PN Junction: Physical and

Mathematical Description of
Operation
Dragica Vasileska
Professor
Arizona State University
• PN Junctions – General Considerations
• Ideal Current-Voltage Characteristics
• Generation and Recombination Currents
• Breakdown Mechanisms
• AC-Analysis and Diode Switching
1. PN-junctions - General Consideration:
• PN-junction is a two terminal device.
• Based on the doping profile, PN-junctions can be
separated into two major categories:
- step junctions
- linearly-graded junctions
ND  N A ND  N A
ax

p-side n-side p-side n-side

Step junction Linearly-graded junction
(A) Equilibrium analysis of step junctions
EC
qVbi (a) Built-in voltage Vbi:
Ei
EF qVbi Ei  E F p  E F  Ei n
EV
p-side nn0 ni expE F  Ei  k BT 
n-side
W p p 0 ni exp Ei  E F k BT 
(x)
qND
+ k BT  p p 0 nn0   N AND 
 
Vbi  ln VT ln
q  n2   n2 
-qNA - x  i   i 
V (x)
Vbi
(b) Majority- minority carrier
relationship:
E (x) x pn0  p p 0 exp Vbi / VT 
 xp xn n p 0 nn0 exp Vbi / VT 
Emax x
(c) Depletion region width:
 Solve 1D Poisson equation using depletion charge
approximation, subject to the following boundary condi-
tions: V ( x p ) 0, V ( xn ) Vbi , E ( xn ) E ( x p ) 0
qN A
V
p-side:p ( x ) 
2k s  0
x  x p 2

qN D
Vn ( x) 
n-side: xn  x 2  Vbi
2k s  0
 Use the continuity of the two solutions at x=0, and
charge neutrality, to obtain the expression for the depletion
region width W:
xn  x p W 
V p (0) Vn (0)   W  2k s  0 ( N A  N D )Vbi
 qN A N D
N A x p  N D xn 
(d) Maximum electric field:
The maximum electric field, which occurs at the
metallurgical junction, is given by:
dV qN A N DW
Emax  
dx x 0 ks0 ( N A  N D )
(e) Carrier concentration variation:
15
10
N A  N D 1015 cm  3
Concentration [cm ]
-3

13
10
Wcalc 1.23 m
Emax( DC ) 9.36 kV / cm
11
10 -3
n [cm ]
-3
p [cm ]
Emax( sim) 8.93 kV / cm
9
10

7
10

5
10
0 0.5 1 1.5 2 2.5 3 3.5
Distance [mm]
 (f) Analytical vs. numerical data

NNAA  N  10 15cm 33

15
 NDD 10 cm
W 1.23 m, E max(DC
calc 1.23 m, Emax(
Wcalc 9.36 kV / cm, E max(sim
DC)) 9.36 kV / cm, Emax(
8.93 kV / cm
sim)) 8.93 kV / cm

15
10 0

Electric field [kV/cm]

14 -2
5x10
r(x)/q [cm ]
-3

-4
0
-6
14
-5x10
-8

15
-10 -10
0 0.5 1 1.5 2 2.5 3 3.5 0 0.5 1 1.5 2 2.5 3 3.5

Distance [mm] Distance [mm]

 NNAA 10 16cm 33, N 1018
16 18cm 33
10 cm , NDD 10 cm
W 0.328 m, E max(DC
calc 0.328 m, Emax(
Wcalc 49.53 kV / cm, E max(sim
DC)) 49.53 kV / cm, Emax(
67 kV / cm
sim)) 67 kV / cm

17
10 10

Electric field [kV/cm]

0
16
5x10 -10
r(x)/q [cm-3]

-20
0 -30
-40
16
-5x10 -50
-60
17
-10 -70
0.6 0.8 1 1.2 1.4 0.6 0.8 1 1.2 1.4

Distance [mm] Distance [mm]

(g) Depletion layer capacitance:
 Consider a p+n, or one-sided junction, for which:
2k s  0 Vbi V 
W
qN D
 The depletion layer capacitance is calculated using:
dQc qN D dW qN D k s  0 1 2(Vbi V )
C    2
dV dV 2(Vbi V ) C qN D k s  0
1 C2
Measurement setup:
1
slope 
ND W
dW
Reverse vac ~
bias Forward bias
V V
Vbi  V
(B) Equilibrium analysis of linearly-graded junction:
1/ 3
 12k s  0 Vbi V 
(a) Depletion layer width: W  
 qa 
qaW 2
(c) Maximum electric field: Emax 
8k s  0
1/ 3
 qak s2 02

(d) Depletion layer capacitance: C  
 12Vbi V 
Based on accurate numerical simulations, the depletion
layer capacitance can be more accurately calculated if Vbi
is replaced by the gradient voltage Vg:
2  a 2 k s  0VT 
Vg  VT ln  3 
3  8qni 
(2) Ideal Current-Voltage Characteristics:
Assumptions:
• Abrupt depletion layer approximation
• Low-level injection  injected minority carrier density
much smaller than the majority carrier density
• No generation-recombination within the space-charge
region (SCR)
(a) Depletion layer:
W
EC
2 2expV / V 
qV np nii expV / VTT 
np  n
E Fp
E Fn nnp (( xxp ))
p p p0 expV
nnp 0 exp V //VVT 
T
EV ppn ((xxn ))
n n  ppn 0 exp
n0 exp
VV //VVT 
T

 xp xn
(b) Quasi-neutral regions:
• Using minority carrier continuity equations, one arrives at
the following expressions for the excess hole and electron
densities in the quasi-neutral regions:
V / VT  ( x  xn ) / L p
p n ( x )  p n 0 (e  1)e
V /V ( x  x p ) / Ln
n p ( x ) n p 0 (e T  1)e

n p (x ) pn (x )
Space-charge Forward bias
region W

pn 0
n p0
x
 xp xn
Reverse bias
• Corresponding minority-carriers diffusion current densities
are: qD p
diff p n0 V / VT  ( x  xn ) / L p
Jp ( x)  (e  1)e
Lp
qDn n p 0 V / VT ( x  x p ) / Ln
J ndiff ( x)  (e  1)e
Ln
Shockley model
diff diff
J tot  J p ( xn )  J n ( x p )
diff drift
majority J diff
p  J drift
p
majority J n  Jn
J tot

diff
minority J ndiff minority J p

x
 xp xn
No SCR generation/recombination
(c) Total current density:
• Total current equals the sum of the minority carrier diffu-
sion currents defined at the edges of the SCR:
I
diff diff
I tot  I p ( xn )  I n ( x p ) Ge Si GaAs
 D p pn 0 Dn n p 0  V / V
qA
 L

L
e T 1

 
 p n 
V
• Reverse saturation current IS:
 D p pn 0 Dn n p 0  2  Dp D 
I s qA   qAni   n 
 L L  L N L N 
 p n   p D n A
(d) Origin of the current flow:

Forward bias: Reverse bias:

W EC
EC Ln qV
qVbi  V  qVbi  V 
qV E Fp
E Fn EV
E Fp
E Fn
EV
Lp
W

Reverse saturation current is

due to minority carriers being
collected over a distance on the
order of the diffusion length.
(e) Majority carriers current:
• Consider a forward-biased diode under low-level injection
conditions:
Quasi-neutrality
nn (x ) requires:
n ( x ) p ( x )
nn 0 n n


pn (x ) This leads
diff to: Dn diff
pn 0 Jn ( x )  J p ( x)
xn
x Dp

• Total hole current in the quasi-neutral regions:

tot diff drift diff
J p ( x) J p ( x)  J p ( x) J p ( x)
• Electron drift current in the quasi-neutral region:

diff  Dn  diff 1 diff

Jn ( x )  J tot   
 1 J p ( x ), E ( x )  J n ( x)
D  qn( x ) n
 p 

drift
Jn (x ) J tot
tot diff drift
J n ( x) J n ( x)  J n ( x)

J ndiff ( x )  J diff
p ( x) diff
J p (x )
x
diff
Jn (x )
(f) Limitations of the Shockley model:
• The simplified Shockley model accurately describes IV-
characteristics of Ge diodes at low current densities.
• For Si and Ge diodes, one needs to take into account
several important non-ideal effects, such as:
 Generation and recombination of carriers within the
depletion region.
 Series resistance effects due to voltage drop in the
quasi-neutral regions.
 Junction breakdown at large reverse biases due to tun-
neling and impact ionization effects.
(3) Generation and Recombination Currents
J scr
 Continuity equation for holes:
p 1 J p
  Gp  Rp
t q x
 Steady-state and no light
genera- p t 0 , G p 0
tion process:
xn
• Space-charge region recombination current:xn
 dJ p ( x ) J p ( xn )  J p (  x p )  q  R p dx
 xp  xp
xn
J scr q  R p dx
 xp
Reverse-bias conditions:
• Concentrations n and p are negligible in the depletion
region:
 ni2 ni  E t  Ei   Ei  E t 
R  ,  g  p exp    n exp 
 p n1   n p1 g  k BT   k BT 

Generation lifetime

• Space-charge region current is actually generation current:

qniW qniW
J scr  J gen   J gen   Vbi  V
g g

• Total reverse-saturation current:

J J s e V / VT
 
 1  J scr      J s  J gen
V VT

• Generation current dominates when ni is small, which is
always the case for Si and GaAs diodes.

EC
I (log-scale)

E Fp
AJ s
V (log-scale) EV
E Fn

AJ gen W

IV-characteristics Generated carriers are

swept away from the
under reverse bias conditions depletion region.
Forward-bias conditions:
• Concentrations n and p are large in the depletion region:

np 2 V / VT
ni e  R
 V /V
ni2 e T  1 
 p n  n1    n p  p1 
• Condition for maximum recombination rate:
n  p ni e
V / 2V
T Recombination lifetime
2 V /V
ni e T ni V / 2VT
Rmax   e ,  rec  p   n
n p  p n  rec
• Estimate of the recombination current:
max qniW V / 2VT
J scr  e
 rec
• Exact expression for the recombination current:
qni  V / 2VT  1 qN D 2Vbin  V 
J scr  e ,   VT , Enp 
rec 2 Enp ks0

• Corrections to the model:

qni  V / mrVT
J scr  e
 rec

• Total forward current:

J J s 
e
V / VT
1qni  V / mrVT
 rec
e 
 J s,eff e
V / VT

1

  ideality factor. Deviations of  from unity represent

an important measure for the recombination current.
• Importance of recombination effects:
Low voltages, small ni  recombination current dominates
Large voltages  diffusion current dominates

log(I)

AJ scr
AJ

V
AJ d
(4) Breakdown Mechanisms
• Junction breakdown can be due to:
 tunneling breakdown
 avalanche breakdown

• One can determine which mechanism is responsible for the

breakdown based on the value of the breakdown voltage
VBD :
 VBD < 4Eg/q  tunneling breakdown
 VBD > 6Eg/q  avalanche breakdown
 4Eg/q < VBD < 6Eg/q  both tunneling and
avalanche mechanisms are responsible
Tunneling breakdown:
• Tunneling breakdown occurs in heavily-doped pn-
junctions in which the depletion region width W is about
10 nm.
Zero-bias band diagram: Forward-bias band diagram:

EFn
EF EFp
EC
EC

EV
EV
W W
Reverse-bias band diagram: • Tunneling current (obtained by
using WKB approximation):
* 3
2m q FcrVA  4 2 m * 3/ 2 
Eg 
It  
exp 
2 2 1/ 2
4  E g  3qFcr 
 
EFp Fcr  average electric field in
EFn the junction
EC
• The critical voltage for
tunneling breakdown, VBR, is
estimated from:
EV
I t (VBR )  10 I S

• With T, Eg and It .

Avalanche breakdown:
• Most important mechanism in junction breakdown, i.e. it
imposes an upper limit on the reverse bias for most diodes.
• Impact ionization is characterized by ionization rates n and
p, defined as probabilities for impact ionization per unit
length, i.e. how many electron-hole pairs have been
generated per particle per unit length:
 Ei 
i  exp  
 qFcr 

- Ei  critical energy for impact ionization to occur

- Fcr  critical electric field
-   mean-free path for carriers
 Avalanche mechanism:

EFp

EFn
EC

EV

Generation of the excess electron-hole

Expanded view of the
pairs is due to impact ionization.
depletion region
• Description of the avalanche process:
dJ n dJ p
 0, 0
Jn J n   n J n dx dx dx
dx dJ n dJ p
J p   n J n dx Jp -
dx dx

Impact ionization initiated by electrons.
J  J n  J p const.

Jn J n   p J p dx Multiplication factors for

dx electrons and holes:
J p   p J p dx Jp J n (W ) J p ( 0)
Mn  , Mp 
J n ( 0) J p (W )
Impact ionization initiated by holes.
• Breakdown voltage  voltage for which the multiplication
rates Mn and Mp become infinite. For this purpose, one
needs to express Mn and Mp in terms of n and p:
x
  
   n   p dx '
1 W
 dJ n 1    n e 0
dx
 dx  n J n   p J p  Mn 0
 dJ   x
 p   n J n   p J p  1 W  
   n   p dx '
 dx 1    p e 0
dx
 Mp 0

The breakdown condition does not depend on which

type of carrier initiated the process.
• Limiting cases:
(a) n=p (semiconductor with equal ionization rates):
 1 W 1
 1  M   n dx  M n  W
 n 0 1   n dx
 0
 W
1  1 1
  p dx  M p  W
 Mp 0
 1   p dx
 0

(b) n>>p (impact ionization dominated by one carrier):

W
 n dx W
M n e 0 1   n dx
0
Breakdown voltages:
(a) Step p+n-junction • For one sided junction we can
make the following approximation:
W Wn  W p Wn
 n
p • Voltage drop across the depletion
region on the n-side:
1 1
Vn  FmaxWn VBD  FmaxW
 Wp Wn 2 2
• Maximum electric field:
 F (x ) qN DW k s 0 2
 Fmax Fmax   VBD  Fmax
k s 0 2 qN D
• Empirical expression for the
breakdown voltage VBD:
3/ 2 -3 / 4
x E
 g   ND   kV 
VBD 60   16 
 cm 
 1.1   10 
(b) Step p+-n-n+ junction
• Extension of the n-layer large:
1
 n  VBD  FmaxWm
p n 2
• Extension of the n-layer small:
 Wp W1 Wm 1 1
VP  FmaxWm  F1 Wm  W1 
 F (x )
2 2
 Fmax • Final expression for the punch-
through voltage VP:
 F1 W1  W1 
VP VBD  2  
x Wm  Wm 
• Doping-dependence of the breakdown voltage VBD:

VBD Width of the n-layer W1 increases

One-sided abrupt
Log-scale
junction
Tunneling breakdown

p+-n-n+
ND
Log-scale

• Temperature dependence:
As temperature increases, lattice scattering increases which
makes impact ionization less probable. As a result of this,
the breakdown voltage increases.
(c) Plane vs. planar or cylindrical junction
• Plane junction:
Maximum electric field:
p+ Q qN DW
Fmax  
k s 0 k s 0
Except for surface effects, this is an
n
ideal junction.

• Planar junction: Maximum electric field:

rj
qN DW  W 
p
+
Fmax  1
k s  0  2 r j 
The smaller the radius rj, the larger
W
n the electric field crowding.
(5) AC-Analysis and Diode Switching
(a) Diffusion capacitance and small-signal equivalent
circuit
• This is capacitance related to the change of the minority
carriers. It is important (even becomes dominant) under
forward bias conditions.
• The diffusion capacitance is obtained from the device
impedance, and using the continuity equation for minority
carriers: dp
2
d p p
n D p n  n
dt 2 p
dx
• Applied voltages, currents and solution for pn:
it
V (t ) V0  V1e , V1  V0 it
it pn ( x, t )  pns ( x )  pn1 ( x )e
J (t )  J 0  J 1e , J 1  J 0
• Equation for pn1(x):
2 2
d pn1 1  i p d pn1 pn1 ( x )
2
 pn1 ( x ) 0  2
 2
0
dx Dp p dx L p'

• Boundary conditions:
pn ( , t )  pn 0  pn1 ( ) 0
 V0  V1eit  pn 0V1  V0 
pn (0, t )  pn 0 exp    pn1 (0)  exp 
 V  VT  VT 
 T 

• Final expression for pn1(x):

pn 0V1  V0   x 
pn1 ( x, t )  exp  exp  
VT V
 T  L 
 p' 
• Small-signal hole current:
dpn1 AqD p pn 0V1  V0 
I1  AqD p  1  i p exp  YV1
dx x 0 L pVT  VT 
• Low-frequency limit for the admittance Y:
AqD p pn 0  V0   1
Y exp   1  i p  Gd  iCdif
L pVT  VT   2 
AqD p pn 0  V0  I s eV0 / VT I dI
Gd  exp     , I  Forward current
L pVT  VT  VT VT dV
1 AqD p pn 0  V0  1 I
Cdif   p exp   p
2 L pVT  VT  2 VT
• RC-constant:
p The characteristic time constant is on
Rd Cdif  the order of the minority carriers lifetime.
2
• Equivalent circuit model for forward bias:
Cdepl

Rs Ls
Cdif

1
Rd 
Gd

• Bias dependence: C Cdif

Cdepl

Va
(b) Diode switching
• For switching applications, the transition from forward bias
to reverse bias must be nearly abrupt and the transit time
short.
• Diode turn-on and turn-off characteristics can be obtained
from the solution of the continuity equations:
d p n 1 1D 1 J p pn
  J p  R p     
dt q q x p

dQ p Qp dQ p Q p
 I p (t )   I (t )  I p (t )  
dt p dt p

Qp(t) = excess hole charge

Valid for p+n diode
Diode turn-on:
• For t<0, the switch is open, and
p+ n
the excess hole charge is:

Q p (t  0) Q p (0 ) 0
t=0
• At t=0, the switch closes, and
we have the following boundary IF
condition:
 
Q p (0 ) Q p (0 ) 0

• Final expression for the excess hole charge:

 t /p   t /p 
Q p (t )  A  Be  p I F 1  e
 
• Graphical representation:
Q p (t ) pn ( x , t )
Slope almost constant

pIF t increasing

pn 0
t x

• Steady state value for the bias across the diode:


p n ( x )  p n 0 e
Va / VT

 1e
 x / Lp

 Q p  Aqpn 0 L p e
Va / VT

1

 IF 
Va VT ln 1  
 IS 
Diode turn-off:
• For t<0, the switch is in position
p+ n
1, and a steady-state situation is
established: t=0
VF
IF  1 2
R VF VR
• At t=0, the switch is moved to
R R
position 2, and up until time t=t1
we have:
pn (0, t )  pn 0  Va 0

• The current through the diode

until time t1 is: V
I R  R
R
• To solve exactly this problem and find diode switching
time, is a rather difficult task. To simplify the problem, we
make the crucial assumption that IR remains constant even
beyond t1.
• The differential equation to be solved and the initial
condition are, thus, of the form:
dQ p Qp  
 IR   , Q p (0 ) Q p (0 )  p I F
dt p
• This gives the following final solution:
 t /p
Q p (t )   p I R   p I F  I R e

• Diode switching time:

 IF 
Q p (trr ) 0  trr  p ln 1  
 IR 
• Graphical representation:
Va (t )

pn ( x , t ) t
Slope almost
constant
t=0
 VR

pn 0 t=ts
IF
ttrr
x
ts trr
t
 0.1I R
tts  switching time
s  switching time
ttrr 
reverse
reverserecovery
recoverytime
time
rr
 IR