4.

Density of States /
Fermi-Dirac Statistics
• Calculation of density of states
• Density of states for specific materials
• Characterization of effective mass
• Rules of filling electronic states
• Derivation of Fermi‐Dirac Statistics
• Intrinsic carrier concentration
• Potential, field, and charge
• E‐k diagram vs. band‐diagram
Suggested readings: Pierret Ch. 4, Neamen Ch. 3, Ch.4.1
EE 5530 C. F. Huang 1
 Carrier Density

Carrier number = Number of states x filling factor

Carrier = moving charges in semiconductors

Total number of occupants

= Number of apartments
x The fraction occupied

EE 5530 C. F. Huang 2
 Density of States
How many density of states per unit energy?

∆k
States between E1 + ∆E & E1 = 2 × ( + k and − k )
δk
∆k
= 2×
2π / Na

States per unit energy @ E1

Na ∆k
=
π ∆E

EE 5530 C. F. Huang 3
 1D DOS
L

Na ∆k Na dk
States per unit energy
= =
π ∆E π dE
L
 2k 2 2m∗ ( E − E0 )
E − E=
0 k
⇒=
2m ∗
2
m*
dk m∗
=
dE 2 2 ( E − E0 )

States per unit energy per unit length

1 dk 1 m∗
≡ DOS = =
π dE π 2 2 ( E − E0 )

=2π/L ×2 if spin is considered

EE 5530 C. F. Huang 4
 1D DOS (continued)

∫
Band I
DOS ( E )dE = constant

EE 5530 C. F. Huang 5
 2D DOS

Show that 2D DOS is a constant independent of energy!

EE 5530 C. F. Huang 6
 3D DOS

States between E1 + ∆E & E1

4 4 3
π (k + ∆k ) − π k
3
V 2
=2( spin ) × 3 3 = k ∆k
2π 2π 2π π 2

L W H
V ∆k
States per unit energy @ E1 = 2 k 2
π ∆E
2 2
 2kx2  k y  2kz 2  2k 2
2
 k 2
2m ( E − E0 )
∗
dk m ∗
E − E= + + =
E − E=
0 k
⇒= 2
⇒ =
0
2 m∗ 2 m∗ 2 m∗ 2 m∗
2m ∗
 dE 2 2 ( E − E0 )

States per unit energy per unit volume

m∗
≡ DOS (3D=
) 2 m ∗
( E − E0 )
π 
2 3

EE 5530 C. F. Huang 7
 3D DOS(continued)

EE 5530 C. F. Huang 8
 DOS of GaAs

Conduction band
mn∗ 2mn∗ ( E − Ec )
gc ( E ) =
π 2 3

Valence band
 m∗ 2 m∗ ( E − E )
 hh hh v

 π 2 3

gv ( E ) = 
 mlh∗ 2mlh∗ ( E − Ev )

 π 2 3

EE 5530 C. F. Huang 9
 Ellipsoidal Bands and DOS
 2 k12  2 k22  2 k32
E − Ec= + + β
2m1 2mt
∗ ∗
2mt∗
α
2 2 2
k k k
1= 1
+ + 2 3
β
 2ml∗ ( E − Ec )   2mt∗ ( E − Ec )   2mt∗ ( E − Ec ) 
  2    2    2  X Nel
     
k12 k22 k32
= + +
α 2
β 2
β2
4 4 3 keff
2
Volume
= k − space N
= el  παβ  π keff
3  3
transform with volume conserved
3
4 2m ( E − EC ) 2m ( E − Ec ) 2m ( E − Ec ) 4  2mn∗ ( E − Ec ) 
∗ ∗ ∗
N el π l
≡ π
t t

3 2 2 2 3  2 
2 ∗2 1
m = N el (m m )
∗
n
3 ∗
l t
3
Density of states effective mass
EE 5530 C. F. Huang 10
 DOS of Ge

N
= el 8=
/2 4
2 ∗2 1
m = 4 (m m )
∗
n
3 ∗
l t
3

Conduction band
mn∗ 2mn∗ ( E − Ec )
gc ( E ) =
π 2 3

EE 5530 C. F. Huang 11
 DOS of Si

N el = 6
2 ∗2 1
m = 6 (m m )
∗
n
3 ∗
l t
3

Conduction band
mn∗ 2mn∗ ( E − Ec )
gc ( E ) =
π 2 3

EE 5530 C. F. Huang 12
 DOS in valence band

Valence band
( E − Ev ) mlh∗ 2mlh∗ ( E − Ev ) m p 2m p ( E − Ev )
∗ ∗
∗
mhh 2mhh
∗

gv ( E ) = + =
π 
2 3
π 
2 3
π 2 3
∗ 3/ 2
=m∗p [(mhh ) + (mlh∗ )3/ 2 ]2 / 3

EE 5530 C. F. Huang 13
 Carrier Density

Carrier number = Number of states x filling factor

DOS

EE 5530 C. F. Huang 14
E‐k diagram and Electronic States
Number of
states per
E-k diagram Density of states unit energy
per unit
volume

EE 5530 C. F. Huang 15
 Rules for filling up the States

In equilibrium
• Pauli Principle: Only one electron per state
• Total number of electrons is conserved NT = ∑i Ni
• Total energy of the system is conserved ET = ∑ i Ei N i

EE 5530 C. F. Huang 16
 Derivation of Fermi‐Dirac Statistics
Considering an illustrative example: 3 Energy Levels

NT
= ∑
= N 5i i

=ET ∑
= E N 12
i i i

2! 5! 7! 2! 5! 7 ! 2! 5! 7!
W203 = ⋅ ⋅ W122 = ⋅ ⋅ W041 = ⋅ ⋅
1!2 ! 0!5! 3! 4! 1!1! 2!3! 5!2! 0!2 ! 4!1! 6!1!
= 35 = 420 = 35
Most probable!

EE 5530 C. F. Huang 17
 Occupation Statistics

EE 5530 C. F. Huang 18
 Occupancy probability

f(E): Occupancy probability at E

EE 5530 C. F. Huang 19
 For general cases
For Si‐states with Ni stated being occupied
maximum

Si !
W =∏ 2! 5! 7!
i ( Si − N i )! N i ! recall that W122 = ⋅ ⋅
1!1! 2!3! 5!2!

ln(W
= ) ∑ [ln S !− ln( S
i
i i − N i )!− ln N i !]
Stirling approx.
≅ ∑ [ Si ln Si − Si − ( Si − N i )ln( Si − N i ) + ( Si − N i ) − N i ln N i + N i ]
i

= ∑ [ S ln S − ( S
i
i i i − N i )ln( Si − N i ) − N i ln N i ]

EE 5530 C. F. Huang 20
 Find Max. with Lagrange‐Multiplier
dln(W )=0

∂lnW
dln(W )= ∑ dN i since
= NT ∑
= N i i constant
∂N i
→ ∑ i dN i =
i
0
  Si  
= ∑ ln  − 1 dN i  − α ∑ dN i − β ∑ Ei dN i
 Ni 
i   i i
since
= ET ∑
=i
EN i i constant
  Si  
∑i ln  N − 1 − α − β Ei dN i → ∑ i Ei dN i =
0
  i  
=0

EE 5530 C. F. Huang 21
 Continued
  Si  
ln  − 1 − α − β Ei  =
0
  Ni  
N 1
f (E) ≡ i = α +β E
Si 1 + e
f max = 1
1
at E EF , f ( EF ) ≡
= ⇒α= + β EF 0
2
Ni 1 1
f ( E=
) = =
S i 1 + e β ( E − EF ) 1 + e ( E − EF ) / k B T

1
at E → ∞, f ≅ f Boltzman
= ( E ) Ae − E / kB=
T
⇒β
k BT

EE 5530 C. F. Huang 22
Another Perspective: Detailed Balance

E1 + E2 = E3 + E4
 f 0 ( E1 ) f 0 ( E2 ) [1 − f 0 ( E3 ) ][1 − f 0 ( E4 ) ] − 
A =0
 f 0 ( E3 ) f 0 ( E4 ) [1 − f 0 ( E1 ) ][1 − f 0 ( E2 ) ]

1
The only solution is f 0 ( E ) =
1 + e β ( E − EF )
Pauli Principle, energy, and number conservation all satisfied

Analogy:

EE 5530 C. F. Huang 23
 Few comments on Fermi‐Dirac
Statistics
• Applies to all spin‐1/2 particles (Pauli exclusive principle)
• Information about spin is not explicit; multiply DOS by 2.
May be more complicated for magnetic semiconductors.
• Coulomb‐interaction (coulomb blockade) among
particles is neglected, therefore it applies to extended
solids, not to small molecules

Physically apart

EE 5530 C. F. Huang 24
 Carrier Distribution

EE 5530 C. F. Huang 25
 Electron Concentration in 3D solids
Etop
n=∫ g c ( E ) f ( E )dE
EC

Etop mn∗ 2mn∗ ( E − EC ) 1

=∫ β ( E − EF )
dE
Ec π 
2 3
1+ e
∞mn∗ 2mn∗ ( E − EC ) 1
≅∫ β ( E − Ec ) β ( Ec − EF )
dE
Ec π  2 3
1+ e e
2
= NC F1 2 (ηc )
π
∞
ξ dξ
where F1 2 (η ) = ∫ ξ −η
, Fermi-Dirac integral of order 1/2
0
1+ e
ξ= E − EC
ηc ≡ β ( EF − EC )
32
 2π mn∗ 
NC ≡ 2  2  , Effective density of conduction band states
 hβ 
EE 5530 C. F. Huang 26
 Boltzmann vs. Fermi‐Dirac Statistics
2
n NC F1 2 (ηc ) → N C eηc if − ηc ≡ β ( EC − EF ) > 3 or EC − EF > 3kT
π

2
F1 2 (ηc )
π

EE 5530 C. F. Huang 27
 Effective Density of States

2
=n NC F1 2 (ηc ) → N C eηc if − ηc ≡ β ( EC − EF ) > 3
π

Delta function with

strength Nceηc

EE 5530 C. F. Huang 28
 pn product

if β ( EC − EF ) > 3 and β ( EV − EF ) < −3

n = N C e − β ( Ec − E F )
p = N V e + β ( Ev − E F )
− β ( E c − Ev ) 3kT
n× p =N C NV e
− β Eg
= N C NV e
3kT

EE 5530 C. F. Huang 29
 Fermi‐Level for Intrinsic
Semiconductors
• Intrinsic Semiconductors: no impurity, no defects
n= p= ni
− β Eg
ni2 = N C NV e
−β E g 2
ni = N C NV e
EF ≡ Ei
p ⇒ N C e − β ( Ec − Ei ) =
n= NV e + β ( Ev − Ei )
Eg 1 NV n=p
Ei
= + ln
2 2β NC

ni = 1010/cm3 in Si at room temp. not enough

to make electronic devices
EE 5530 C. F. Huang 30
 Intrinsic Carrier Density

EE 5530 C. F. Huang 31
 E‐k diagram vs. band‐diagram

EE 5530 C. F. Huang 32
 Position Resolved E‐k Diagram

m0

m*

EE 5530 C. F. Huang 33
 E‐k Diagram vs. Band‐diagram

EE 5530 C. F. Huang 34
 Potential, Field and Charge

P.E=. Ec − Eref

Eref
− qV = Ec − Eref

dV 1 dEc
ε=
− =
dx q dx

ρ dε d 2V
= = − 2
ε s dx dx

EE 5530 C. F. Huang 35
 Effective Density of States

2
=n NC F1 2 (ηc ) → N C eηc if − ηc ≡ β ( EC − EF ) > 3
π

EE 5530 C. F. Huang 36
 E‐k diagram vs. band‐diagram

EE 5530 C. F. Huang 37
 Alternative Expression for n

By defintion,
= i when E F
n n= Ei
ni = N C e − β ( Ec − Ei )
− β ( Ec − E F ) − β ( Ec − Ei ) − β ( Ei − EF ) β ( EF − Ei )
=n N=
C e N C e = e ni e

Similarly,

p = ni e β ( Ei − EF )

EE 5530 C. F. Huang 38