Chapter 2-2.

Carrier properties

Mass like charge is a very basic property of electrons and holes.

The mass of electrons in a semiconductor may be different than its
mass in vacuum.

Effective mass concept

dv dv
F = −qΕ = m0 F = −qΕ = m*n
dt dt 1
 Effective mass

• Electrons moving inside a semiconductor crystal will collide with

semiconductor atoms, thereby causing periodic deceleration of the
carriers
• In addition to applied electric field, the electrons also experience
complex field forces inside the crystals
• The effective mass can have different values along different
directions
• The effective mass will be different depending on the property we
are observing. So you can have conductivity effective mass,
density of states effective mass, etc.

2
3
 Carrier numbers in intrinsic materials
Intrinsic semiconductor or pure semiconductor has equal numbers
of electrons and holes at a particular temperature.

Number of electrons/cm3 [n] = number of holes/cm3 [p]

Why is n = p?

This is an intrinsic property of the semiconductor and is called

intrinsic carrier concentration, ni

At T = 300 K, ni = 2  106 / cm3 in GaAs

1  1010 / cm3 in Si
2  1013 / cm3 in Ge

How large is this compared to the number of Si atoms/cm3?

What happens to ni at higher temperature? At 0 K?
4
 Extrinsic semiconductors

Elements in column V of the periodic table have 5 electrons in

their outer shell (one more than Si)! This can be easily released,
thus increasing the net free electrons in the Si crystal.

Elements of column III of the periodic table have only 3 electrons

in their outer shell (one less than Si)! To complete the bond, the
atom can accept an electron by breaking a bond somewhere else,
thus creating a broken bond, or a hole.
5
Two-dimensional representation of Si lattice

6
Visualization of (a) donors and (b) acceptors

Phosphorus (P) atom Boron (B) atom

7
 Pseudo-hydrogen atom model for donors

Instead of m0, we have to use mn*.

Instead of o, we have to use Ks o.

Ks is the relative dielectric constant

of Si (Ks, Si = 11.8).

m0 q 4
EH = − = − 13.6 eV (see page 24 of text)
2 ( 4  0  ) 2
* * 2
mn q 4
mn  0 
Ed = − = − 13.6 eV    − 0.05 eV
2 (4π Ks 0 ) 2 m0  Ks0 

This is an approximate value. More accurate values are given next. 8

 Binding energies for dopants

Questions:
How much energy is required to break a Si-Si bond?
How much energy is required to break the 5th electron from As in
Si?
How much energy is required to break a Si-Si bond when that
bond is adjacent to a B atom?
Does the freeing of an electron from a donor atom create an extra
hole? 9
Energy-band model for donors

10
Temperature effects on donors and acceptors

11
 Some items to remember

• Ionization energy of electrons in hydrogen

• Bond and band models of semiconductors
• Intrinsic semiconductors and intrinsic carrier concentrations.
• Extrinsic semiconductors
• Dopants: Donors and acceptors
• n-type material
• p-type material
• Majority carriers
• Minority carriers
• Band models of donors and acceptors
• Band gap energies of Si, Ge and GaAs

12
 Chapter 2-3. States and carrier distributions

So far we have concentrated on carrier properties of qualitative

nature. We also need to:

• Determine the carrier distribution with respect to energy in

different bands.
• Determine the quantitative information of carrier
concentrations in different bands.

Two concepts will be introduced to determine this:

• Density of states
• Fermi-Dirac distribution and Fermi-Level

13
 Density of states

• There are 4 states per atom or 4  51022 / cm3 states in each

of conduction and valence bands of Si.
• The distribution of these states in the bands are not uniform,
but follows a distribution function given by the following
equations.

14
Dependence of DOS near band edges

15
 More on density of states (DOS)

• gc(E) dE represents the # of conduction band states/cm3 lying in

the energy range between E and E + dE
• gv(E) dE represents the # of valence band states/cm3 lying in the
energy range between E and E + dE
• More states are available available away from the band edges,
similar to a seating arrangement in a football field
• Units for gc(E) and gv (E): # per unit volume per unit energy,
i.e., # / (cm3 eV)
• Energy bands are drawn with respect to electron energies

16
 Fermi-Dirac distribution and the Fermi-level

Density of states tells us how many states exist at a given energy

E. The Fermi function f(E) specifies how many of the existing
states at the energy E will be filled with electrons. The function
f(E) specifies, under equilibrium conditions, the probability that an
available state at an energy E will be occupied by an electron. It is
a probability distribution function.

EF = Fermi energy or Fermi level

k = Boltzmann constant = 1.38  10−23 J/K
= 8.6  10−5 eV/K
T = absolute temperature in K
17
 Distribution function for gas molecules

Example: Gas molecules follow a different distribution

function: The Maxwell-Boltzmann distribution

− E2
−( E2 − E1 )
n2 e kT
= = e kT
n1 − E1
e kT

Let us look at the Fermi-Dirac distribution more closely.

18
 Fermi-Dirac distribution: Consider T → 0 K

1
For E > EF : f ( E  EF ) = = 0
1 + exp (+)

1
For E < EF : f ( E  EF ) = =1
1 + exp (−)

EF

0 1 f(E) 19
 Fermi-Dirac distribution: Consider T > 0 K

If E = EF then f(EF) = ½
E  EF + 3kT  E − EF 
If then exp    1
 kT 
 − ( E − EF ) 
Thus the following approximation is valid: f ( E ) = exp  
 kT 

i.e., most states at energies 3kT above EF are empty.

 E − EF 
If E  EF − 3kT then exp    1
 kT 
 E − EF 
Thus the following approximation is valid: f ( E ) = 1 − exp  
 kT 
So, 1 −f(E) = Probability that a state is empty, decays to zero.
So, most states will be filled.
kT (at 300 K) = 0.025eV, Eg(Si) = 1.1eV, so 3kT is very small
in comparison.
20
Temperature dependence of Fermi-Dirac distribution

21
Exercise 2.3

22
 Example
Assume that the density of states is the same in the conduction band
(CB) and valence band (VB). Then, the probability that a state is
filled at the conduction band edge (EC) is equal to the probability
that a state is empty at the valence band edge.
Where is the Fermi level located?
1 1 1
f ( EC ) = 1 − f ( EV ) =1 − =
EC − E F EV − EF EF − EV
1 + e kT 1 + e kT 1 + e kT

EC + EV
EC − EF = EF − EV EF =
2
This corresponds to intrinsic material, where the # of electrons at EC
= # of holes (empty states) at EV. Note that the probability within
the band gap is finite, but there are no states available, so electrons
cannot be found there. 23
 Equilibrium distribution of carriers

Distribution of carriers = DOS  probability of occupancy

= g(E) f(E)

(where DOS = Density of states)

Total number of electrons in CB (conduction band) =

E top
n0 =  g C ( E ) f ( E ) dE
EC

Total number of holes in VB (valence band) =

g V ( E ) (1 − f ( E ) ) dE
EV
p0 = 
EBottom

24
Fermi-level positioning and carrier distributions

25
 Visualization of carrier distribution

• One way to convey the carrier distribution is to draw the

following diagram. This diagram represents n-type material since
there are more electrons than holes.

26
 Visualization of carrier distribution (continued)

Another more useful way to convey the carrier distribution is to

draw the following band diagrams. The position of EF with respect
to Ei is used to indicate whether is n-type, p-type or intrinsic.

27