SD Carrier 1 PDF

Semiconductor Devices
Carrier Statistics
M. B. Patil
mbpatil@ee.iitb.ac.in
www.ee.iitb.ac.in/~sequel
Department of Electrical Engineering

Indian Institute of Technology Bombay
M. B. Patil, IIT Bombay

Carrier statistics in equilibrium
Ec
Eg
Ev

Ec
Eg
Ev
* The term “carrier” refers to mobile entities, viz., electrons in the conduction band (or
simply “electrons”) and vacancies in the valence band (or simply “holes”).

Ec
Eg
Ev
* We are interested in the carrier densities, i.e., electron density (n) and hole density (p),
because they are responsible for carrying a current. (The nuclei and core electrons of the
silicon atoms do not contribute to conduction.)

Ec
Eg
Ev
* We are interested in the carrier densities, i.e., electron density (n) and hole density (p),
because they are responsible for carrying a current. (The nuclei and core electrons of the
silicon atoms do not contribute to conduction.)
* We will first consider a semiconductor in equilibrium, i.e., without an external
perturbation such as an applied voltage, a magnetic field, or optical illumination.

Electron density (n) in equilibrium
E
conduction band
Ec gc (E)
Eg
Ev gv (E)
valence band

* The electron density depends on two factors: E

conduction band
Ec gc (E)
Eg
Ev gv (E)
valence band


conduction band
- How many states are available in the conduction band for the Ec gc (E)
electrons to occupy?
Eg
Ev gv (E)
valence band


conduction band
- What is the probability that a given state (at energy E ) is
occupied?
Eg
Ev gv (E)
valence band


conduction band
occupied?
Eg
* The “density of states” function gc (E ) gives the number of states
available per unit energy per unit volume.
Ev gv (E)
valence band


conduction band
occupied?
Eg
* The “density of states” function gc (E ) gives the number of states
available per unit energy per unit volume.
(mn∗ )3/2 2 (E − Ec )
p
gc (E ) = , E > Ec , where
π 2 ~3 gv (E)
Ev
mn∗ ≡ electron effective mass = 1.18 m0 for silicon at T = 300 K, valence band
m0 = free electron mass = 9.1 × 10−31 Kg,
~ = h/2π, with h (Planck constant) = 6.63 × 10−34 J-s.

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

T = 300 K. conduction band
Ec gc (E)
Eg
Ev
valence band


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
Eg
Ev
valence band


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
√ Eg
2 (mn∗ )3/2 (∆E )3/2
=
π 2 ~3 3/2
Ev
valence band


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
√ Eg
2 (mn∗ )3/2 (∆E )3/2
=
π 2 ~3 3/2
√
16 2 π mn ∆E 3/2
∗
=
3 h2 Ev
valence band


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
√ Eg
2 (mn∗ )3/2 (∆E )3/2
=
π 2 ~3 3/2
√
16 2 π mn ∆E 3/2
∗
=
3 h2 Ev
3/2 valence band
1.18 × 9.1 × 10−31 × 50 × 10−3 × 1.6 × 10−19

= 23.7 ×
(6.63 × 10−34 )2


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
√ Eg
2 (mn∗ )3/2 (∆E )3/2
=
π 2 ~3 3/2
√
16 2 π mn ∆E 3/2
∗
=
3 h2 Ev
3/2 valence band
1.18 × 9.1 × 10−31 × 50 × 10−3 × 1.6 × 10−19

= 23.7 ×
(6.63 × 10−34 )2
= 23.7 × 2.73 × 1024 /m3 = 6.5 × 1019 /cm3 .


Ec gc (E)
Ec +∆E (mn∗ )3/2
p
2 (E − Ec )
Z
N0 = dE
Ec π 2 ~3
√ Eg
2 (mn∗ )3/2 (∆E )3/2
=
π 2 ~3 3/2
√
16 2 π mn ∆E 3/2
∗
=
3 h2 Ev
3/2 valence band
1.18 × 9.1 × 10−31 × 50 × 10−3 × 1.6 × 10−19

= 23.7 ×
(6.63 × 10−34 )2
= 23.7 × 2.73 × 1024 /m3 = 6.5 × 1019 /cm3 .
N 0 would be the number of electrons per unit volume (in the conduction band)
if the states in the range Ec < E < Ec + ∆E were all occupied (and the rest of
the states unoccupied). The real picture is different.

E 1
conduction band
Ec gc (E)
0.8
f
0.6
Eg
0.4
0.2
Ev 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT

E 1
conduction band
Ec gc (E)
0.8
f
0.6
Eg
0.4
0.2
Ev 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
* The number of electrons in the interval E to (E + dE ) is not gc (E )dE but gc (E )f (E )dE .

E 1
conduction band
Ec gc (E)
0.8
f
0.6
Eg
0.4
0.2
Ev 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT

* f (E ) is the probability that the state at E is occupied.

E 1
conduction band
Ec gc (E)
0.8
f
0.6
Eg
0.4
0.2
Ev 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT

* f (E ) is the probability that the state at E is occupied.
* The probability depends on the “Fermi level” EF which typically lies in the forbidden gap.

E
conduction band
Ec (1)
gc (E) EF
(2)
EF
(3)
EF
Eg
(4)
EF
(5)
EF
(6)
gv (E) EF
Ev
valence band
f(E)
0 0.5 1

E
conduction band
Ec (1)
gc (E) EF
(2)
EF
(3)
EF
Eg
(4)
EF
(5)
EF
(6)
gv (E) EF
Ev
valence band
f(E)
0 0.5 1
* A change in the Fermi level causes the probability function to shift, and therefore the
carrier concentrations (n and p) change substantially with EF .

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
(E − EF )/kT (E − EF )/kT

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
1
* f (E ) is given by the Fermi function: f (E ) = where
1 + e (E −EF )/kT
k = 1.38 × 10−23 −5
J/K (or 8.62 × 10 eV/K) is the Boltzmann constant.

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
1
1 + e (E −EF )/kT
k = 1.38 × 10−23 −5
* At E = EF , f (E ) = 1/2.

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
1
1 + e (E −EF )/kT
k = 1.38 × 10−23 −5
* At E = EF , f (E ) = 1/2.
* If E − EF kT , e (E −EF )/kT 1, and f (E ) → 0.

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
1
1 + e (E −EF )/kT
k = 1.38 × 10−23 −5
* At E = EF , f (E ) = 1/2.
If E − EF −kT , e (E −EF )/kT 1, and f (E ) → 1.

1
E 1 fMB fMB
conduction band
Ec gc (E) f
0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
1
1 + e (E −EF )/kT
k = 1.38 × 10−23 −5
* At E = EF , f (E ) = 1/2.
If E − EF −kT , e (E −EF )/kT 1, and f (E ) → 1.
* For E − EF > 3 kT , e (E −EF )/kT > 20, which is much larger than 1. We then have
f (E ) ≈ fMB (E ) = e −(E −EF )/kT , the Maxwell-Boltzmann function.
E 1 fMB 1 T=0K
conduction band 100 K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)
* Low T : e (E −EF )/kT varies rapidly with E .

High T : e (E −EF )/kT varies slowly with E .

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)

→ f (E ) becomes broader as T increases.

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)

→ f (E ) becomes broader as T increases.
* Because of the significant variation of f (E ) with temperature, we can expect the electron
density to have a significant temperature dependence.

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)
√ Z p
∞ (mn∗ )3/2 2 ∞ (mn∗ )3/2 √ η 1/2
Z ∞
E − Ec
Z
3/2
n= gc (E )f (E )dE = 2 3 (E −E )/kT
dE = 2 3
2 (kT ) dη
Ec π ~ Ec 1 + e F π ~ 0 1 + e η−ηc

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)
√ Z p
∞ (mn∗ )3/2 2 ∞ (mn∗ )3/2 √ η 1/2
Z ∞
E − Ec
Z
3/2
n= gc (E )f (E )dE =
2 3 (E −E )/kT
dE = 2 3
2 (kT ) dη
Ec π ~ Ec 1 + e F π ~ 0 1 + e η−ηc
2 EF − Ec
→ n = Nc √ F1/2 (ηc ) with ηc = .
π kT

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)
√ Z p
∞ (mn∗ )3/2 2 ∞ (mn∗ )3/2 √ η 1/2
Z ∞
E − Ec
Z
3/2
n= gc (E )f (E )dE =
2 3 (E −E )/kT
dE = 2 3
2 (kT ) dη
Ec π ~ Ec 1 + e F π ~ 0 1 + e η−ηc
2 EF − Ec
→ n = Nc √ F1/2 (ηc ) with ηc = .
π kT
η 1/2
Z ∞
* F1/2 (ηc ) = dη, is called the “Fermi-Dirac integral of order 1/2.”
0 1 + e η−ηc

E 1 fMB 1 T=0K
Ec gc (E)
0.8 0.8 300 K 200 K
f 400 K
0.6 0.6
f
Eg
0.4 0.4
0.2 0.2
Ev 0 0
valence band
-6 -4 -2 0 2 4 6 -0.2 -0.1 0 0.1 0.2
(E − EF )/kT (E − EF ) (eV)
√ Z p
∞ (mn∗ )3/2 2 ∞ (mn∗ )3/2 √ η 1/2
Z ∞
E − Ec
Z
3/2
n= gc (E )f (E )dE =2 3 (E −E )/kT
dE = 2 3
2 (kT ) dη
Ec π ~ Ec 1 + e F π ~ 0 1 + e η−ηc
2 EF − Ec
→ n = Nc √ F1/2 (ηc ) with ηc = .
π kT
η 1/2
Z ∞
* F1/2 (ηc ) = dη, is called the “Fermi-Dirac integral of order 1/2.”
0 1 + e η−ηc
∗ 3/2
mn kT
* Nc = 2 is called the “effective” density of states for the conduction band.
2π~2

1
E 1 fMB fMB
conduction band
Ec gc (E) f
3 kT 0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
When the Fermi level is below Ec − 3kT , f (E ) ≈ fMB (E ), and

√ Z
(mn∗ )3/2 2 ∞ p
n = 2 3
E − Ec e −(E −EF )/kT dE
π ~ Ec
∗ 3/2
√ Z ∞
(mn ) 2 3/2 √ −η −(Ec −EF )/kT
= 2 3
(kT ) ηe e dη.
π ~ 0

1
E 1 fMB fMB
conduction band
Ec gc (E) f
3 kT 0.8
f 0.1
0.6
Eg
0.4
0.01
0.2
Ev 0
valence band 0.001
-6 -4 -2 0 2 4 6 0 1 2 3 4 5 6
When the Fermi level is below Ec − 3kT , f (E ) ≈ fMB (E ), and

√ Z
(mn∗ )3/2 2 ∞ p
n = 2 3
E − Ec e −(E −EF )/kT dE
π ~ Ec
∗ 3/2
√ Z ∞
(mn ) 2 3/2 √ −η −(Ec −EF )/kT
= 2 3
(kT ) ηe e dη.
π ~ 0
√ √ √
∞
√ π (mn∗ )3/2 2 π −(Ec −EF )/kT
Z
η e −η dη = →n= 2 3
(kT )3/2 e = Nc e −(Ec −EF )/kT .
0 2 π ~ 2

Hole density (p) in equilibrium
E 1
conduction band
Ec gc (E)
0.8
1−f
f
0.6
Eg
0.4
0.2
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
(mp∗ )3/2
p
2 (Ev − E ) Ev
Z
gv (E ) = , E < Ev , p= gv (E ) (1 − f (E )) dE .
π 2 ~3 −∞

E 1
conduction band
Ec gc (E)
0.8
1−f
f
0.6
Eg
0.4
0.2
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
(mp∗ )3/2
p
2 (Ev − E ) Ev
Z
gv (E ) = , E < Ev , p= gv (E ) (1 − f (E )) dE .
π 2 ~3 −∞
(mp∗ )3/2
p
Z Ev 2(Ev − E )
p= dE
π 2 ~3 −∞ 1 + e −(E −EF )/kT

E 1
conduction band
Ec gc (E)
0.8
1−f
f
0.6
Eg
0.4
0.2
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
(mp∗ )3/2
p
2 (Ev − E ) Ev
Z
gv (E ) = , E < Ev , p= gv (E ) (1 − f (E )) dE .
π 2 ~3 −∞
(mp∗ )3/2
p
Z Ev 2(Ev − E )
p= dE
π 2 ~3 −∞ 1 + e −(E −EF )/kT
(mp∗ )3/2 √ ∞ η 1/2 Ev − E Ev − EF

Z
= 2 (kT )3/2 dη, with η = and ηv =
π 2 ~3 0 1 + e η−ηv kT kT

E 1
conduction band
Ec gc (E)
0.8
1−f
f
0.6
Eg
0.4
0.2
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
(mp∗ )3/2
p
2 (Ev − E ) Ev
Z
gv (E ) = , E < Ev , p= gv (E ) (1 − f (E )) dE .
π 2 ~3 −∞
(mp∗ )3/2
p
Z Ev
2(Ev − E )
p= dE
π 2 ~3−∞ 1 + e −(E −EF )/kT
(mp∗ )3/2 √ ∞ η 1/2 Ev − E Ev − EF

Z
= 2 3
2 (kT )3/2 dη, with η = and ηv =
π ~ 0 1 + e η−ηv kT kT
mp kT 3/2
∗
2

= Nv √ F1/2 (ηv ), where Nv = 2 is the effective density of states for the valence band.
π 2π~2
E 1
conduction band f
Ec gc (E)
0.8
1−f
0.6
Eg
0.4
e−(EF −E)/kT
0.2
3 kT
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
When EF > Ev + 3 kT , 1 − f (E ) can be approximated using the Maxwell-Boltzmann function.

1 − f (E ) ≈ e −(EF −E )/kT → p = Nv e −(EF −Ev )/kT .

E 1
conduction band f
Ec gc (E)
0.8
1−f
0.6
Eg
0.4
e−(EF −E)/kT
0.2
3 kT
Ev gv (E) 0
valence band
-6 -4 -2 0 2 4 6
(E − EF )/kT
When EF > Ev + 3 kT , 1 − f (E ) can be approximated using the Maxwell-Boltzmann function.

1 − f (E ) ≈ e −(EF −E )/kT → p = Nv e −(EF −Ev )/kT .
STOP

EF > Ec − 3 kT
(degenerate semiconductor) !
2 Ec − EF Ei
n = Nc √ F1/2 (ηc ), ηc = −
π kT Ev Ev + 3 kT Ec − 3 kT Ec
Ec p = N e−(EF −Ev )/kT
v
3 kT 1022
pMB nMB
Ev + 3 kT < EF < Ec − 3 kT 1018

(non-degenerate semiconductor)
pFermi
n,p (cm−3 )
nFermi
n = N e−(Ec −EF )/kT
c 10 14
p = Nv e−(EF −Ev )/kT
1010
3 kT T = 300 K
Ev EF < Ev + 3 kT
(degenerate semiconductor) 106
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
n = N e−(Ec −EF )/kT
c !
EF (eV)
2 EF − Ev Electron and hole densities in silicon
p = Nv √ F1/2 (ηv ), ηv = −
π kT

Ei
Ev Ev + 3 kT Ec − 3 kT Ec
1022 valence forbidden

forbidden gap conduction
nMB band
pMB gap band
1020 1020
1018 T = 300 K T = 300 K
pFermi
n,p (cm−3 )
nFermi 1019 1019

1014
p (cm−3 )
n (cm−3 )
1018 1018
1010
Ev Ec
T = 300 K
6 17 17
10 10 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 −0.1 −0.05 0 0.05 0.1 0.15 0.95 1 1.05 1.1 1.15 1.2
EF (eV) EF (eV) EF (eV)
Ec
3 kT
Eg
3 kT
Ev
Ei

nMB band
pMB gap band
1020 1020
1018 T = 300 K T = 300 K
pFermi
n,p (cm−3 )
nFermi 1019 1019

1014
p (cm−3 )
n (cm−3 )
1018 1018
1010
Ev Ec
T = 300 K
6 17 17
10 10 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 −0.1 −0.05 0 0.05 0.1 0.15 0.95 1 1.05 1.1 1.15 1.2
Ec
3 kT
2 Ec − EF

* nFermi = Nc √ F1/2 (ηc ), ηc = − * nMB = Nc e −(Ec −EF )/kT
π kT
Eg
3 kT
Ev
Ei

nMB band
pMB gap band
1020 1020
1018 T = 300 K T = 300 K
pFermi
n,p (cm−3 )
nFermi 1019 1019

1014
p (cm−3 )
n (cm−3 )
1018 1018
1010
Ev Ec
T = 300 K
6 17 17
10 10 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 −0.1 −0.05 0 0.05 0.1 0.15 0.95 1 1.05 1.1 1.15 1.2
Ec
3 kT
2 Ec − EF

π kT
2 F − Ev

E
Eg * p Fermi = Nv √ F1/2 (ηv ), ηv = − * p MB = Nv e −(EF −Ev )/kT
π kT
3 kT
Ev
Ei

nMB band
pMB gap band
1020 1020
1018 T = 300 K T = 300 K
pFermi
n,p (cm−3 )
nFermi 1019 1019

1014
p (cm−3 )
n (cm−3 )
1018 1018
1010
Ev Ec
T = 300 K
6 17 17
10 10 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 −0.1 −0.05 0 0.05 0.1 0.15 0.95 1 1.05 1.1 1.15 1.2
Ec
3 kT
2 Ec − EF

π kT
2 F − Ev

E
π kT
* For EF < (Ec − 3 kT ), nMB ≈ nFermi
3 kT
Ev
Ei

nMB band
pMB gap band
1020 1020
1018 T = 300 K T = 300 K
pFermi
n,p (cm−3 )
nFermi 1019 1019

1014
p (cm−3 )
n (cm−3 )
1018 1018
1010
Ev Ec
T = 300 K
6 17 17
10 10 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 −0.1 −0.05 0 0.05 0.1 0.15 0.95 1 1.05 1.1 1.15 1.2
Ec
3 kT
2 Ec − EF

π kT
2 F − Ev

E
π kT
* For EF < (Ec − 3 kT ), nMB ≈ nFermi
* For EF > (Ev + 3 kT ), p MB ≈ p Fermi
3 kT
Ev
Carrier statistics in equilibrium: Example
For silicon at T = 300 K and in equilibrium (with Nc = 2.8 × 1019 cm−3 ,

Nv = 1.04 × 1019 cm−3 , Eg = 1.12 eV),


Nv = 1.04 × 1019 cm−3 , Eg = 1.12 eV),
(a) Find EF for which n and p are equal. This Fermi level is called Ei , the
“intrinsic” Fermi level.


Nv = 1.04 × 1019 cm−3 , Eg = 1.12 eV),
(b) Obtain expressions for n and p in terms of (Ei − EF ) (instead of (Ec − EF )
and (EF − Ev )). Assume non-degenerate conditions.


Nv = 1.04 × 1019 cm−3 , Eg = 1.12 eV),
(b) Obtain expressions for n and p in terms of (Ei − EF ) (instead of (Ec − EF )
and (EF − Ev )). Assume non-degenerate conditions.
(c) Plot gc (E ) f (E ) and gv (E ) [1 − f (E )] versus E for
EF = Ei + 20 meV,
EF = Ei + 10 meV,
EF = Ei ,
EF = Ei − 10 meV,
EF = Ei − 20 meV.

Ei
22
10
pMB nMB
The condition n = p is satisfied when EF is about (Ev + Ec )/2.
1018
→ We can use MB statistics, i.e.,
pFermi
n,p (cm−3 )
nFermi
Nc e −(Ec −EF )/kT = Nv e −(EF −Ev )/kT 1014
1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
conduction band
Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
Nv
→ −Ec + EF + EF − Ev = kT log .
Nc 1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
conduction band
Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
Nv
Nc 1010
1 kT Nv T = 300 K
→ EF = (Ec + Ev ) + log .
2 2 Nc 10 6
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
conduction band
Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
Nv
Nc 1010
1 kT Nv T = 300 K
→ EF = (Ec + Ev ) + log .
2 2 Nc 10 6
mp∗ −0.2 0 0.2 0.4 0.6 0.8 1.0 1.2

1 3
Nv /Nc = (mp∗ /mn∗ )3/2 → EF ≡ Ei = (Ec + Ev ) + kT log ∗ . EF (eV)
2 4 mn
conduction band
Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
Nv
Nc 1010
1 kT Nv T = 300 K
→ EF = (Ec + Ev ) + log .
2 2 Nc 10 6
mp∗ −0.2 0 0.2 0.4 0.6 0.8 1.0 1.2

1 3
2 4 mn
conduction band
* The second term in the above equation is about −7.3 meV, Ec
i.e., the intrinsic Fermi level Ei is located 7.3 meV below the
centre of the band gap. Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
Nv
Nc 1010
1 kT Nv T = 300 K
→ EF = (Ec + Ev ) + log .
2 2 Nc 10 6
mp∗ −0.2 0 0.2 0.4 0.6 0.8 1.0 1.2

1 3
2 4 mn
conduction band
* The second term in the above equation is about −7.3 meV, Ec
i.e., the intrinsic Fermi level Ei is located 7.3 meV below the
centre of the band gap. Eg
Ei
* If Nc and Nv were equal, Ei would be exactly at the centre of
the band gap.
Ev
valence band

Ei
22
10
pMB nMB
When EF = Ei , we have n = p ≡ ni , the “intrinsic carrier
1018
concentration.”
pFermi
n,p (cm−3 )
The actual electron concentration (for a given Fermi level EF ) can nFermi
be written in terms of ni as follows. 1014
n = Nc e −(Ec −EF )/kT , ni = Nc e −(Ec −Ei )/kT .

1010
n
→ = e (EF −Ei )/kT → n = ni e (EF −Ei )/kT . T = 300 K
ni
6
Similarly, for the hole concentration p, we obtain 10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
p = ni e (Ei −EF )/kT .
conduction band
Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
concentration.”
pFermi
n,p (cm−3 )

1010
n
ni
6
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
conduction band
* for EF > Ei , n > p. Ec
Eg
Ei
Ev
valence band

Ei
22
10
pMB nMB
1018
concentration.”
pFermi
n,p (cm−3 )

1010
n
ni
6
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)
conduction band
* for EF > Ei , n > p. Ec
* For EF < Ei , n < p.
Eg
Ei
Ev
valence band

E
conduction band
Ec gc (E)
Eg
Ev gv (E)
valence band
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3
1
* f (E ) =
1 + e (E −EF )/kT
Z ∞
* n= gc (E )f (E )dE
Ec
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞
E
conduction band
Ec gc (E) gc (E) f(E)
Eg
Ev gv (E)
valence band
EF
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3
1
* f (E ) = gv (E) [1 − f(E)]
1 + e (E −EF )/kT
Z ∞
Ec EF = Ei
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞
E
conduction band
Eg
Ev gv (E)
valence band
EF EF
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3
1
* f (E ) = gv (E) [1 − f(E)]
1 + e (E −EF )/kT
Z ∞
Ec EF = Ei + 10 meV EF = Ei
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞
E
conduction band
Ec
Eg
Ev gv (E)
valence band EF EF EF
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3 Ev
1
* f (E ) = gv (E) [1 − f(E)]
1 + e (E −EF )/kT
Z ∞
Ec EF = Ei + 20 meV EF = Ei + 10 meV EF = Ei
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞
E
conduction band
Ec
Eg
Ev gv (E)
valence band EF EF EF EF
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3 Ev
1
* f (E ) = gv (E) [1 − f(E)]
1 + e (E −EF )/kT
Z ∞
Ec EF = Ei + 20 meV EF = Ei + 10 meV EF = Ei EF = Ei − 10 meV
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞
E
conduction band
Ec
Eg
Ev gv (E)
valence band EF EF EF EF EF
(mn∗ )3/2
p
2 (E − Ec )
* gc (E ) =
π 2 ~3
(mp∗ )3/2 2 (Ev − E )
p
* gv (E ) =
π 2 ~3 Ev
1
* f (E ) = gv (E) [1 − f(E)]
1 + e (E −EF )/kT
Z ∞
Ec EF = Ei + 20 meV EF = Ei + 10 meV EF = Ei EF = Ei − 10 meV EF = Ei − 20 meV
Z Ev
* p= gv (E ) (1 − f (E )) dE
−∞ M. B. Patil, IIT Bombay
Temperature dependence of ni
n = Nc e −(Ec −EF )/kT , p = Nv e −(EF −Ev )/kT .

When EF = Ei , n = p = ni , i.e.,
n = ni = Nc e −(Ec −Ei )/kT , p = ni = Nv e −(Ei −Ev )/kT


→ n p = ni2 = Nc Nv e −(Ec −Ev )/kT = Nc Nv e −Eg /kT



√
→ ni = Nc Nv e −Eg /2kT
kT 3/2 −Eg /2kT

= 2 (mn∗ mp∗ )3/4 e
2π~2
≡ K T 3/2 e −Eg /2kT
1020
K
1015
ni
1010
105
ni (cm−3 )
n = ni = Nc e −(Ec −Ei )/kT , p = ni = Nv e −(Ei −Ev )/kT T3/2
100
10−5
√
→ ni = Nc Nv e −Eg /2kT 10−10 e−Eg /2kT
3/2
kT
= 2 (mn∗ mp∗ )3/4 e −Eg /2kT 10−15
2π~2 200 300 400 500
T (K)
≡K T 3/2 e −Eg /2kT
1020
K
1015
ni
1010
105
ni (cm−3 )
100
10−5
√
3/2
kT
= 2 (mn∗ mp∗ )3/4 e −Eg /2kT 10−15
2π~2 200 300 400 500
T (K)
≡K T 3/2 e −Eg /2kT
T (◦ C) ni (cm−3 )
25 8.1 × 109
35 1.7 × 1010
45 3.5 × 1010
55 6.9 × 1010
65 1.3 × 1011
75 2.3 × 1011
1020
K
1015
ni
1010
105
ni (cm−3 )
100
10−5
√
3/2
kT
= 2 (mn∗ mp∗ )3/4 e −Eg /2kT 10−15
2π~2 200 300 400 500
T (K)
≡K T 3/2 e −Eg /2kT
T (◦ C) ni (cm−3 )
For silicon, near room temperature, ni nearly doubles
with every 10 ◦ C rise in temperature. 25 8.1 × 109
35 1.7 × 1010
45 3.5 × 1010
55 6.9 × 1010
65 1.3 × 1011
75 2.3 × 1011
How to obtain EF
Ei
1022
pMB nMB
1018
pFermi
n,p (cm−3 )
nFermi
14
10
1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)

How to obtain EF
Ei
1022
pMB nMB
* So far, we have assumed a certain EF (with respect to Ec
and Ev ) and obtained n and p.
1018
pFermi
n,p (cm−3 )
nFermi
14
10
1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)

How to obtain EF
Ei
1022
pMB nMB
1018
* In practice, we only have information such as Nc , Nv , Eg , T ,
and doping densities (Na and Nd ). pFermi
n,p (cm−3 )
nFermi
14
10
1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)

How to obtain EF
Ei
1022
pMB nMB
1018
* In practice, we only have information such as Nc , Nv , Eg , T ,
and doping densities (Na and Nd ). pFermi
n,p (cm−3 )
nFermi
* We now want to consider the reverse problem of finding EF 10 14
(and n, p), given the above data.
1010
T = 300 K
6
10
−0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
EF (eV)

Charge considerations in equilibrium: intrinsic semiconductor
conduction band
Ec
Eg
Ev
valence band
T=0K
* At 0 K, the positive charge due to the atomic cores balances the negative charge
due to valence electrons.
conduction band
Ec
Eg
Ev
valence band
T=0K
* As temperature increases, some of the valence electrons become free, i.e., they enter
the conduction band, leaving behind positively charged holes in the valence band.
conduction band
Ec
Eg
Ev
valence band
T=0K T = 300 K
conduction band
Ec
Eg
Ev
valence band
T=0K T = 300 K
* The number of electrons in the conduction band is equal to the number of holes in
the valence band.

conduction band
Ec
Eg
Ev
valence band
T=0K T = 300 K
* The number of electrons in the conduction band is equal to the number of holes in
the valence band.
* Also, their densities must be equal since the electrostatic potential is constant (no
electric field) → n = p.
Charge considerations in equilibrium: doped semiconductor
conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
* When there are donor or acceptor atoms in the lattice, we have the following charged species.

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
- electrons in the conduction band (density n)

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
- holes in the valence band (density p)

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
- ionised donor atoms (density Nd+ )

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
- ionised acceptor atoms (density Na− )

conduction band
Ec
Ed
Eg A
Ea D
Ev
valence band
- ionised acceptor atoms (density Na− )
* If the doping densities (Na or Nd or both) are uniform in space, charge neutrality in equilibrium requires
−qn + qp + qNd+ − qNa− = 0 → n + Na− = p + Nd+ .

Charge considerations in equilibrium
conduction band
Ec
Ed
Eg
Ea
Ev
valence band
conduction band
Ec
Ed
Eg
Ea
Ev
valence band
* Let the donor density be Nd . Some of the donor atoms donate their electrons and acquire a net positive
charge; the others remain neutral. → Nd = Nd+ + Nd0 .
conduction band
Ec
Ed
Eg
Ea
Ev
valence band
* Similarly, Na = Na− + Na0 .
conduction band
Ec
Ed
Eg
Ea
Ev
valence band
Nd+ 1 Na− 1
* The ratios Nd+ /Nd and Na− /Na are given by = , = .
Nd 1+ 2 e (EF −Ed )/kT Na 1 + 4 e (Ea −EF )/kT
Ed Ec
1
conduction band
Ec T = 50 K
0.8
Ed 100 K
0.6 200 K
d /Nd
Eg 300 K
N+
0.4
Ea
Ev 0.2
valence band
0
0.8 0.9 1 1.1 1.2
EF (eV)
Nd+ 1 Na− 1
Ed Ec
1
conduction band
Ec T = 50 K
0.8
Ed 100 K
0.6 200 K
d /Nd
Eg 300 K
N+
0.4
Ea
Ev 0.2
valence band
0
0.8 0.9 1 1.1 1.2
EF (eV)
Nd+ 1 Na− 1
* Nd+ /Nd → 1 if EF is sufficiently below Ed .
Ed Ec Ev Ea
1 1
conduction band
Ec T = 50 K T = 50 K
0.8 0.8
Ed 100 K 100 K
0.6 200 K 0.6 200 K
d /Nd
a /Na
Eg 300 K
300 K
N+
N−
0.4 0.4
Ea
Ev 0.2 0.2
valence band
0 0
0.8 0.9 1 1.1 1.2 -0.1 0 0.1 0.2 0.3
EF (eV) EF (eV)
Nd+ 1 Na− 1
Ed Ec Ev Ea
1 1
conduction band
Ec T = 50 K T = 50 K
0.8 0.8
Ed 100 K 100 K
0.6 200 K 0.6 200 K
d /Nd
a /Na
Eg 300 K
300 K
N+
N−
0.4 0.4
Ea
Ev 0.2 0.2
valence band
0 0
0.8 0.9 1 1.1 1.2 -0.1 0 0.1 0.2 0.3
EF (eV) EF (eV)
Nd+ 1 Na− 1
* Na− /Na → 1 if EF is sufficiently above Ea .
Calculation of n and p in equilibrium
conduction band
Ec
Ed
Eg
Ea
Ev
valence band

* n + Na− = p + Nd+ .
conduction band
Ec
Ed
Eg
Ea
Ev
valence band

* n + Na− = p + Nd+ .
conduction band Na Nd
Ec * Nc e −(Ec −EF )/kT + = Nv e −(EF −Ev )/kT + .
Ed 1 + 4 e (Ea −EF )/kT 1 + 2 e (EF −Ed )/kT
Eg
Ea
Ev
valence band

* n + Na− = p + Nd+ .
* We can take Ev as a reference → Ev = 0, Ec = Eg .
Eg
Ea
Ev
valence band

* n + Na− = p + Nd+ .
Eg
* This is a nonlinear equation in EF and must be solved iteratively.
Ea
Ev
valence band

* n + Na− = p + Nd+ .
Eg
* Note that Eg depends on the temperature. For silicon,
Ea
Ev Eg (T ) = Eg (0) − αT 2 /(β + T ),
valence band
with Eg (0) = 1.17 eV, α = 4.73 × 10−4 eV/K, and β = 636 K.

* n + Na− = p + Nd+ .
Eg
* Note that Eg depends on the temperature. For silicon,
Ea
Ev Eg (T ) = Eg (0) − αT 2 /(β + T ),
valence band
with Eg (0) = 1.17 eV, α = 4.73 × 10−4 eV/K, and β = 636 K.
* Let us look at the results obtained for a few representative values of Nd and Na , with
Ec − Ed = 45 meV, Ea − Ev = 45 meV.

n and p in equilibrium
Ec
* With Nd = 1015 cm−3 and Na = 0, 1 Ed Ec
- At room temperature (300 K),
Energy (eV)
EF
n ≈ Nd , and p n.
EF
Ea Ev
Ev
0
2
Nd = 1015 /cm3 Nd =0
Density (×1015 /cm3 )

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Energy (eV)
EF
n ≈ Nd , and p n.
- Since Nd+ + p = n, we have EF
Nd+ ≈ Nd , i.e., complete Ev
Ev Ea
ionisation of the donor atoms. 0
2
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Energy (eV)
EF
n ≈ Nd , and p n.
Ev Ea
2
* With Na = 1015 cm−3 and Nd = 0,
Nd = 1015 /cm3 Nd =0

- At room temperature (300 K), Na =0 Na = 1015 /cm3
p ≈ Na , and n p.
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Energy (eV)
EF
n ≈ Nd , and p n.
Ev Ea
2
* With Na = 1015 cm−3 and Nd = 0,
Nd = 1015 /cm3 Nd =0

p ≈ Na , and n p.
1
- Since Na− + n = p, we have n p
Na− ≈ Na , i.e., complete p n
ionisation of the acceptor atoms. ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Energy (eV)
EF
n ≈ Nd , and p n.
Ev Ea
2
* With Na = 1015 cm−3 and Nd = 0,
Nd = 1015 /cm3 Nd =0

p ≈ Na , and n p.
1
- Since Na− + n = p, we have n p
Na− ≈ Na , i.e., complete p n
ionisation of the acceptor atoms. ni ni
* In fact, the condition of complete 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
ionisation is valid over a wide range of T (K) T (K)
temperatures, called the “extrinsic”
temperature region.

Ec
* Note that one of the carrier densities 1 Ec
Ed
is much larger than the other (in the
Energy (eV)
extrinsic region). The more abundant EF
carrier is called the “majority carrier,”
and the other carrier is called the
EF
“minority carrier.” Ev
Ev Ea
0
2
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Ed
Energy (eV)
EF
Ev Ea
* For Nd = 1015 cm−3 and Na = 0, 0
electrons are the majority carriers. 2
For Na = 1015 cm−3 and Nd = 0, Nd = 1015 /cm3 Nd =0

holes are the majority carriers. Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
Ed
Energy (eV)
EF
Ev Ea
* For Nd = 1015 cm−3 and Na = 0, 0
electrons are the majority carriers. 2
For Na = 1015 cm−3 and Nd = 0, Nd = 1015 /cm3 Nd =0

holes are the majority carriers. Na =0 Na = 1015 /cm3
* A semiconductor with electrons as 1 n p
majority carriers is called an n-type
semiconductor. p n
A semiconductor with holes as ni ni
majority carriers is called a p-type 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
semiconductor.
T (K) T (K)

Ec
1 Ed Ec
* At low temperatures, a significant
Energy (eV)
EF
fraction of impurity atoms remains
neutral, causing a reduction in n or p.
EF
Ea Ev
Ev
0
2
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
1 Ed Ec
Energy (eV)
EF
EF
* The carriers remain “frozen” at the
Ea Ev
impurity sites, i.e., electrons remain Ev
0
bound to donors, and holes (vacancies)
2
remain bound to acceptors.
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
1 Ed Ec
Energy (eV)
EF
EF
* The carriers remain “frozen” at the
Ea Ev
impurity sites, i.e., electrons remain Ev
0
bound to donors, and holes (vacancies)
2
remain bound to acceptors.
Nd = 1015 /cm3 Nd =0

* This effect is called the carrier Na =0 Na = 1015 /cm3
“freeze-out” effect, and it can be a
limiting factor in low-temperature 1 n p
operation of semiconductor devices.
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
1 Ed Ec
* At high temperatures, the intrinsic
Energy (eV)
EF
carrier concentration ni becomes large
and starts dominating.
EF
Ea Ev
Ev
0
2
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
1 Ed Ec
Energy (eV)
EF
EF
* As a result, n and p become
Ea Ev
comparable (and larger than Nd or Ev
0
Na ).
2
Nd = 1015 /cm3 Nd =0

Na =0 Na = 1015 /cm3
1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec
1 Ed Ec
Energy (eV)
EF
EF
* As a result, n and p become
Ea Ev
comparable (and larger than Nd or Ev
0
Na ).
2
* This region is called the “intrinsic Nd = 1015 /cm3 Nd =0

region,” and it must be avoided for a Na =0 Na = 1015 /cm3
semiconductor device to work as
intended. 1 n p
p n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

n and p in equilibrium: compensated semiconductor
Ec Ec
1 Ed Ed
* If both types of dopants are present, EF
Energy (eV)
the dopant with the larger density
dominates.
EF
Ea Ev Ea Ev
0
2

Nd = 1015 /cm3 Nd = 0.2 × 1015 /cm3
15
Na = 0.2 × 10 /cm 3
Na = 1015 /cm3
1
n p p
n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec Ec
1 Ed Ed
Energy (eV)
dominates.
EF
* If Nd Na , the semiconductor is
Ea Ev Ea Ev
n-type, and n = Nd − Na (in the
0
extrinsic temperature range).
2

Nd = 1015 /cm3 Nd = 0.2 × 1015 /cm3
15
Na = 0.2 × 10 /cm 3
Na = 1015 /cm3
1
n p p
n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec Ec
1 Ed Ed
Energy (eV)
dominates.
EF
Ea Ev Ea Ev
0
2
* If Na Nd , the semiconductor is

Nd = 1015 /cm3 Nd = 0.2 × 1015 /cm3
p-type, and p = Na − Nd (in the 15
Na = 0.2 × 10 /cm 3
Na = 1015 /cm3
1
n p p
n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Ec Ec
1 Ed Ed
Energy (eV)
dominates.
EF
Ea Ev Ea Ev
0
2
* If Na Nd , the semiconductor is

Nd = 1015 /cm3 Nd = 0.2 × 1015 /cm3
p-type, and p = Na − Nd (in the 15
Na = 0.2 × 10 /cm 3
Na = 1015 /cm3
1
* A semiconductor with both types of
dopants is called a “compensated” n p p
semiconductor. n
ni ni
0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
T (K) T (K)

Computation of n and p at room temperature in equilibrium

* We have seen that Nd+ ≈ Nd and Na− ≈ Na at room temperature. This makes the
computation of n and p much easier.

* Charge neutrality → n + Na− = p + Nd+ → n + Na ≈ p + Nd .

* Also, assuming non-degenerate conditions, we have
n p = Nc e −(Ec −EF )/kT × Nv e −(EF −Ev )/kT
= Nc Nv e −(Ec −Ev )/kT
= Nc Nv e −Eg /kT
= ni2 (T ).

* Also, assuming non-degenerate conditions, we have
n p = Nc e −(Ec −EF )/kT × Nv e −(EF −Ev )/kT
= Nc Nv e −(Ec −Ev )/kT
= Nc Nv e −Eg /kT
= ni2 (T ).
* The above two equations can be solved to obtain n and p.

Computation of n and p at room temperature in equilibrium: example
In a silicon sample with Nd = 5 × 1016 cm−3 , find the equilibrium electron and hole
concentrations at T = 300 K (ni = 1010 cm−3 at 300 K).

Solution: n + Na− = p + Nd+ → n ≈ p + Nd

ni2
Using np = ni2 , i.e., p = , we get
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2

ni2
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2
+ sign gives a physically meaningful result, viz.,
ni2 1020
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016

ni2
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2
ni2 1020
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016
Since n p, this is an n-type sample.

ni2
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2
ni2 1020
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016
Calculation of EF : n = Nc e −(Ec −EF )/kT

ni2
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2
ni2 1020
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016
Nc 2.8 × 1019
→ Ec − EF = kT log = (0.0259 eV) log = 0.1636 eV.
n 5 × 1016

ni2
n q
n2 Nd ± Nd2 + 4ni2
n = i + Nd → n2 − nNd − ni2 = 0 → n = .
n 2
ni2 1020
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016
Nc 2.8 × 1019
n 5 × 1016
(Solving the charge neutrality equation exactly gives n = 4.9 × 1016 cm−3 ,
Ec − EF = 0.1641 eV.)

ni2
n q Ec
n2 Nd ± Nd2 + 4ni2 Ed
n = i + Nd → n2 − nNd − ni2 = 0 → n = . EF
n 2
ni2 1020 Eg
n ≈ Nd = 5 × 1016 cm−3 , p = = = 2 × 103 cm−3 .
n 5 × 1016
Calculation of EF : n = Nc e −(Ec −EF )/kT Ev
Nc 2.8 × 1019
n 5 × 1016
(Solving the charge neutrality equation exactly gives n = 4.9 × 1016 cm−3 ,
Ec − EF = 0.1641 eV.)

Variation of EF with doping density (silicon, 300 K)
1 Ec 1 Ec
T = 300 K
EF
Energy (eV)
EF
T = 300 K
Ev Ev
0 0
1014 1015 1016 1017 1014 1015 1016 1017
Nd (/cm3 ) Na (/cm3 )

1 Ec 1 Ec
T = 300 K
EF
Energy (eV)
EF
T = 300 K
Ev Ev
0 0
1014 1015 1016 1017 1014 1015 1016 1017
Nd (/cm3 ) Na (/cm3 )
* n-type semiconductor: EF gets closer to Ec as Nd is increased.

1 Ec 1 Ec
T = 300 K
EF
Energy (eV)
EF
T = 300 K
Ev Ev
0 0
1014 1015 1016 1017 1014 1015 1016 1017
Nd (/cm3 ) Na (/cm3 )

* p-type semiconductor: EF gets closer to Ev as Na is increased.

1 Ec 1 Ec
T = 300 K
EF
Energy (eV)
EF
T = 300 K
Ev Ev
0 0
1014 1015 1016 1017 1014 1015 1016 1017
Nd (/cm3 ) Na (/cm3 )

Nc
* n ≈ Nd = Nc e −(Ec −EF )/kT → Ec − EF = kT log ≈ kT (log Nc − log Nd ).
n
→ (Ec − EF ) varies linearly with log Nd .

1 Ec 1 Ec
T = 300 K
EF
Energy (eV)
EF
T = 300 K
Ev Ev
0 0
1014 1015 1016 1017 1014 1015 1016 1017
Nd (/cm3 ) Na (/cm3 )

Nc
* n ≈ Nd = Nc e −(Ec −EF )/kT → Ec − EF = kT log ≈ kT (log Nc − log Nd ).
n
→ (Ec − EF ) varies linearly with log Nd .
Nv
* p ≈ Na = Nv e −(EF −Ev )/kT → EF − Ev = kT log ≈ kT (log Nv − log Na ).
p
→ (EF − Ev ) varies linearly with log Na .

SD Carrier 1 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SD Carrier 1 PDF

Uploaded by

Copyright:

Available Formats

Semiconductor Devices

Department of Electrical Engineering

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

* The electron density depends on two factors: E

M. B. Patil, IIT Bombay

* The electron density depends on two factors: E

M. B. Patil, IIT Bombay

* The electron density depends on two factors: E

M. B. Patil, IIT Bombay

* The electron density depends on two factors: E

M. B. Patil, IIT Bombay

* The electron density depends on two factors: E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

Calculate the number of states N 0 between Ec and Ec + 50 meV for silicon at E

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

* The number of electrons in the interval E to (E + dE ) is not gc (E )dE but gc (E )f (E )dE .

M. B. Patil, IIT Bombay

* The number of electrons in the interval E to (E + dE ) is not gc (E )dE but gc (E )f (E )dE .

M. B. Patil, IIT Bombay

* The number of electrons in the interval E to (E + dE ) is not gc (E )dE but gc (E )f (E )dE .

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

* Low T : e (E −EF )/kT varies rapidly with E .

M. B. Patil, IIT Bombay

* Low T : e (E −EF )/kT varies rapidly with E .

M. B. Patil, IIT Bombay

* Low T : e (E −EF )/kT varies rapidly with E .

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

When the Fermi level is below Ec − 3kT , f (E ) ≈ fMB (E ), and

M. B. Patil, IIT Bombay

When the Fermi level is below Ec − 3kT , f (E ) ≈ fMB (E ), and

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay

M. B. Patil, IIT Bombay