Semiconductor Optoelectronics

Lecture 4: Nonequilibrium Excess Carriers in Semiconductors

Asst. Prof. Dr. Ghusoon Mohsin Ali

4th year Electronics & Communication
Department of Electrical Engineering
College of Engineering
Mustansiriyah University
 Text book
Semiconductor Physics
and Devices
Basic Principles
Third Edition

Donald A. Neamen

Chapter 6: Nonequilibrium Excess Carriers

in Semiconductors

Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display.
 Continuity Equations

Let us consider the flux of particles into and out of a small box.
Assume the flow of particles only occurs in 1-D along the x-axis and the box
is a differential volume with dimension dx by dy by dz.

Assume our particles are holes. Then using a first-order Taylor expansion we
can relate the flux of particles into the box to the flux of particles out of the box as
Fpx
Fpx  x  dx   Fpx x   dx
x
 Continuity Equations

The net increase in particles within the volume would be the difference of the
two fluxes (multiplied by the surface area)
   

 

Fpx  x   Fpx  x  dx  dy

 dz  
F px
dx  dy dz
Area  x 
But the net increase in particles inside the box could also be written
p
dx dy dz
t Volume

So
p Fpx
dx dy dz   dx dy dz
t x
This only represents the buildup or decrease of particles in the box due to
different flow rates however.
 Continuity Equations

We also need to account for the effects of generation and recombination that
may be occurring within the box. These two phenomena would also
contribute to an increase or decrease of the particle concentration within the
box. Including these terms we have:

p Fpx p
dxdy dz   dxdy dz  g p dxdy dz  dxdy dz
t x  pt
p is the density of holes and pt is the combined hole lifetime (it includes both
the thermal equilibrium carrier lifetime and the excess carrier lifetime).
Dividing through by dx·dy·dz we reach the Continuity Equations.

p Fp
  gp 
p
Holes 
t x  pt
n Fn
  gn 
n
Electrons 
t x  nt
 Time-Dependent Diffusion Equations

Our current densities are given by

p
J p  e p pE  eD p
x
n
J n  e n nE  eDn
x
By dividing the current density by the unit of charge we obtain particle flux:

Jp p
 F   p pE  D p


x
p
e
Jn n
 Fn    n nE  Dn


 e x
 Time-Dependent Diffusion Equations

By substituting these expressions into the Continuity Equations we get:

  pE 

p Fp p 2 p p
  gp    p  Dp 2  g p 
t x  pt x x  pt
n Fn n  nE   2n n
  gp    n  Dn 2  g n 
t x  nt x x  nt
Both p (or n) and E can be functions of position so we need to use the chain rule

  pE  p E
E p
x x x
 nE  n E
E n
x x x
 Time-Dependent Diffusion Equations

Thus
p  p E  2 p p
  p  E p   Dp 2  g p 
t  x x  x  pt
n  n E   2n n
  n  E n   D  g 
t   x 2  pt
n n
 x x 
If we assume we have a homogeneous semiconductor (the doping concentration
of electrons and holes is uniform throughout the semiconductor), then

px   p0  px 
nx   n0  nx 
and our partial derivatives of p(x) and n(x) just become partial derivatives of
p(x) and n(x).
 Time-Dependent Diffusion Equations

Thus we get the time-dependent diffusion equations for electrons and holes in
a homogeneous semiconductor:

p   2 p   p  E  p

 Dp   p E  p   g 
t x    pt
p
 
2
x x
n   2 n   n  E  n
 Dn  n  E n   gn 
t x 2
 x x   nt
 Example 4

The concentration of holes in a semiconductor is given by p(x)= 5 ×1015 e-x/Lp

cm-3 for x≥0. Determine the hole diffusion current density at

a) x=0
b) x=Lp

if the material is silicon with Dp=10 cm2/s and Lp=50 m

a) x=0

b) x=Lp
 Example 5

The steady-state electron distribution in silicon can be approximated by a

linear function of x. The maximum electron concentration occurs a x=0 and is
n(0)=2×1016 cm-3.

At x=0.012 cm, the electron concentration is 5×1015 cm-3. If there electron

diffusion coefficient is Dn=27 cm2/s, determine the electron diffusion current
density
 Ambipolar Transport

• Since the internal E-field creates a force attracting the electrons and holes, this E-
field will hold the pulses of excess electrons and holes together.
• The negatively charged electrons and positively charged holes then will drift or
diffuse together with a single effective mobility or diffusion coefficient.
• This phenomenon is called Ambipolar diffusion or Ambipolar transport.
 Ambipolar Transport Equations

Since electrons and holes are generated in pairs

n p
Similarly, Rn   Rp   R
 nt  pt
And since they are generated in pairs (assuming charge neutrality), n = p.
Thus we can write the time-dependent diffusion equations as:

 p   n   2 n    n  E 

  Dp   p E  p  g R
t t x 2
 x x 
 n   2 n    n  E 
 Dn  n  E n  g R
t x 2
 x x 
Multiplying the first equation by nn and the second equation by pp we get:
 Ambipolar Transport Equations

n   2 n   n  E 

n n   n nD p  n  p nE  p    n n g  R 
t x 2
 x x 
n   2 n   n  E 
p p   p pDn    p
n p  E  n    p pg  R 
t x 2
 x x 
E
Adding these we eliminate the term
x

n n   p p   n nDp   p pDn  2  n  p  p  nE n  n n   p p g  R

n   2 n 
t x x

We can simplify/rewrite this by dividing through by (nn+pp).

 Ambipolar Transport Equations

The Ambipolar Transport Equation

n   2
n  n 
D '
 E
'
gR
t x 2
x
 n nD p   p pDn n  p  p  n
D 
'
 
'
n n   p p n n   p p
Dn Dp kT
However, since  
n p q

HW Dn D p n  p 
D 
'

Dn n  D p p
 Ambipolar Transport Equations

Dn D p n  p  n  p  p  n
D 
'
 
'

Dn n  D p p n n   p p
For “strongly” p-type (or n-type) material under low level injection this
reduces considerably.

For p-type, low-level injection For n-type, low-level injection

p0  n0 n0  p0
p0  n n0  n
D '  Dn D'  Dp
 '  n  '   p
 Ambipolar Transport Equations

Also for low-level injection (small # of excess carriers) the probability of

recombining will be almost constant (the chance of hitting a majority carrier won’t
change much), so nt = n and pt = p.
For generation and recombination we have a combination of thermal-equilibrium
generation/recombination and excess carrier generation/recombination. Looking
at just electrons we have:
  
g  R  g n  Rn  Gn 0  g n'  Rn 0  Rn' 
However, for thermal equilibrium we know Gn0 = Rn0 so
n
g  R  g n'  Rn'  g n' 
 n0
and we would have a similar expression for the holes.
 Ambipolar Transport Equations

Combining all this gives our final Ambipolar Transport Equations:

 2 n  n  n n 
Dn  n E g 
'
 p  type
x 2
x  n0 t
 2 p  p  p p 
Dp  pE g 
'
 n  type
x 2
x  p0 t

These equations simply say that there can be a change in our excess carrier
concentation over time because:
1. They diffuse away.
2. They drift away.
3. More are generated.
4. They recombine.
 Quasi-Fermi Energy Levels

The thermal-equilibrium electron and hole concentrations are functions of the

Fermi energy level. We can write
 ( EC  EF )   EF  EFi 
no  N C exp    ni exp  
 kT   kT 
 ( EF  Ev )   ( EF  EFi ) 
po  N v exp    ni exp  
 kT   kT 
 Quasi-Fermi Levels

The Fermi Level is constant in a sample under equilibrium. However, with

generation or current flowing (even in steady-state) we are in nonequilibrium
so EF is no longer a constant.
However, we still want to know (or relate) total electron and hole
concentrations. This can be done by considering a “shift” in EF that represent
the “More” or “Less” p or n nature of the semiconductor. These new levels
are called EFn and EFp, the Quasi-Fermi Levels.

n  n0  n  ni e  EFn  EFi  kT
EFi  E Fp  kT
p  p0  p  ni e
 Quasi-Fermi Energy Levels

If excess carriers are created in a semiconductor, we are no longer in thermal

equilibrium and the Fermi energy is strictly no longer defined. If n and p are
excess electron and hole concentrations respectively, we write

 E  EFi 
no   n  ni exp  Fn 
 kT 

 ( EFp  EFi ) 
po   p  ni exp  
 kT 
 Quasi-Fermi Energy Levels

Question: Consider an n-type semiconductor at T=300K with carrier concentration

of n0=1015 cm-3, ni=1010 cm-3, and p0=105 cm-3. In nonequilibrium, assume that the
excess carrier concentrations are n=p= 1013 cm-3

The Fermi level for thermal equilibrium can be determined

 n0 
EF  EFi  kT ln    0.2982 eV
 ni 
In nonequilibrium, quasi Fermi level for electrons and holes becomes

 n  δn 
EFn  EFi  kT ln  o   0.2984 eV
 ni 
 Example 6

Question: Consider n-type silicon doped at Nd = 5 × 1015 cm-3. It is found that

EFn-EF=1.02 × 10-3 eV. The value of ni=1.5 × 1010 cm-3 . Determine
a)What is the excess carrier concentration
b)Determine EFn-EFi and EFi-Efp. Assume n=p
For n-type EFn  EF   EFn  EFi    EF  EFi 
 no   n   no   no   n ni 
 kT ln    kT ln    kT ln   
 ni   ni   ni no 
 n n 
 kT ln  o 
 no 
 5 1015   n 
0.00102   0.0259  ln  
 5  1015

 0.00102 
5 10   n  5 10 exp 
15 15
   n  2  1014
cm -3

 0.0259 
 Example 6

Question: Consider n-type silicon doped at Nd = 5 × 1015 cm-3. It is found that

EFn-EF=1.02 × 10-3 eV. The value of ni=1.5 × 1010 cm-3 . Determine
a)What is the excess carrier concentration
b)Determine EFn-EFi and EFi-Efp. Assume n=p

 n n   p 
EFn  EFi  kT ln  o  EFi  EFp  kT ln  
 ni   i 
n
 5 1015  2 1014   2 1014 
  0.0259  ln     0.0259  ln  10 
 1.5  1010
  1.5  10 
 0.33038eV  0.2460 eV