
= 0 ; otherwise (10)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.15
ly, t
(
e wav
) =
nction becomes
) ∑
√ ( ) ex ) exp
= = ( )
( ) de ghtly de
( )
cr
−
eased o
( )
r
,
incre
which he diffe
( )
re
−
nce
( ) =
8 [ −
i
+
s giv
) ]
−
(
2 + =
4 + −
2
2 +
y
= / = ′ = /2
is theo
an
neve
).
r the
Near
sary
(
to write
) = exp )+ exp ( + )
an are
4)
ts. Th
=
eqn
. ear the
( )
b
−
oundar
( )
y
+ ( )
±
]
Su
ex
stituting eqn (14
) + [ ( +
) in eq
) −
n (1) w
( )
e
( ) ] exp ( + ) =
Consequent h e fu
( p(− (11)
With, 1 and (12)
The energy is sli ased slightly with respect to pending
on t en by
=
8
= − (13)
At or near the zone boundar
− − −
Where an integer. The perturbation ry there for break down whe wave
vector is an integral multiple of ± d the wave is reflected (k= − a
zone boundary, it is neces
( (1
Where d constan e (14) means that one Fourier coefficient
will be large n b obtain
[ p(
+ 0 (15)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.16
) or by exp ( + ) ] an
ultaneo
( )
us
−
eqns.
( )] + = [ ∗
+[ ( + ) − ( )]
0
= 0
ic eqn h
( ) =
two so
(
lu
+
tion
)] ( ) − ( + )} + | ]
=
[{
an ( ) = ( +
4|
). T
= + | |
= − | |
= − =
If we multiply eqn (15) either with exp( [− d integrate,
we obtain the following sim
(16)
From a non trial solution, the determinant of coefficient must vanish. The resulting
quadrat as s
2 [ ± (17)
At the zone boundary, ± d herefore the two energies
are
Thus the energy gap at the first zone boundary (n=1) is ∆ 2
Fig. 1 Energy band in the NFE model (a) the extended zone scheme, (b) the
reduced zone scheme
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.17
d the s
⁄
econ
| =
ndarie
=
n
( )- ( + ) ≫ , ) ≅ ( ) ( + ), i
The Fig (1) shows the energy gap at the first an d zone bou s i
extended and reduced zone schemes. Also in the case,| 1, so that at ±
the wave function is a superposition of two plane waves travelling opposite directions
and thus represents a standing wave.
Near k = 0, the difference in the unperturbed energies is so large that the
perturbation effect is negligible and hence there is no effect on the curve in this region.
When then ( .e., away
from Bragg reflection, the energy values are practically plane waves, the same as that of
free electrons.
ZONE SCHEMES
The wave vector representation of a plane wave Eigen state is simple and
unambiguous while the representation of the Bloch states are not so because the Bloch
function is not a simple plane wave but a modulated plane wave. To represent such sates,
three different schemes are commonly used. They are called the extended zone scheme
(or the Brillouin zone scheme), the reduced zone scheme and the periodic zone scheme
(or the repeated zone scheme). All the three schemes represent the identical physical
behavior.
Extended Zone Scheme
Consider a one dimensional lattice in which the energy of an electron is being
slowly increased so that the value of k is also increased. When the value of k becomes
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.18
hat th
=
e Bragg
=
large enough, the wavelength becomes small enough as k = 2π/λ, the electron will
suffer a Bragg reflection following the Bragg condition 2d sin θ = nλ.
For a one dimensional lattice, d= a so t reflection will occur at
where ±1, ±2, …..
As a result of these reflections energy gaps are developed in the free electron
parabola as shown in the Fig.1. This representation of energy as a function of k is known
as extended zone scheme. Physically this scheme is very close to the free electron scheme
and differs only at the zone boundaries. However, like free electron case, the k-space is
infinite in the extended zone scheme but is dissected by the planes of energy discontinuity
in to segments called Brillouin zones. The extended zone scheme represents various
Brillouin zones in k-space.
Fig.1 Extended zone scheme
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.19
y fr
(
ee ele
) =
tron wave is g
( )
iven by
ssion i
( )
s
=
8
te eq
(
n (1)
) =
n the form,
( ) = ( + ) exp( )
′ ) ex
−
) = exp( p(−
Reduced Zone Scheme
Fig.2 Reduced zone scheme
For a free electron case, the ordinar c
(1)
And the corresponding energy expre
(2)
A plot between E versus k gives the free electron parabola (Fig.3).
Let us now wri i
(3)
Where ′= k+g is the reduced value of the original wave vector k. Since, the second term
on the right hand side of eqn (3) is periodic function in the lattice. It is the limiting case
of a vanishingly small periodic potential under this assumption one must
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.20
parab
=
la
⁄
in k-
(Fi e energ
( ′) =
y express
( +
ion i
) =
he uced zone
2 +
sche me
= 0
( ) i s
place
=
the part
/ as
t red
take in to account the symmetry requirement of the periodicity. The general demand
of the periodicity implies that the possible electron states are not restricted to a simple
o space, but can be found equally well on parabolas shifted by a vector
g. 3b). Th n becomes
(4)
Where , ±1, ±2, … ..
Since, the behavior of periodic in k-space and hence follows the
symmetry properties, it is sufficient to represent this in the first zone only. In order to
achieve this, let us dis of the parabola of interest linearly by the
appropriate multiple of 2 shown by arrows in Fig. 3b. The E versus k
curves for several values of n reduced into the first zone for a simple cubic lattice with
vanishing potential are shown in Fig.3c. Each value of n defines a Brillouin zone in the
reduced zone scheme representation.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.21
versus plo
porta
val
elec
app
a crystal
/ .
be
Fig. 3 (a) Energy as a function of wave vector for a free electron (b) Dispersion
curve with the translational symmetry and the various bands (c) The same
dispersion curve in the first zone only
Periodic Zone Scheme
In the preceding section, we obtained the t in reduced zone
scheme by translating the desired portion of the neighbouring parabolas into the first zone
as the energies obey translational symmetry. A reverse process is equally probable.
Accordingly, we can translate the desired portion of the first zone to any or every other
zone, we can obtain a periodically repeated zone as shown in Fig. 4. This construction is
known as periodic or repeated zone scheme. This representation leads us to a very im
nt result that the trons in have like free electrons
for most of the ues except when roaches
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.22
ril are
;
one given b
2 4 ± ; …
y:
… ± = 0;
/
±
= / , f o We cut th
/
e series
= − /
of
is n nted
/
as an
. Th
nnecte
, th
to
With
in
Fig. 4 Periodic Zone Scheme
Band model of metal, semiconductor and insulator
Consider a linear crystal constructed of an even number N of primitive cells of
lattice constant a. In order to count states we apply periodic boundary conditions to the
wave functions over the length of the crystal. The allowed values of the electron wave
vector k in the first B louin z
f at r this is the zone boundary. The point
− ot to be cou independent point because it is co d
by a reciprocal lattice vector with e total number of points is exactly e number
of primitive cells.
Each primitive cell contributes exactly one independent value of each
energy band. This result carries over into three dimensions. account taken of the two
independent orientations of the electron spin, there are 2 dependent orbitals in
each energy band. If there is a single atom of valence, one in each primitive cell, the
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.23
band can be half filled with electrons. If each atom contributes two valence electrons to
the band, the band can be exactly filled. If there are two atoms of valence, one in each
primitive cell, the band can also be exactly filled.
Metals and Insulators
If the valence electrons exactly fill one or more bands, leaving others empty, the
crystal will be an insulator. An external electric field will not cause current flow in an
insulator. (We suppose that the electric field is not strong enough to disrupt the electronic
structure.) Provided that a filled band is separated by energy gap from the next higher
band, there is no continuous way to change the total momentum of the electrons if every
accessible state is filled. Nothing changes when the field is applied. A crystal can be an
insulator only if the number of valence electrons in a primitive cell of the crystal is an
even integer.
If a crystal has an even number of valence electrons per primitive cell, it is
necessary to consider whether or not the bands overlap in energy. If the bands overlap in
energy, then instead of one filled band giving an insulator, we can have two partly filled
bands giving a metal (Fig. 1) the alkali metals and the noble metals have one valence
electron per primitive cell, so that they have to be metals. The alkaline earth metals have
two valence electrons per primitive cell; they could be insulators, but the bands overlap
in energy to give metals, but not very good metals. Diamond, silicon, and germanium
each have two atoms of valence four, so that there are eight valence electrons per
primitive cell; the bands do not overlap, and the pure crystals are insulators at absolute
zero.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.24
is co
(th
Fig. 1 Occupied states and band structures giving (a) an insulator, (b) a metal or a semimetal
because of band overlap, and (c) a metal because of electron concentration. In (b) the
overlap need not occur along the same directions in the Brillouin zone. If the overlap is
small, with relatively few states involved, we speak of a semimetal.
INTRINSIC SEMICONDUCTORS
Intrinsic semiconductors are those materials having an energy gap of the order of
1 eV. Hence the conduction process in these materials is achieved as a result of thermal
excitation of electrons across the energy gap. Pure chemical substances such as silicon
and germanium (elements of group IV of the periodic Table), GaAs, InSb (III-V
compounds), SiC (a IV-IV compound) and PbS (a IV-VI compound) are some
examples of intrinsic semiconductors. At absolute zero, the valence band of an intrinsic
semiconductor mpletely filled and the conduction band which is
separated by a distance e band gap) from the valence band is empty. For this
reason, at absolute zero, an intrinsic semiconductor behaves as an insulator and has zero
conductivity. Fig. la shows a simplified schematic diagram of an intrinsic
semiconductor.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.25
electrons
availab
However, as the temperature is gradually increased, the of the
valence band get excited. For some electrons, the thermal energy le at the
room temperature is sufficient to surmount the forbidden energy gap and move into the
conduction band (Fig. 1b). The excitation of electrons from the valence band to the
conduction band leaves an equal number of vacancies (called holes) in the valence
band. Both, electrons in the conduction band and holes in the valence band serve as
charge carriers and contribute to the electrical conductivity.
Fig. 1 Intrinsic semiconductor: (a) at absolute zero the valence band is completely
filled by electrons and the conduction band completely empty; (b) at temperature
above absolute zero some of the electrons from the valence band are excited to the
conduction band; holes appear in the valence band and free electrons in the
conduction band
EXTRINSIC SEMICONDUCTORS
A real semiconducting crystal grown in a laboratory always contains some
impurity atoms (in addition to other imperfections), which create their own energy
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.26
is
levels termed as impurity levels. These levels may occupy positions anywhere in the
allowed and forbidden bands of the semiconductors. However, frequently some known
impurities (in terms of both qualities and quantities) are introduced intentionally to
impart specific properties to the host material. The resulting semiconducting material is
known as extrinsic semiconductor, whose properties depend on the type of the impurity
introduced into the host material. For this purpose, normally a few parts per million of
suitable impurity material is added to the melt during the growth of semiconductor
crystal. For silicon (and germanium) two types of impurities are used. These are from
group Ill and V of the periodic Table, i.e. trivalent elements such as born and indium
and pentavalent elements such arsenic and phosphorus.
CARRIER CONCENTRATION IN SEMICONDUCTORS
Intrinsic Semiconductors
As we know that in semiconductors the magnitude of the energy gap,
very small, therefore some of the electrons lying on the top of the valence band gain
sufficient thermal energy (which is the most usual source, other energy sources may be
electrical or radiant) even at moderate temperatures and get excited into the conduction
band leaving behind equal number of holes in the valence band (in a pure
semiconductor under thermal equilibrium the number of electrons and holes are equal
since they are produced as electron-hole pairs). These holes attract electrons, in the
process one of the neighboring electrons will jump to fill the vacancy.
When an electric field is applied, the electrons drift towards the positive
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.27
he sam
are p l l ec
an
m co
( )
ndi
d tempe
( ) (w
r
f ma
( )
ter
. T
( ) = ( ) =
and to int
take
= . S
( )
o tha
=
t,
( ) =
electrode and constitute a current. Similarly, the holes move towards the negative
electrode and constitute a current. However, since the electrons and holes have opposite
charges, the current due to both is in t e direction. Their drift
velocities under the action of an ap ied e tric field
= μ (1)
= μ
where the subscripts d refer to electrons and holes, respectively. The
proportionality factor µ is known as the mobility (the freedom of charge movement) and
is a function of thermal vibration of the atoms and the crystal structure. Under equilibriu
tions, in intrinsic semiconductors, the rate of generation of electron hole
pair epends on the nature of material and the ature. To maintain a
constant concentration of pairs, the rate of recombination hich depends on the
properties o ial and is proportional to the concentrations of two charges) must be
equal to hen
(2)
where , refer rinsic electron and hole densities, respectively. The
proportionality constant s into account the charge densities and properties of the
semiconductor material.
Now for pair generation,
= (3)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.28
an are
o
=
Since the densities d inherent property of a semiconductor at a given
temperature, then the product r the square of the intrinsic charge density, also has a
fixed value at a given temperature and depends only on the nature of material. This result
is analogous to the constant solubility product in chemistry.
Extrinsic Semiconductors
In the preceding section, we observed that when a small quantity of suitable
impurity element from group V of the periodic Table is added to pure material, n-type
semiconductor results, where electrons act as majority charge carriers. Similarly, the
addition of suitable group III element produces p-type semiconductor, where holes are
majority charge carriers.
In both n and p type semiconductors, the electrons which create the ions (donor
impurity atoms become positively ionized after donating electrons and acceptor atoms
become negatively ionized) receive relatively higher energy than needed to create
electron-hole pairs. Therefore, it is possible to establish conditions for conduction by
holes in p-types semiconductors, and conduction by electrons in n-type semiconductors
at temperature much below than the temperature at which intrinsic conduction by holes
and electrons occur. Since no charge is added or removed during the above processes, the
crystal as a whole is assumed to be electrically neutral.
Because of the ease with which the electrons from the donor impurity can be excited
to the conduction band, it is reasonable to assume that these (donor impurity) atoms are
totally ionized at usual ambient temperatures. Hence, the density of conduction
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.29
h ge neu rality
+
of th
= an
ar
i
electrons, ND can be assumed equal to the density of donor impurity atoms. Likewise, the
acceptor states may assume to be filled completely in a p-type semiconductor at usual
temperatures, and so the density of holes in the valence band, NA is taken as equal to the
density of acceptor atoms.
As a matter of fact, the concentration of impurity atoms is very low (usually one
atom per 106-1010 host atoms) as compared to the parent semiconductor atom and
therefore impurity doping does not appreciably affect the thermal generation and
recombination process. Making use of the argument that the recombination is
proportional to concentrations of two charges, for p-type material, the eqn (3) becomes
= (4)
It is to be noted that when the concentration of majority charges is increased over its
intrinsic value by doping, the concentration of minority charges is found to decrease to
maintain the concentration product, ni2, constant. Based on above considerations, the
amount of reduction in the minority charges can be determined by the use of an
equation expressing the overall c e material as
Now, making use of the equality np = n 2 we have
(5)
So that from eqn (5) we have
d = (6)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.30
− + = − +
− − ) + =
uation are
−
2 ±
[( −
4
) ] +
= −
2 ±
[( −
4
) ] +
0, he ho
+
le den sity
1 + (
2
/ )
ms, th
+
2
appr
2
1 +
ate h
2
le den
= 2
+ 2
1 + 2
+ ≅
The roots of the above eq
(7)
Similarly
(8)
For an intrinsic semiconductor, where NA = ND = 0, from eqs. (8) and (9), we
obtain n = p = ni, as they should be. These are general equations for charge densities and
let us see how they change under special cases:
Case I: n-type Semiconductors
In this case, ND >> ni and NA ≅ then t is
− 4
= (9)
The second term under the radical is less than unity and may be expanded as a power
series. Neglecting the higher power ter e oxim o sity can be found as
= − = (10)
Similarly, eqn (9) provides the electron density in n-type material as
(11)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.31
then the
≅
ch
≅ + ≅
Case II: p-type Semiconductors
In this case, NA >> ni and ND ≅0, arge densities can be found as
(12)
(13)
The density of majority carriers is thus approximately equal to the density of
impurity atom (as in eqs. 11 and 13) at usual ambient temperatures. On the other hand,
because of recombination process the density of minority carriers remains significantly
below the intrinsic ni, or pi, level.
THE P-N JUNCTION
The pn junction consists of an abrupt discontinuity or a graded distribution of the
two types of doping impurity across a region of a particular semiconductor sample. The
electrostatic conditions at the junction can be obtained if we consider the special case
where a piece of p-type material is brought into intimate contact with a piece of n-type
material. The energy-band diagram at the moment of contact is given in Fig. 1 (a).
Fig. 1 Energy-band diagram for an abrupt pn junction: (a) before contact (b) at thermal
equilibrium
brup
( ) = 0
t jun ion, N
(
a an
) =
d Nd are constants, inde
− < <
( ) =
,
, ( ) = 0 < <
0
ities a
=
pressed in
=
t
1
er ms of the condition, if the internal field is ex electrostatic potential:
Owing to the concentration gradient, electrons diffuse from the conduction band
of the n-type region into that of the p-type region and recombine with the free holes by
dropping into the valence band. The space charge, produced by the negatively charged
acceptors in the p-type side and the positively charged donors left behind in the n-type
side, exerts a repulsive force on further charges crossing the junction, and inhibits further
flow. In terms of the energy-band diagram, the Fermi energy of the p-type region is raised
with respect to that of the n-type region until the Fermi levels become equal, as in
Fig.1(b). The transferred charge which terminates on the immobile impurity atoms in the
transition region produces an electric field which is directed from the n-type side to the
p-type side. This built-in potential barrier Vint corresponds to the difference between the
Fermi levels on the n and p sides before contact was made.
Let Nd and Na be the concentrations of charged donor and acceptor impurities, and
xd, xa be the thicknesses of the space charge regions on each side of the junction
respectively. For an a ct pendent of x:
0
and we may assume that all impurities are ionized, so that thermal equilibrium carrier
dens t the edges of the space charge regions will be taken as n0 (xd) = Nd and p0
− Na. The built-in potential barrier results from the thermal equilibrium
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.32
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.33
= ( ) − ( −
a
)
x = x
l
d,
n
one o
(
(−
btai
)
) =
(− ) . Henc e
= l n
o
: terial of
=
diele c Ԑ
< <
= for <
0
ions of zer o f
=
ld at both extre
=
=
On integration over the transition region, from x = -x to ns:
(1)
We can take n0 (xd) = Nd and n0 (-xa) = = , the thermal
equilibrium potential barrier, whose n-type side is positive with respect to the p-type
side can be written as:
(2)
The variation of the potential in the depletion region is btained by solving the
Poisson equation within the ma tric constant
for −
− 0 < (3)
subject to the boundary condit ie mities:
0 (4)
Since the pn junction is made from separate pieces of p-type and n-type material, we must
have the same charge per unit area on each side of the transition region, due to charge
separation:
(5)
On the p-type side of the junction (-xa <x <0), integrating Eq.(3) and bearing in mind the
boundary condition (4), we obtain the field distribution:
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.34
= ( + )
( ) = ( + 2
) + (7
nt ca
( )
n b
=
e eliminate
( +
2
d by
) f
ing th
<
zero
<
(− ) (0)
or −
2
n sid
( )
e (
=
x < xd):
( −
2 ) +
ontinuity at x
( )
=0
=
allo
2
ws us to
+
e v
2
a lua
−
te the
(
2
const
− )
ant, g iving:
0 < < (1
( ) = 2
+ 2
= 2
( + )
(6)
Integrating Eq.(6) gives:
)
where the arbitrary consta defin e of potential at - xa :
0 (8)
= 0 and = (9)
Similarly we obtain on the 0<
−
and the c
0)
From Eq.(10) it follows that:
The built-in potential barrier Vint = V(xd) -V(-xa) = V(xd) can be written, using the
condition (5), as:
(11)
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.35
= + for
= +
= +
we obt
=
2 ( + )
= 2
at therm
+
al eq
From Eq.(5), introducing the total width of the depletion region, we
have
(12)
On substitution into Eq. (11) ain:
(13)
so that the width of the depletion region uilibrium can be written as:
(14)
Fig. 2 Energy-band diagrams for a pn junction: (a) forward bias (b) reverse bias
A pn junction is in a state of dynamic equilibrium and can be described, in terms
of the energy-band diagram shown in Fig.1, as a balance between majority carriers,
moving from n-type to p-type side and climbing the potential barrier, and an equal number
of minority carriers, moving down the barrier from the p-type to n-type side. When a
potential difference V is applied across the junction, so that the p-type
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.36
side is biased positively with respect to the n-type side (V >0 or forward bias), the barrier
to diffusion is lowered, as in Fig. 2 (a), resulting in large forward currents. Reversing the
polarity of the applied voltage (V <0 or reverse bias), the barrier to diffusion is raised, as
in Fig. 2(b), so that there are only the minority carriers near the transition region that
produce a small reverse current.
Energy-band diagrams of inhomogeneous and homogeneous semiconductors
It is often convenient to represent inhomogenous semiconductors by energy- band
diagrams which visualize the electrical behaviour described by the carrier transport
equations. Most phenomena can be understood if we use a one-dimensional effective
energy-band model, shown in Fig.1, where Ec is the lowest energy in the conduction band
and Ev is the highest energy in the valence band for any possible crystallographic
direction. In other words, Eg = Ec - Ev is the minimum thermal energy required to excite
electrons from the valence band into the conduction band and the x axis represents an
averaged direction through the crystal. This convention is illustrated in Fig.1for a
homogenous semiconductor.
Fig.1 The band structure in a homogenous semiconductor
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.37
tion for
(∆
th
) =
e exce
−
s elec
+ (
∆
ron co
)µ
ncentration
+ µ
∆
(∆ ) +
( ∆ )
, µ rep
he mot
(∆ ) =
of the
−
xc
∆
ol oncentration
+ µ
∆p
(∆ ) +
( ∆ )
, µ rep
For an inhomogenous semiconductor at thermal equilibrium, the Fermi energy is
always represented in such diagrams by a horizontal line, as in Fig. 2. Any local increase
or decrease in the carrier potential energy with respect to the average effective value
will be represented by an upward or downward bending of the bands. It is apparent that,
in an energy-band representation, the electrons move towards the bottom of the
conduction band, while the holes move towards the top of the valence band.
Fig. 2 The energy-band diagram for an inhomogenous semiconductor
at thermal equilibrium
Steady-state diffusion of excess minority carriers
Equation of mo s t
∆ (1)
Where, , and resent the generated electrons, mobility, diffusion time
and diffusion constant of electrons.
Equation for t ion e ess h e
+ (∆ )µ (2)
Where, , and resent the generated holes, mobility, diffusion time and
diffusion constant of holes.
Unit-2 Theory of metals and semiconductors, Dr P Sagayaraj 2.38
les in
=
ion e
(∆
q
) ∆ − =
usion length f
=
or
+
ctron
= nd E=0
(∆
in eq
) −
(1)).
=
∆
= +
= i
n.
For the steady-state motion of excess ho a n-type semiconductor we must take
d(∆p)/dt = 0 in Eq.(2). If we assume that 0 and that the applied field is E
= 0, Eq.(2) reduces to the diffus uation:
0 (3)
where it is usual to introduce a diff holes as:
(4)
Eqn (3) has the general solution:
∆ = (5)
Where, a and b are arbitrary constants. In a similar manner, the steady-state diffusion
equation for excess ele s in a p-type semiconductor may be derived from eqn (1) as:
(take d(∆n)/dt = 0; 0 a
0 (6)
The solution is:
∆ (7)
where, s the diffusion length for electrons and c and d are arbitrary
constants.