Chapter 4 PN and Metal-Semiconductor Junctions

4.1   Building Blocks of the PN Junction Theory

– V +
I
Donor ions

N P
N-type
P-type I
diode
symbol

V
Reverse bias Forward bias

PN junction is present in perhaps every semiconductor device.

Slide 4-1
Chapter 4 PN and Metal-Semiconductor Junctions

4.1   Building Blocks of the PN Junction Theory

– V +
I
Donor ions

N P
N-type
P-type I
diode
symbol

V
Reverse bias Forward bias

PN junction is present in perhaps every semiconductor device.

Slide 4-2
 4.1.1   Energy Band Diagram of a PN Junction
N-region P-region
Ef is constant at
(a) Ef
equilibrium
Ec

Ec Ef
Ec and Ev are known
(b)
Ev relative to Ef
Ev
Ec

(c) Ef Ec and Ev are smooth,

Ev the exact shape to be
determined.
Neutral Depletion Neutral
N-region layer P-region
A depletion layer
Ec
exists at the PN
(d) Ef junction where n  0
Ev and p  0.

Slide 4-3
Slide 4-4
 4.1.2 Built-in Potential

(a)
N-type P-type
NNd d NNa a

Ec
q bi
(b)
Ef
Ev

V
bi
(c)

xN 0 xP x

Can the built-in potential be measured with a voltmeter?

Slide 4-5
 4.1.2 Built-in Potential

 q A kT kT N c
N-region n  N d  Nce  A ln
q Nd
2
ni  q B kT kT N c N a
P-region n   Nce B ln 2
Na q ni

kT  N c N a Nc 
bi  B  A   ln  ln 

q  ni
2
Nd 


kT N d N a
bi  ln 2
q ni

Slide 4-6
 4.1.3 Poisson’s Equation
Gauss’s Law:
a A
ar e

E (x) E (x + x)
s: permittivity (~12o for Si) 
: charge density (C/cm3)

x

x

Poisson’s equation

Slide 4-7
 4.2 Depletion-Layer Model
(a)
N
4.2.1
N
Field andN Potential
P
in the Depletion Layer
d a

(b) Neutral Region Depletion Layer Neutral Region

On the P-side of the
N P depletion layer,  = –qNa
xnN 0 xpP
 dE  qN a
dx s
qNd
(c) xpP qN a qN a
x E ( x )   x  C  ( x  x)
xnN
–qNa s 1
s P

 E On the N-side,  = qNd

(d) qN d
E ( x)  ( x - xN )
x
s
xnN 0 xpP
V
bi

(e) Slide 4-8

(a)
N
4.2.1 Field Nd Potential in theNDepletion
and P
Layer
a

(b) N eut ra l Re gion D eple tion La yer N e utral R egi on

N P
–xnN 0 xpP

The electric field is continuous at x = 0.

qN
Nda |xP| = Nd|xP|
(c) xp
of the junction is depleted more? x
Which side –x
n
–qN a

A one-sided junction is called a N+P junction or P+N junction

 E

Slide 4-9
 qNd
(c) xp
4.2.1x Field–qNand Potential
n
x
in the Depletion Layer
a

 E On the P-side,
qN a
V ( x)  ( xP  x ) 2
2 s
(d)

xn 0 xp x
N P

bi
V
Arbitrarily choose the
(e) voltage at x = xP as V = 0.
x
xnN xpP
On the N-side,
Ec
qN d
 bi , built-in potential V ( x)  D  ( x  xN )2
(f)
Ef 2 s
Ev
qN d
 bi  ( x  xN )2
2 s

Slide 4-10
 (a)
N Nd Na P
4.2.2 Depletion-Layer Width

(b) N eut ra l Re gion D eple tion La yer N e utral R egi on

N P
–xnN 0 xpP
V is continuous at x = 0 
2 sbi  1 1 
xP  xN  Wdep    
q  Na Nd 
qN
If Na >> Nd , as in a P+N junction,
d
(c) xp
2 sbi x
Wdep 
qN d
–xxn N
–qN a
|x P||xN|N d Na  0

What about a N+P junction?  E

1 1 1 1
Wdep  2 s bi qN where   
N N d N a lighter dopant density
(d)
Slide 4-11
–xn 0 xp x
EXAMPLE: A P+N junction has Na=1020 cm-3 and Nd
=1017cm-3. What is a) its built in potential, b)Wdep , c)xN ,
and d) xP ?

Solution:
kT N d N a 10 20  1017 cm 6
a) bi  ln 2
 0.026V ln 20 6
1V
q ni 10 cm
1/ 2
2 sbi  2 12  8.85 10 1 
14

b) Wdep    19
  0.12 μm
 1.6 10 10
17
qN d 

xN  Wdep  0.12 μm
c)
xP  xN N d N a  1.2  104 μm  1.2 Å  0
d)

Slide 4-12
Slide 4-13
Slide 4-14
Slide 4-15
Slide 4-16
Slide 4-17
Slide 4-18
Slide 4-19
Volume charger density, electric field, and electrostatic potential
of a PN junction as a function of bias

Slide 4-20
 4.3 Reverse-Biased PN Junction
V
+ –

N P

Ec 2 s (bi  | Vr |) 2 s  potential barrier

qbi
Wdep  
qN qN
Ec Ef
Ef Ev

Ev
(a) V = 0 1 1 1 1
  
Ec
N N d N a lighter dopant density
qbi + qV Efp
Ev
• Does the depletion layer
Ec qV
Efn
widen or shrink with
increasing reverse bias?
Ev
(b) reverse-biased

Slide 4-21
 4.4 Capacitance-Voltage Characteristics

N Nd Na P

Conductor Insulator Conductor

Wde p

Reverse biased PN junction is a s

capacitor. (Modeled as a parallel-
Cdep A
Wdep
plate capacitor with capacitance.)

• Is Cdep a good thing?

• How to minimize junction capacitance?

Slide 4-22
 4.4 Capacitance-Voltage Characteristics

1/C dep 2 Capacitance data

2
1 Wdep 2(bi  V )
 
Cdep
2
A 2 s
2
qN S A2

Slope = 2/qN sA2

Vr
– bi Increasing reverse bias

• From this C-V data can Na and Nd be determined?

Slide 4-23
 EXAMPLE: If the slope of the line in the previous slide is
2x1023 F-2 V-1, the intercept is 0.84V, and A is 1 m2, find the
lighter and heavier doping concentrations Nl and Nh .

Solution:
N l  2 /( slope  q s A2 )
 2 /(2 10 23  1.6 10 19  12  8.85 10 14  10 8 cm 2 )
 6  1015 cm 3

q 2 0.84
kT N h N l ni kTbi 10 20 0.026
bi  ln 2
 N h  e  e  1 .8  1018
cm 3

q ni Nl 6  1015

Slide 4-24
 4.5 Junction Breakdown
I

Forward Current
V B, breakdown
voltage
V
R
Small leakage
Current
A
P N

R C
A
3.7 V
IC
Zener diode
B
D

A Zener diode is designed to operate in the breakdown mode.

Slide 4-25
 4.5.1 Peak Electric Field

Neutral Region
increasing
reverse bias
N+ Na P
0 xp
1/ 2
(a)  2qN 
E
Ep  E(0)    (bi  | Vr |)
Ep  s 
increasing reverse bias

 sEcrit 2
x VB   bi
xp 2qN
(b)

Slide 4-26
 4.5.2 Tunneling Breakdown

Dominant if both sides of

a junction are very heavily
doped.
Filled States - Empty States
Ec

Ev H / εp
JGe
I

V
  6
Ep Ecrit 10 V/cm
Breakdown

Slide 4-27
 4.5.3 Avalanche Breakdown
• impact ionization: an
Ec
energetic electron generating
original
electron electron and hole, which can
Efp also cause impact ionization.
Ev
• Impact ionization + positive
feedbackavalanche breakdown
sEcrit 2
electron-hole VB 
pair generation 2qN

Ec 1 1 1
Efn VB   
N Na Nd

Slide 4-28
Slide 4-29
 4.6 Forward Bias – Carrier Injection

VV=0
=0 Forward
Forward biased
biased
V
I=0 – +

Ec N P
 qbi
- Ec
Ef qbi –qV
Ev
Ef n qV Efp
Ev

Drift and diffusion cancel out +

Minority carrier injection

Slide 4-30
 4.6 Forward Bias –
Quasi-equilibrium Boundary Condition
 ( Ec  E fn ) / kT  ( Ec  E fp ) / kT ( E fn  E fp ) / kT
n ( xP )  N c e  Nce e
( E fn  E fp ) / kT
 nP 0 e  nP 0 e qV / kT

EEcc

EEfnfnfn • The minority carrier

EEfpfp densities are raised
EEv v by eqV/kT
• Which side gets more
xx carrier injection?
xN0 x0 P
N P

Slide 4-31
4.6 Carrier Injection Under Forward Bias–
Quasi-equilibrium Boundary Condition

2
ni qV
n(xP )  nP 0 e 
qV kT kT
e
Na
2
ni qV
p (xP)  p N 0 e 
qV kT kT
e
Nd

n( xP )  n( xP )  nP 0  nP 0 (e qV kT
 1)
p( x N )  p ( x N )  p N 0  p N 0 (e qV kT
 1)

Slide 4-32
 EXAMPLE: Carrier Injection
A PN junction has Na=1019cm-3 and Nd=1016cm-3. The applied
voltage is 0.6 V.

Question: What are the minority carrier concentrations at the

depletion-region edges?

Solution: n ( xP )  nP 0e  10  e0.6 0.026  1011 cm -3

qV kT

p( x N )  p N 0 e qV kT  10 4  e 0.6 0.026  1014 cm -3

Question: What are the excess minority carrier concentrations?

n( xP )  n( xP )  nP 0  1011  10  1011 cm -3
Solution:
p( x N )  p( x N )  p N 0  1014  104  1014 cm -3

Slide 4-33
 4.7 Current Continuity Equation

J p ( x) J p ( x  x) p
A  A  A  x 
q q 
aA
a re
Jp (x) Jp ( x +  x )
J p ( x  x)  J p ( x) p
 q
p x 

Volume = A·x
dJ pp
 q
dx 
x

x

Slide 4-34
 4.7 Current Continuity Equation
dJ p p
 q Minority drift current is
dx  
negligible;
d2p p Jp= –qDpdp/dx
qD p 2  q
dx p

d p
2
p p d 2 n n
  2 2
 2
dx 2
D p p L p dx Ln

Lp and Ln are the diffusion lengths

L p  D p p Ln  Dn n

Slide 4-35
4.8 Forward Biased Junction-- Excess Carriers

P + N
d 2 p p
2
 2
xP dx Lp
-x N
0
x p()  0
p( x N )  p N 0 (e qV / kT  1)
x / Lp  x / Lp

p ( x)  Ae  Be
  x  xN  / L p
p( x)  p N 0 (e qV / kT  1)e , x  xN

Slide 4-36
4.8 Excess Carrier Distributions

1.0

P-side N-side
N a = 1017 cm -3 Nd = 2 1017 cm-3
0.5
nP ' ex /L n pN '  e–x /L p

– 3L n –2L n –Ln 0 L p 2L p 3L p 4Lp

 x  xN  / L p

p ( x )  p N 0 (e qV / kT
 1)e , x  xN
n( x )  nP 0 (e qV / kT  1)e  x  xP  / Ln , x  xP
Slide 4-37
 EXAMPLE: Carrier Distribution in Forward-biased PN Diode

N-type P-type
Nd = 5  cm-3 Na = 101 7 cm -3
Dp =12 cm 2/s Dn=36.4 cm2 /s
p = 1  s n = 2  s

• Sketch n'(x) on the P-side.

2
ni 1020 0.6 0.026
n( xP )  nP 0 (e qV kT
 1)  (e qV / kT
 1)  17 e  1013 cm 3
Na 10
1013cm-3
N-side P-side
n’ ( = p’ )


p´ ( = n’ ) 2
x
Slide 4-38
EXAMPLE: Carrier Distribution in Forward-biased PN Diode

• How does Ln compare with a typical device size?

Ln  Dn n  36  2 10 6  85 μm

• What is p'(x) on the P- side?

Slide 4-39
 4.9 PN Diode I-V Characteristics
Jtotal
Jtotal Jp = Jtotal – Jn Jn = Jtotal – Jp

JnP JnP
JpN JpN
x
P-side 0 N-side P-side 0 N-side

dp( x) Dp  x  xN  Lp
J pN   qD p q p N 0 (e qV kT
 1)e
dx Lp

dn( x) D
J nP  qDn  q n nP 0 (e qV kT
 1)e  x  xP  Ln

dx Ln
 Dp Dn  qV

Total current  J pN ( x N )  J nP ( xP )  q p q n P 0 ( e kT
 1)
 L N0 Ln 
 p 
 J at all x
Slide 4-40
Slide 4-41
 The PN Junction as a Temperature Sensor

I  I 0 (e qV kT
 1)

 Dp D 
I 0  Aqni  
2
 n
L N 
 p d Ln N a 

What causes the IV curves to shift to lower V at higher T ?

Slide 4-42
 4.9.1   Contributions from the Depletion Region
n  p  ni e qV / 2 kT
Net recombination (generation) rate :
ni qV / 2 kT
(e  1)
 dep

qniWdep
I  I 0 (e qV / kT
 1)  A (e qV / 2 kT  1)
τ dep

Space-Charge Region (SCR) current

qniWdep
I leakage  I 0  A
τ dep Under forward bias, SCR current is an extra
current with a slope 120mV/decade

Slide 4-43
 4.10 Charge Storage

1013 cm -3 QI
N-side P-side

n' I  Q s
Q  I s
p’ 2
x

What is the relationship between s (charge-storage time)

and  (carrier lifetime)?

Slide 4-44
 4.11 Small-signal Model of the Diode
I
1 dI d d
G   I 0 (e qV / kT
 1)  I 0 e qV / kT
R dV dV dV
V R C
q kT
 I 0 (e qV / kT )  I DC /
kT q
What is G at 300K and IDC = 1 mA?

Diffusion Capacitance:
dQ dI kT
C s   sG   s I DC /
dV dV q

Which is larger, diffusion or depletion capacitance?

Slide 4-45
 Bn Increases with Increasing Metal Work Function

Vacuum level, E0

Si = 4.05 eV  M : Work Function

 qM of metal

qBn  Si : Electron Affinity of Si

Ec

Ef

Theoretically,
Ev Bn=M – Si

Slide 4-46
 4.16 Schottky Barriers
Energy Band Diagram of Schottky Contact

Depletion
Metal layer Neutral region

qBn
Ec
Ef
N-Si • Schottky barrier height, B ,
Ev is a function of the metal
Ec
material.

P-Si • B is the most important

Ef
parameter. The sum of qBn
qBp Ev
and qBp is equal to Eg .

Slide 4-47
 Schottky barrier heights for electrons and holes

Metal Mg Ti Cr W Mo Pd Au Pt
f Bn (V) 0.4 0.5 0.61 0.67 0.68 0.77 0.8 0.9
f Bp (V) 0.61 0.5 0.42 0.3
Work
Function 3.7 4.3 4.5 4.6 4.6 5.1 5.1 5.7
y m (V)

Bn + Bp  Eg

Bn increases with increasing metal work function

Slide 4-48
 Fermi Level Pinning

Vacuum level, E0 • A high density of

energy states in the
Si = 4.05 eV
 q bandgap at the metal-
M semiconductor interface
pins Ef to a narrow
qBn Ec
range and Bn is
+ 
Ef typically 0.4 to 0.9 V

• Question: What is the

Ev
typical range of Bp?

Slide 4-49
 Schottky Contacts of Metal Silicide on Si

Silicide: A silicon and metal compound. It is conductive

similar to a metal.

Silicide-Si interfaces are more stable than metal-silicon

interfaces. After metal is deposited on Si, an annealing step is
applied to form a silicide-Si contact. The term metal-silicon
contact includes and almost always means silicide-Si contacts.

Silicide ErSi1.7 HfSi MoSi2 ZrSi2 TiSi2 CoSi2 WSi2 NiSi2 Pd2Si PtSi
BnBn (V) 0.28 0.45 0.55 0.55 0.61 0.65 0.67 0.67 0.75 0.87
BpBp (V) 0.55 0.49 0.45 0.45 0.43 0.43 0.35 0.23

Slide 4-50
 Using C-V Data to Determine B

qbi qbi  q Bn  ( Ec  E f )
qBn
Ec
Nc
Ef  q Bn  kT ln
Nd
Ev 2 s (bi  V )
Wdep 
qN d
qBn q(bi + V) s
C A
qV Wdep
Ec
Ef
Question:
How should we plot the CV
Ev data to extract bi?

Slide 4-51
 Using CV Data to Determine B

1/C2
1 2(bi  V )

C 2
qN d  s A2

qBn qbi
V Ec
bi Ef

Once bi is known, can

E
be determined using v

Nc
qbi  qBn  ( Ec  E f )  qBn  kT ln
Nd
Slide 4-52
 4.17 Thermionic Emission Theory
vthx
-
q( B  V) Ec
q B
N-type Efm qV Efn
V Metal Silicon

Ev

x
3/ 2
 2mn kT 
n  N c e  q ( B V ) / kT  2 2  e  q ( B V ) / kT
 h 
vth  3kT / mn vthx   2kT / mn
1 4qmn k 2 2  q B / kT qV / kT
J S M   qnvthx  3
T e e
2 h
 J 0 e qV / kT , where J o  100e  q B / kT A/cm2

Slide 4-53
4.18 Schottky Diodes

Forward
biased
V=0

V
Reverse bias Forward bias
Reverse
biased

Slide 4-54
 4.18 Schottky Diodes

I 0  AKT 2 e  q B / kT
4qmn k 2
K 3
 100 A/(cm 2
 K 2
)
h
I  I S  M  I M  S  I 0 e qV / kT  I 0  I 0 (e qV / kT  1)
Slide 4-55
 4.19 Applications of Schottly Diodes
I I Schottky diode I  I 0 (e qV / kT  1)
I 0  AKT 2 e  q B / kT
BB PN junction
PN junction
diode

V
V

• I0 of a Schottky diode is 103 to 108 times larger than a PN

junction diode, depending on B . A larger I0 means a smaller
forward drop V.
• A Schottky diode is the preferred rectifier in low voltage,
high current applications.
Slide 4-56
 Switching Power Supply

PN Junction Schottky
rectifier Transformer rectifier
100kHz
110V/220V Hi-voltage Hi-voltage Lo-voltage 50A
DC MOSFET AC AC 1V DC
AC inverter
utility
power

feedback to modulate the pulse width to keep Vout = 1V

Slide 4-57
 4.19 Applications of Schottky diodes

Question: What sets the lower limit in a Schottky diode’s

forward drop?

• Synchronous Rectifier: For an even lower forward drop,

replace the diode with a wide-W MOSFET which is not
bound by the tradeoff between diode V and leakage current.

• There is no minority carrier injection at the Schottky

junction. Therefore, Schottky diodes can operate at higher
frequencies than PN junction diodes.

Slide 4-58
4.20   Quantum Mechanical Tunneling

Tunneling probability:

8 2 m
P  exp( 2T 2
(VH  E ) )
h

Slide 4-59
4.21 Ohmic Contacts

Slide 4-60
 4.21 Ohmic Contacts
Silicide N+ Si
2 s Bn
Wdep  Bn
Bn – V
qN d - -
- - Ec , Ef Efm V
Ec , Ef

Tunneling
probability: Ev
 H Bn Nd Ev
Pe x x

T  Wdep / 2   sBn / 2 qN d
4
H   s mn / q
h
1  H ( Bn V ) /
kT / 2mn e
Nd
J S M  qN d vthx P  qN d
2

Slide 4-61
 4.21 Ohmic Contacts

1 H Bn / N d
 dJ S  M  2e H Bn / Nd
Rc     e Ω  cm 2
 dV  qvthx H N d

Slide 4-62
 4.22 Chapter Summary

Part I: PN Junction

kT N d N a The potential barrier

bi  ln 2 increases by 1 V if a 1 V
q ni
reverse bias is applied

2 s  potential barrier
depletion width Wdep 
qN

junction capacitance s
Cdep A
Wdep

Slide 4-63
 4.22 Chapter Summary

• Under forward bias, minority carriers are injected

across the jucntion.

• The quasi-equilibrium boundary condition of

minority carrier densities is:
n( x p )  nP 0 e qV kT

p ( x N )  p N 0 e qV kT

• Most of the minority carriers are injected into the

more lightly doped side.

Slide 4-64
 4.22 Chapter Summary

• Steady-state • Minority carriers

continuity equation: diffuse outward  e–|x|/Lp
and e–|x|/Ln
d 2 p p p • Lp and Ln are the
  2
dx 2
D p p L p diffusion lengths

I  I 0 (e qV kT
 1)
L p  D p p
 Dp Dn 
I 0  Aqni
2
 
L N 
 p d Ln N a 

Slide 4-65
 4.22 Chapter Summary

Charge storage: Q  I s

Diffusion capacitance: C   sG

kT
Diode conductance: G  I DC /
q

Slide 4-66
 4.22 Chapter Summary

Part III: Metal-Semiconductor Junction

I 0  AKT 2 e  qB / kT
•Schottky diodes have large reverse saturation current, determined by the
Schottky barrier height B, and therefore lower forward voltage at a given
current density.

•Ohmic contacts relies on tunneling. Low resistance contact requires

low B and higher doping concentration.

4
(  B  s mn / qN d )
Rc  e h
Ω  cm 2

Slide 4-67
 Bn Increases with Increasing Metal Work Function

Vacuum level, E0

Si = 4.05 eV
 q
M
Ideally,
Bn=M – Si
q  Bn Ec

Ef

Ev

Slide 4-68