CHAP #03 - PART 1

PN Junctions in Thermal Equilibrium

In this lecture you will learn:

• Junctions of P and N doped semiconductors

• Electrostatics of PN junctions in thermal equilibrium

• Built-in junction potential, junction electric field, depletion regions, and

all that…

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

Review: Potential of a Doped Semiconductor

What are the values of potentials in N-doped and P-doped semiconductors ??

N-doped Semiconductors (doping density is Nd ):

The potential in n-doped semiconductors is denoted by: n

no  x   Nd Example:
q n  x  Suppose,
 Nd  ni e KT Nd  1017 cm- 3 and ni  1010 cm-3
KT N  KT N 
 n  log d   n  log d    0.41 Volts
q  ni  q  ni 

P-doped Semiconductors (doping density is Na ):

n-doped semiconductors is denoted by: p
The potential in p

po  x   Na Example:
q p  x  Suppose,

 Na  ni e KT Na  1017 cm- 3 and ni  1010 cm-3
KT N  KT N 
 p   log a   p   log a    0.41 Volts
q  ni  q  ni 
ECE 315 – Spring 2005 – Farhan Rana – Cornell University

1
 Can One Measure the Potential Difference
between Doped Semiconductors Directly?

P-doped N-doped
p n
Na Nd

Answer: No! K có dòng chuyển dời của hạt mang điện

For the explanation, you need to wait till the end of this handout…….

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

A Junction of P and N Doped Semiconductors

What do we know (before the junction is formed):

P-doped N-doped
Na Nd

Doping: Doping:
Na Nd

Carrier densities: Carrier densities:

p po  N a nno  Nd

ni2 ni2
n po  pno 
Na Nd

Potential: Potential:
KT N  KT N 
  p   log a    n  log d 
q  ni  q  ni 

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

2
Metal Step 1: Diffusion of Electron and Holes
contact
P-doped N-doped
Na Nd

0 x
Carrier diffusion
Metal wire

Holes are more in number on the P-side, so they will diffuse into the N-side from the
P-side

--- On reaching the N-side, the hole will find an electron and recombine with it

Electrons are more in number on the N-side, so they will diffuse into the P-side from
the N-side

--- On reaching the P-side, the electron will find a hole and recombine with it

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

Step 2: Creation of “Depletion” Regions

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

0 x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qN d

Both the N-doped and P-doped materials were charge neutral before the junction was
formed

As the holes diffuse from the P-side into the N-side, they leave behind negatively
charged acceptor atoms in a region near the interface on the P-side which becomes
“depleted” of holes

As the electrons diffuse from the N-side into the P-side, they leave behind positively
charged donor atoms in a region near the interface on the N-side which becomes
“depleted” of electrons

As the carrier keep diffusing, the widths of the depletion regions keep increasing ……..

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

3
 Diffusion Currents

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

0 x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qN d

P-doped J pdiff N-doped

Na J ndiff Nd

0 x

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

Generation of Electric Field

Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

0 x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qNd

The charge densities in the depletion regions generate an electric field

As the depletion regions grow in thickness, the magnitude of the electric field also
increases

Question: How do we figure out how big this electric field is?........

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

4
 Generation of Electric Field
Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
 x po 0 x no
x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qNd

Assume that the thickness of the depletion region on the P-side is: x po

Assume that the thickness of the depletion region on the N-side is: x no

Gauss’s Law on the P-side: Gauss’s Law on the N-side:

dE

qN a
s
 E  x po   0 dE

qNd
 E  x no   0
dx dx s
 E x   
qNa
s
x  x po   E x   
qNd
 x no  x 
s

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

Generation of Electric Field

Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
 x po 0 x no
x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qN d

Field on the P-side: Field on the N-side:

E x   
qN a
s
x  x po  E x   
qNd
 x no  x 
E x  s

 x po 0 x
x no

Field varies linearly with distance

The field must be continuous at the junction: qN a x po  qNd x no

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

5
 Drift Currents
Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
 x po 0 x no
x

As the carriers diffuse, and the depletion regions grow in thickness, and the electric
field increases, the drift currents due to the electric field also become non-zero…………

P-doped J pdrift N-doped

Na J ndrift Nd
 x po 0 x no
x

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

Step 3: Establishment of Equilibrium

As the carriers diffuse, and the depletion regions grow in thickness, and the electric
field increases, the drift currents due to the electric field also increase …………and as
the depletion region width increases, the diffusion currents decrease………

A stage is reached when the carrier diffusion currents are exactly balanced by the
(oppositely directed) drift currents:

P-doped J pdiff N-doped

Na J ndiff Nd
 x po 0 x no x

P-doped J pdrift N-doped

Na J ndrift Nd
 x po 0 x no
x

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

6
 Step 3: Establishment of Equilibrium
P-doped J pdiff N-doped
Na J ndiff Nd
 x po 0 x no x

P-doped J pdrift N-doped

Na J ndrift Nd
 x po 0 x no
x

In thermal equilibrium:
J p  x   J pdrift  x   J pdiff  x   0 J n  x   J ndrift  x   J ndiff  x   0

---- The electron diffusion current is balanced by the equal and opposite electron drift
current
---- The hole diffusion current is balanced by the equal and opposite hole drift
current

So the net currents of both the electrons as well as the holes go to zero …. and
equilibrium is established!!

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

A PN Junction in Equilibrium
Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
 x po 0 x no
x
Region depleted of holes Region depleted of electrons

Charge density:  qN a Charge density:  qNd

In equilibrium, there is no NET current (of either electrons or the holes)

But, in equilibrium, there is an electric field…….right at the junction!

E x 

 x po 0 x
x no

A non-zero electric field implies a spatially varying electrostatic potential:

d  x 
E x   
dx
ECE 315 – Spring 2005 – Farhan Rana – Cornell University

7
 A PN Junction in Equilibrium
Electric field

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
 x po 0 x no
x
A non-zero electric field implies an electrostatic potential:
E x 

 x po 0 x
x no
Điện thế là
d  x  một loại thế năng
E x   
dx

Potential on the P-side: Potential on the N-side:

KT N  KT N 
  p   log a    n  log d 
q  ni  q  ni 

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

A PN Junction in Equilibrium: Electrostatic Potential

Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x
A non-zero electric field implies an electrostatic potential:
E x 

 x po 0 x
E x   
qN a
s
x  x po  x no
E x   
qNd
 x no  x 
s

d  x  d  x 
 E  x   E  x 
dx dx
  x   p 
qN a
2 s
x  x po 2    x   n 
qNd
2 s
 x no  x 2

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

8
 A PN Junction in Equilibrium: Electrostatic Potential
Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x

 x   p 
qN a
2 s
x  x po 2 n 
KT N 
log d 
q  ni 
 x 
n

 x po 0 x no x
p
KT N 
p   log a 
q  ni  qNd
  x   n   x no  x 2
2 s
Potential varies quadratically with distance

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

The Built-In Potential

Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x

 x 
n

B
 x po 0 x no x
p
B  n   p 0,7 với Si

Built-In Junction Potential: Example:

B  n   p Nd  10171/cm3 N a  10171/cm3
N N  B  n   p
KT
 log d 2 a 
q KT N N 
 ni   log d 2 a   0.83 Volts
q  ni 
ECE 315 – Spring 2005 – Farhan Rana – Cornell University

9
 A PN Junction in Equilibrium: Depletion Region Widths
Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x

The electric field and the potential must be continuous at the junction

Field and Potential on the P-side: Field and Potential on the N-side:

E x   
qN a
s
x  x po  E x   
qNd
 x no  x 
s

 x   p 
qN a
2 s
x  x po 2   x   n 
qNd
2 s
 x no  x 2
qN a x po  qNd x no

qN a
p 
2 s
x po 2  n  qN d x
2 s
no 
2

These two equations can be solved to find xpo and xno

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

A PN Junction in Equilibrium: Depletion Region Widths

Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x

The result is:

2  s B  Nd  2  s B  Na 
x po    x no   
qN a  N a  Nd  qNd  N a  Nd 

Observations:

i) If Nd >> Na:
2  s B 2  s B  Na  In an asymmetrically
x po   x no    doped junction, the
qN a qNd  Nd  depletion region on
ii) If Na >> Nd: the lightly doped
2  s B  Nd  2  s B side is larger
x po     x no 
qN a  Na  qNd
ECE 315 – Spring 2005 – Farhan Rana – Cornell University

10
 A PN Junction in Equilibrium: A Summary
Electric field

P-doped - - - - + + + + N-doped
  p - - - - + + + +   n
Na - - - - + + + + Nd
 x po 0 x no
x

E  x  Electric Field

 x po 0 x
x no

2 q B  N d N a 
Emax   
 s  N a  Nd 
Potential
 x 
n
B

 x po 0 x no x
p
B  n   p

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

A PN Junction Process
P-dopant
(Ion implantation)

P-doped

A Si wafer (N-doped) A Si wafer (N-doped)

Metal Metal Metal

P-doped P-doped

A Si wafer (N-doped) A Si wafer (N-doped)

Depletion region

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

11
 Can One Measure the Potential Difference
between Doped Semiconductors Directly?

P-doped M M n

p
N-doped

Metals also have a

potential: M

There is contact potential at the metal-semiconductor interface

Contact potential is just the built-in potential at the metal-semiconductor interface

After accounting for the contact-potential, the potential difference measured by the
voltmeter is zero!

ECE 315 – Spring 2005 – Farhan Rana – Cornell University

12
 CHAP #03 - PART 2

Biased PN Junction Diodes and Current Flow

In this lecture you will learn:

• Biased PN junction diodes (forward biased and reverse biased PN diodes)

• Depletion capacitance of PN junction diodes
• Minority and majority carrier distributions in a biased PN junction diodes
• Carrier transport and current flow in biased PN junction diodes

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Review: A PN Junction Diode in Thermal Equilibrium

• You have already seen a PN Junction diode in thermal equilibrium:

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 x 
n

B
 x po 0 x no x
p
B  n   p
E x 

 x po 0 x
x no
2  s B  Nd  2  s B  Na 
x po   N  N  x no   
qN a  a d  qNd  N a  Nd 
• In thermal equilibrium no net current flows in either left or right direction

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

1
 Drift and Diffusion Currents in Thermal Equilibrium
P-doped J pdiff N-doped
Na J ndiff Nd
 x po 0 x no x

P-doped J pdrift N-doped

Na J ndrift Nd
 x po 0 x no
x

In thermal equilibrium:

● The electron diffusion current is balanced by the equal and opposite electron drift
current
● The hole diffusion current is balanced by the equal and opposite hole drift current

So the net currents of both the electrons as well as the holes are zero!

In thermal equilibrium:
J p  x   J pdrift  x   J pdiff  x   0 J n  x   J ndrift  x   J ndiff  x   0

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Some Thermal Equilibrium Relations

P-doped J ndrift N-doped
Na J ndiff Nd
 x po 0 x no x

• Total electron current is zero in thermal equilibrium:

d n x 
J n  x   q n x  n E  x   q Dn 0
dx
In thermal equilibrium, these two components balance each other exactly at every
point in space so that there is no total electron current anywhere

d n x 
 q n x  n E  x   q Dn
dx
d  x  d n x 
  n x  n   Dn
dx dx
n d   x  d logn x 
 
Dn dx dx Mật độ e
q  x 
q d   x  d logn x 
   n x   ni e KT
KT dx dx
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

2
 Some Thermal Equilibrium Relations
P-doped J pdrift N-doped
Na J pdiff Nd
 x po 0 x no x

• Total hole current is zero in thermal equilibrium:

d p x 
J p  x   q p  x   p E  x   q Dp 0
dx
In thermal equilibrium, these two components balance each other exactly at every
point in space so that there is no total hole current anywhere
d p x 
 q p  x   p E  x   q Dp
dx
d  x  d p x 
  p x   p  Dp
dx dx
 p d   x  d logp x 
  
Dp dx dx Mật độ p
q  x 
q d   x  d logp x  
    p x   ni e KT
KT dx dx
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Carrier Concentrations in Thermal Equilibrium

q  x 
n  x   ni e KT

q  x 

p  x   ni e KT

Another way to write the same equations is:

q   x 2   x1 
nx 2   nx1  e KT

q   x 2   x1 

p x 2   p x1  e KT

In thermal equilibrium, electron and hole concentrations at different points are

related exponentially to the potential difference at these points

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

3
 A PN Junction Diode in Thermal Equilibrium
• Minority carrier concentrations at the edges of the depletion region:
 x po x no
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 x 
n

B
 x po 0 x no x
p

q p q 
q  p  n  q q p 
q n  p 
ni2 ni2
 
n
 n  
n  x po   ni e KT  ni e KT e KT
p x no    ni e KT  ni e KT e KT
Na Nd

q  p n  
q  n  p 
 n x no  e KT 
 p  x po e  
KT

qB q B
 
 Nd e KT
 Na e KT

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Voltage Drops in Resistive Networks

V1 V2 V3

R1 R2 R3

Most of the voltage drops across the largest resistor in series

Suppose:
R2  R1
R2  R3

Then:
R2
V2  V V
R1  R2  R3

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

4
 A Forward Biased PN Junction Diode
• Now apply a forward bias with an external voltage source VD:

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD > 0
All the applied bias is assumed
=> Chênh lệch thế năng giảm
to fall across the depletion region
and reduces the built-in potential   x  n
across it by VD  xp  V
B B D
0 xn x
p

2  s B  VD   Nd  E x  2  s B  VD   N a 
xp    xn   
qN a  N a  Nd  qNd  N a  Nd 

 xp 0 x
x n Vùng hiếm nhỏ lại
 x po x no
• In forward bias, the depletion regions shrink, and the electric field in the junction
also decreases in magnitude
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

A Reversed Biased PN Junction Diode

• Now apply a bias with an external voltage source VD (where VD < 0):

P-doped - - - - - + + + ++ N-doped
- - - - - + + + ++
Na - - - - - + + + ++ Nd
+ -
VD < 0
All the applied bias is assumed
to fall across the depletion region
and increases the built-in  x  n
potential across it by -VD  x p
B
0 xn  B  VD x
p

2  s B  VD   Nd  E x  2  s B  VD   N a 
xp    xn   
qN a  N a  Nd  qNd  N a  Nd 

 xp 0 x
xn
 x po x no
• In reverse bias, the depletion regions become larger, and the electric field in the
junction also increases in magnitude
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

5
 Junction Depletion Region Capacitance
A PN junction in equilibrium:
 x po x no
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD = 0

A PN junction in forward bias (depletion region shrinks):

P-doped - - - + + + N-doped
- - - + + +
Na - - - + + + Nd
+ - Có tính chất như tụ điện
VD > 0
 xp xn

A PN junction in reverse bias (depletion region widens):

P-doped - - - - - + + + ++ N-doped
- - - - - + + + ++
Na - - - - - + + + ++ Nd
+ -
VD < 0
 xp xn
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Junction Depletion Region Capacitance

Area = A
 xp xn
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD
 x p  x p  x n  x n

P-doped - - - + + + N-doped
- - - + + +
Na - - - + + + Nd
+ -

VD +  VD

Q j  Added charge density

Q j  qN a x p

 Q j  qNd x n
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

6
 Junction Depletion Region Capacitance
Area = A
 xp xn
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD
 x p  x p  x n  x n

P-doped - - - + + + N-doped
- - - + + +
Na - - - + + + Nd
+ -

+ VD +  VD
-
+ -
+ -
+ -
VD +  VD
Junction Depletion Capacitance:
Điện dung của mối nối
Q j dQ j d  qN a x p A  d  qNd x n A  sA
Cj     
 VD dVD dVD dVD x p  x n 
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Junction Depletion Region Capacitance

Area
 xp xn
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD

Since:

2  s B  VD   N a  Nd 
x p  xn   N N 
q  a d 

The depletion region capacitance becomes: Zero voltage junction

capacitance

sA q s  N a Nd  C jo
Cj   
x p  x n   A 
2 B  VD   N a  Nd  V
1 D
B

The depletion capacitance is mostly contributed

by the side with the lower doping that has the
larger depletion region thickness
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

7
 Junction Breakdown in Reverse Bias
 xp xn
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD < 0

In the presence of very large electric fields, the generation rate G of electrons and
holes can increase dramatically from the equilibrium value GoKhi điện trường tăng, e và holes
được gia tốc và phá vỡ cấu trúc PN
In the presence of large electric fields, electrons and holes accelerate to large
velocities and then give off their kinetic energies to create more electrons and holes
(impact ionization) which in turn accelerate and create even more electrons and
holes leading to an avalanche effect and resulting in very large currents

The minimum electric field at which breakdown occurs is called the “breakdown
field”

In Silicon, the breakdown field is around 3x105 V/cm

PN diodes exhibit breakdown under large reverse biases - when the maximum field in
the junction becomes equal to the breakdown field value

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Circuit Symbol for a PN Junction Diode

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD

+ -
VD

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

8
 Current Flow in a Forward Biased PN Junction Diode
A PN junction in equilibrium:
 x po x no
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD = 0
A PN junction in forward bias (junction field decreases and depletion region shrinks):

P-doped - - - + + + N-doped
- - - + + +
Na - - - + + + Nd
+ -
VD > 0
 xp xn

E x 
2 q B  VD   N a Nd 
Emax   
  N a  Nd 

 xp 0 x
xn
 x po x no

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Bias -> Thay đổi điện trường

Current Flow in a Forward Biased PN Junction Diode: Balance

of Drift Diffusion Currents is Broken
Drift current of both electrons and holes is reduced in forward bias (because the
electric field is reduced)

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD
Diffusion>Drift
Diffusion current of both electrons and holes in approximately the same as in
equilibrium

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD
• Consequently, the total electron and hole currents are no longer zero; diffusion
exceeds drift!!

• There is net current due to electron diffusion flow from the N-side to the P-side, and
due to hole diffusion flow from the P-side to the N-side

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

9
 Carrier Injection in a Forward Biased PN Junction Diode

• Minority carrier concentrations at the edges of the depletion region in forward

bias:
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD

 x 

 B  VD
 xp 0 xn x

q B VD  q B VD 


n  x p   Nd e

p x n   Na e
KT KT

qVD q VD
ni2 ni2
 e KT  e KT
Na Nd
In forward bias, the minority carrier concentrations increase exponentially at the
edges of the depletion region
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Fundamental Assumptions In Modeling Carrier Transport

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD
• In forward bias, the main obstacle to current flow is not the depletion region but
carrier diffusion in the N- and the P-sides

This assumption implies that it is enough to study current flow in the N- and P-sides
and not worry about what happens inside the depletion region

• The N-side and the P-side are “quasi-neutral”

The word “quasi-neutral” implies that there is almost no net charge densities inside
these regions and, by Gauss’s law, almost zero electric fields inside these regions

This assumption is not 100% accurate. The physical reason behind this assumption
is that when a material is highly conducting (like the N- and P-sides) the electric field
inside it is usually small. How small is small…….see the next slide….

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

10
 Modeling Minority Carrier Diffusion (N-side)
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD
Consider the N-side first:
• The holes diffuse from the P-side, cross the depletion region, and enter the N-side

• On the N-side, the holes are the minority carriers

• The dynamics of holes on the N-side in steady state are described by the Shockley
0
equations:
 p x  1 J p  x 
GR
t q x
 p x   p x 
0

J p  x   q p  x   p E  x   q Dp  q Dp
x x
• The electric fields in the quasi-neutral regions are assumed to be small enough
that they may be neglected in modeling minority carrier transport. Therefore,
minority carriers flow by diffusion (not drift).

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (N-side)

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD

1 J p  x   2 p x 
0 GR  G  R  Dp
q x x 2
• Let the total hole density on the N-side be written as:

p x   pno  p'  x  pno 

ni2
Nd
Equilibrium hole density Excess hole density
p'  x 
•Then the generation-recombination term becomes: GR  
p
•And we get:

 2 p'  x  p'  x  Diffusion equation for

 0 the excess hole density
x 2 Dp p

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

11
 Modeling Minority Carrier Diffusion (N-side)
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD

 2 p'  x  p'  x 
• We need to solve the second order diffusion equation:  0
x 2 Dp p
with the boundary condition:
ni2
q VD
ni2 ni2  q VD 
p'  x n   p  x n   pno  e KT    e KT  1
Nd Nd N d  
 
• But it is a second order differential equation, so we need a second boundary
condition!

The minority carrier lifetime p is assumed to be zero at the metal contacts.

• Consequently, there cannot be any excess hole density at the right metal
contact. This gives us the second boundary condition:

p' Wn  x n   0
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (N-side)

We need to solve:  q VD
ni2 
p'  x n    e KT  1
 2 p'  x  p'  x   Nd 
 0  
x 2 Dp p p' Wn  x n   0

Define a “minority carrier diffusion length” Lp for holes as: Lp  Dp p

We need to solve:  q VD 
ni2
 e KT  1
p'  x n  
 p'  x  p'  x 
2  Nd 
 2 0  
x 2 Lp p' Wn  x n   0

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

12
 Modeling Minority Carrier Diffusion (N-side)

We need to solve:  q VD 
ni2
 e KT  1
p'  x n  
 2 p'  x  p'  x   Nd 
 2 0  
x 2 Lp p' Wn  x n   0
And the solution is:  W  xn  x   W  xn  x 
sinh n  sinh n 
 Lp  2  qVD   Lp 
  
p' x  p' x n    
ni  KT
e 1   
 Wn  Nd    Wn 
sinh    sinh 
 Lp   Lp 
   
p'  x 

 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

The Minority Carrier Diffusion Length (N-Side)

The minority carrier diffusion length Lp is the average length a hole injected into
the N-side will diffuse before it finds an electron and recombines with it

• Long Base Limit:

If Lp << Wn then pretty much all the holes injected into the N-side recombine with
electrons before they are able to cross the N-side

• Short Base Limit:

If Lp >> Wn then pretty much all the holes injected into the N-side do not recombine
with the electrons and are able to cross the N-side

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

13
 Modeling Minority Carrier Diffusion (N-side)

Long Base Limit:

If Lp << Wn then the solution is:

 W  xn  x 
sinh n  x  x 
ni2  qVD   Lp  n 2  qVD   L n
p'  x    e KT  1    i  e KT  1e p
Nd    Wn  Nd  
  sinh   
 Lp 
 

p'  x 

 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (N-side)

Short Base Limit:

If Lp >> Wn then the solution is:
 W  xn  x 
sinh n 
 qVD
ni2    n 2  qVD 
p'  x    e KT  1  Lp   i  e KT  1 Wn  x n  x 
Nd   W  Nd   Wn 

  sinh n   
 Lp 
 
Could have just solved:
 2 p'  x 
0
x 2
p'  x 

 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

14
 Charge Neutrality and Majority Carrier Distribution (N-side)

• The assumption of quasi-neutrality implies that charge neutrality in the

conducting N-doped region holds even when out of equilibrium

• Let the excess majority carrier distribution (i.e. of electrons on the N-side) be
written as:
n x   nno  n'  x  nno  Nd
Equilibrium electron density Excess electron density

• Then charge neutrality implies: n'  x   p'  x 

n'  x 
p'  x 
 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (P-side)

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD
Consider the P-side now:
• The electrons diffuse from the N-side, cross the depletion region, and enter the P-side

• On the P-side, the electrons are the minority carriers

• The dynamics of electrons on the P-side in steady state are described by the Shockley
0
equations:
 n x  1 J n  x 
GR
t q x
0
 n x   n x 
J n  x   q n x  n E  x   q Dn  q Dn
x x
• The electric fields in the quasi-neutral regions are assumed to be small enough
that they may be neglected in modeling minority carrier transport. Therefore,
minority carriers flow by diffusion (not drift).

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

15
 Modeling Minority Carrier Diffusion (P-side)
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD
1 J n  x   2n x 
0 GR  G  R  Dn
q x x 2
• Let the total electron density on the P-side be written as:

ni2
n x   npo  n'  x  npo 
Na
Equilibrium electron density Excess electron density
n'  x 
•Then the generation-recombination term becomes: GR  
n
•And we get:

 2 n'  x  n'  x  Diffusion equation for

 0 the excess electron
x 2 Dn n density

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (P-side)

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD

 2 n'  x  n'  x 
• We need to solve the second order diffusion equation:  0
 q VD  x
2
Dn n
ni2
with the boundary condition: n'  x p    Na
 e KT

 1

 
• But it is a second order differential equation, so we need a second boundary
condition!

The minority carrier lifetime n is assumed to be zero at the metal contacts.

• Consequently, there cannot be any excess electron density at the left metal
contact. This gives us the second boundary condition:

n'  Wp  x p   0

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

16
 Modeling Minority Carrier Diffusion (P-side)

We need to solve:  q VD 
ni2
n'  x p    e KT  1
 2 n'  x  n'  x   Na 
 0  
x 2 Dn n n'  Wp  x p   0

Define a “minority carrier diffusion length” Ln for electrons as: Ln  Dn n

We need to solve:  q VD 
n2
 n'  x  n'  x 
2 n'  x p   i  e KT  1
 2 0 Na  
x 2  
n'  W p  x p   0
Ln

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Modeling Minority Carrier Diffusion (P-side)

We need to solve:  q VD 
ni2
 n'  x  n'  x 
2 n'  x p    e KT  1
 2 0  Na 
x 2  
n'  W p  x p   0
Ln

And the solution is:  Wp  x p  x   Wp  x p  x 

sinh  2  qVD  sinh 
n'  x   n'  x p     
Ln n   Ln
 i e KT 1
 Wp  Na    Wp 
sinh    sinh 
 Ln   Ln 

n'  x 

 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

17
 The Minority Carrier Diffusion Length (P-Side)

The minority carrier diffusion length Ln is the average length an electron injected
into the P-side will diffuse before it finds a hole and recombines with it

• Long Base Limit:

If Ln << Wp then pretty much all the electrons injected into the P-side recombine
with holes before they are able to cross the P-side

• Short Base Limit:

If Ln >> Wp then pretty much all the electrons injected into the P-side do not
recombine with the holes and are able to cross the P-side

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Charge Neutrality and Majority Carrier Distribution (P-side)

• The assumption of quasi-neutrality implies that charge neutrality pretty much

holds even when out of equilibrium
• Let the excess majority carrier distribution (i.e. of holes on the P-side) be written
as:
p x   ppo  p'  x  ppo  Na
Equilibrium hole density Excess hole density

• Then charge neutrality implies: p'  x   n'  x 

n'  x 
p'  x 
 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

18
 Minority Carrier Current Flow

n'  x  p'  x 

 Wp  x p  xp 0 xn Wn  x n x

Electron current on the P-side: Hole current on the N-side:

   p'  x 
Jn  x   J ndiff  x   q Dn  n' x J p  x   J pdiff  x   q Dp
x W  x  x x W  x  x 
 p p   n n 
cosh
 qVD  cosh 
D p 
qVD   Lp 
J n  x   qni2
Dn  e KT  1  Ln  J p  x   qni2 e KT  1  
N a Ln    Wp  N d Lp   W 
  sinh    sinh n 
 Ln  L 
 p

Jn  x  Jp x 

 Wp  x p  xp 0 xn Wn  x n
x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Total Current Flow

• The total current in steady state is the sum of electron and hole currents and is
independent of position:
JT  J n  x   J p  x 
• So we can compute the total current in forward bias if we know the total
electron current (drift and diffusion components) and the total hole current at any
one location in the device – wherever that location might be.

• Assumption: the minority carrier diffusion currents inside the depletion region
are constant (valid if there is no recombination in the depletion region), i.e.:
 
Jn  x p  Jn  x n  Jp xn   Jp  x p  
Jn  x  Jp x 

 Wp  x p  xp 0 xn Wn  x n x

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

19
 Total Current Flow
• Now using: JT  J n  x   J p  x 
with “x” anywhere inside the depletion region, we can calculate the total current

JT

Jn  x  Jp x 

 Wp  x p  xp 0 xn Wn  x n x
 qVD 
J n  x p   qni2
Dn  e KT  1 coth W p 
  L 
N a Ln
   n 
D p  
qVD
W 
J p  x n   qni2 e KT  1 coth n 
N d Lp   L 
   p

JT  J n  x n   J p  x n   J n  x p   J p  x n 

 W   
qVD
 D  Wp  Dp
JT  qni2  n coth   coth n   e KT  1
 N a Ln  L   
  Ln  Nd Lp  p   
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Majority Carrier Current Flow

• So far we have assumed that minority carrier current is entirely due to diffusion
• With the above assumption, we were able to calculate the total current

JT

Jn  x  Jp x 

 Wp  x p  xp 0 xn Wn  x n x
• Now we want to see how the majority carriers contribute to the total current
• Since we already know the total current everywhere, and the minority carrier
current everywhere, the difference must be the majority carrier current

JT

Jn  x  Jp x 
 Wp  x p  xp 0 xn Wn  x n x
• The majority carriers flow by both drift and diffusion

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

20
 Area “A” Current Flow in a PN Junction Diode

ID

 W   
qVD
 D  Wp  Dp
JT  qni2  n coth   coth n   e KT  1
 N a Ln     
  Ln  Nd Lp  Lp   
ID

 qVD 
I D  AJT  Io  e KT  1
 
 
 D  Wp  Dp W 
Io  qni2 A n coth   coth n  
 N a Ln  L 
  n 
L N L
d p  p 

VON VD

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Carrier Flow Diagram in Steady State Forward Bias

Recombination

P-doped N-doped
- - - - + + + +
- - - - + + + +
- - - - + + + +

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD

Holes (majority carriers) flowing via drift and diffusion

Holes (minority carriers) flowing via diffusion

Electrons (minority carriers) flowing via diffusion (không có drift)

Electrons (majority carriers) flowing via drift and diffusion

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

21
 Current Flow in a Reverse Biased PN Junction Diode: Balance
of Drift Diffusion Currents is Broken
Drift current of both electrons and holes increases in reverse bias (because the
junction electric field increases)

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD<0
Diffusion current of both electrons and holes in approximately the same as in
equilibrium
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD<0
• Consequently, the total electron and hole currents are no longer zero; drift exceeds
diffusion!!

• There is net current due to electron drift flow from the P-side to the N-side, and due
to hole drift flow from the N-side to the P-side

• But the P-side dos not have very many electrons and the N-side does not have very
many holes!!
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

A Reverse Biased PN Junction Diode

• Minority carrier concentrations at the edges of the depletion region in reverse

bias:
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ -
VD

 x 

 B  VD
 xp 0 xn x

q B VD  q B VD 

 
n  x p   Nd e KT p x n   Na e KT

q VD
n2
qVD ni2
 i e KT  e KT
Na Nd
In reverse bias, the minority carrier concentrations decrease exponentially at the
edges of the depletion region
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

22
 Excess Minority Carrier Distributions in Reverse Bias

n'  x  p'  x 

 Wp  x p  xp 0 xn Wn  x n
x

Same equations, as in the forward bias case, work in the reverse bias:

 Wp  x p  x   W  xn  x 
sinh n 
ni2  qVD  sinh   qVD
ni2   Lp 
 e KT  1  Ln  p'  x    e KT  1  
n'  x  
Na    Wp  Nd    Wn 
  sinh    sinh 
 Ln   Lp 
 

Jn  x  Jp x 

 Wp  x p 0 x
 xp xn Wn  x n

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Total Current Flow in Reverse Bias

• Now using: JT  J n  x   J p  x 
with “x” anywhere inside the depletion region, we can calculate the total current

Jp x 
Jn  x 
x

 Wp  x p JT Wn  x n
 xp 0 xn

 qVD 
J n  x p   qni2
Dn  e KT  1 coth W p 
  L 
N a Ln
   n  D p 
qVD  W 
J p  x n   qni2 e KT  1 coth n 
N d Lp   L 
   p

JT  J n  x n   J p  x n   J n  x p   J p  x n 

 W   
qVD
 D  Wp  Dp
JT  qni2  n coth   coth n   e KT  1
 N a Ln  L   
  Ln  Nd Lp  p   
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

23
 Current Flow in a PN Junction Diode
Area “A”

 W   
qVD
 D  Wp  Dp
JT  qni2  n coth   coth n   e KT  1
 N a Ln     
  Ln  Nd Lp  Lp   
ID
 qVD 
I D  AJT  Io  e KT  1
 
 
 D  Wp  Dp W 
Io  qni2 A n coth   coth n  
 N a Ln  L 
  n 
L N L
d p  p 

VON VD
Reverse bias
Junction breakdown Forward bias

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Carrier Flow Diagram in Steady State Reverse Bias

Generation

P-doped N-doped
- - - - + + + +
- - - - + + + +
- - - - + + + +

 Wp  x p  xp 0 xn Wn  x n x
+ -
VD

Holes (majority carriers) flowing via drift and diffusion

Holes (minority carriers) flowing via diffusion

Electrons (minority carriers) flowing via diffusion

Electrons (majority carriers) flowing via drift and diffusion

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

24
 CHAP #03 - PART 3

Large and Small Signal Modelling of PN Junction Diodes

In this lecture you will learn:

• Circuit models of PN junction diodes

• Small signal models of PN junction diodes
• Junction resistance and capacitances
• Light emitting PN junction diodes (LEDs)

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Current Flow in a PN Junction Diode

Area “A”

 W  x n   
qVD
 D  Wp  x p  Dp
JT  qni2  n coth   coth n   e KT  1
 N a Ln     
  Ln  N d Lp  Lp   
ID

 qVD 
I D  AJT  Io  e KT  1
 
 
 D  Wp  x p  Dp  W  xn  
Io  qni2 A n coth   coth n 
 N a Ln  L 
  Ln  N d Lp  p 

VON VD
Reverse bias
Junction breakdown Forward bias

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

1
 Simplest Circuit Model for a PN Junction Diode
VD
+ -

R For most diodes:

0.4  VON  0.8
V
Load Line: ID ID

V  I D R VD V Slope  
V  VD R
 ID 
R

VON V VD

Solution for current:

If V ≤ VON: If V > VON:

VD  V VD  VON
V  VON
ID  0 ID 
R
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Better Circuit Model a PN Junction Diode

VD
ID
ID Slope
dI D
 gd
dVD
->Tăng tuyến tính

VON VD

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

2
 Circuit Example for a PN Junction Diode
VD
+ -

V
ID
Load Line: ID

V  I D R VD V Slope
V  VD R dI D
 gd
 ID 
R dVD

VON V VD
Solution for current:
If V ≤ VON: If V > VON:
V gd R ->Thành phần tuyến tính
VD  V VD   VON
1  gd R 1  gd R
ID  0 gd
ID  V  VON 
1  gd R
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Small Signal Model of a PN Junction Diode: Junction Conductance

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
ID + id + - + -
VD vd

 qVD 
I D  Io  e KT  1
 
 
 q VD v d   I
 I D  id  Io  e KT  1  I D  D v d  .........  I D  gd v d
  VD
 
 id  gd v d (tính hiệu nhỏ, f thấp)

q I D  Io  qI D
qVD
1 I qI
gd   D  oe KT  
rd VD KT KT KT
In strong forward
bias
Differential resistance
Differential conductance

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

3
 Small Signal Model of a PN Junction Diode: Junction Conductance

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ - + -
VD vd
ID + id

I D  i d  I D  gd v d
 id  gd v d

1
rd 
gd

vd
id

Small signal circuit model of a PN diode

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Small Signal Model of a PN Junction Diode: Junction Depletion

Capacitance  xp 0 xn
P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ - + -
VD vd
ID + id

At high frequencies, part of the current id flows through the junction but part of it
also charges up the junction capacitance
dv d sA
Cj 
id  gd v d C j
dt x p  x n 
rd

Cj

vd
id
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

4
 Small Signal Model of a PN Junction Diode: Diffusion Capacitance

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ - + -
VD vd
ID + id

There is also charge stored in the quasi-neutral regions that changes as the junction
voltage is varied (negative and positive charge stored at the same location!!)
Minority
n'  x  p'  x  carriers

p'  x  n'  x 
 Wp  x p  xp 0 xn Wn  x n x
Majority
carriers
Charge stored:

Wn  x n xp
Qd  qA  p'  x dx  qA  p'  x dx
xn Wp  x p

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Small Signal Model of a PN Junction Diode: Diffusion Capacitance

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ - + -
VD vd
ID + id
Wn  x n xp
Charge stored: Qd  qA  p'  x dx  qA  p'  x dx
xn Wp  x p
Qd
Diffusion Capacitance: Cd 
VD

  Wp  W  
 2 cosh   1 cosh n   1
2 qVD 2 L 
q A KT  ni  Ln  n  p 
Cd  e  N Ln  i Lp
KT  W p  N W  
 a sinh  d
sinh n  
  Ln  L  
 p 

Increases exponentially with bias!

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

5
 Small Signal Model of a PN Junction Diode: Total Capacitance

P-doped - - - - + + + + N-doped
- - - - + + + +
Na - - - - + + + + Nd
+ - + -
VD vd
ID + id

At high frequencies, part of the current id flows through the junction but part of it
also charges up the junction capacitance and the diffusion capacitance

id  gd v d  C j  Cd 
dv d Cd
dt
rd

Cj

vd
id
ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Capacitances of a PN Junction Diode

Total Capacitance: C  C j  Cd

Cd

rd

Cj

vd
id

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

6
 Small Signal Model of a PN Junction Diode in Revere Bias

1 q I D  Io 
gd   0
rd KT

Cd  0

Cj

vd
id

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

Breadboard Wiring: Good Wiring

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

7
Breadboard Wiring: Bad Wiring

ECE 315 – Spring 2006 – Farhan Rana – Cornell University

ECE 315 – Spring 2006 – Farhan Rana – Cornell University