1 Esaki Diode

 When the concentration of impurity atoms in a pn-diode is very high, the

depletion layer width is reduced to about 10 nm. Classically, a carrier must
have an energy at least equal to the potential-barrier height in order to cross
the junction. However, quantum mechanics indicates that there is a nonzero
probability that a particle might penetrate through a barrier as thin as that
indicated above. This phenomenon is called tunneling, and because of this, these
high-impurity density p-n devices are called tunnel diodes, or Esaki diodes.
 The condition that the barrier be less than 10 nm thick is a necessary but
not a sucient condition for tunneling. It is also required that occupied energy
states exist on the side from which the electrons tunnel and that allowed empty
states exist on the other side.
 The energy-band diagram for a heavily - doped diode under open circuit
and reverse bias conditions is shown below.

Figure 1: Esaki diode

1
 (a) Under zero-bias conditions, there are no lled states on one side of the
junction which are at the same energy as the empty allowed states on the other
side. Hence, there can be no ow of charge in either direction across the junction,
and the current is zero, an obviously correct conclusion for an open - circuited
diode.
(b) If a reverse bias is a applied, the height of the barrier is increased above
the open-circuit value E0 . Hence, the n-side levels must shift downward with
respect to the p-side levels. In this case, there are some energy states in the
valence band on the p-side that lie at the same level as allowed empty states
in the conduction band of the n-side. Hence, these electrons might tunnel from
the p to the n - side, giving rise to the reverse diode current. As the magnitude
of the reverse bias increases, the heavily - shaded are grows in size, causing the
reverse current to increase, as shown by section (1) in the Figure below.

Figure 2: IV characteristics.

 Consider now the forward bias case when the potential barrier is decreased
below E0 . The n-side level shifts upward with respect to those on the p-side,
and the energy-band picture for this situation is indicated in Figure 3.

2
Figure 3: Forward Bias

3
 It is clear that there will be occupied states in the conduction band on the
n-side which are at the same energy as the allowed empty sates (holes) in the
valence band on the p-side. Hence electrons will tunnel from the n to the p
material giving rise to the forward current of section (2). As the forward bias
increases further, the condition in (b) is reached, giving rise to maximum current
value. If more forward bias is applied, the situation in (c) is obtained and the
tunneling current decreases, giving to section (3). At even larger bias, there will
not be empty allowed states on one side of the junction at the same energy as
the occupied states on the other side and the tunneling current drops to zero.
 In addition to the quantum-mechanical current described above, the regu-
lar pn-junction current is being collected at larger voltages, giving rise to section
(4).
 Mathematical analysis of the tunneling current is based on the results
described in the previous sections. In other words, one can use the WKB ap-
proximation to calculate the tunneling coecient (or probability) and then use
the result into the Landauer expression for the current.
 To calculate the tunneling coecient, we consider the case when E? = 0
and Ez 6= 0, and we approximate the barrier height by EG :

Figure 4: Tunneling barrier

Z
 

V (Z ) = EG 1 , W (1)
The attenuation of the barrier is thus
Z Z0 r 2m
2 = 2 ~2
[V (Z ) , EZ ]dZ
0

4
 s
Z Z0
2m E 1 , Z , E dZ
   

= 2 ~2 G
W Z
0
s
Z Z
0 2mE  Z , EZ dZ

G
= 2 ~2
1 , W E
0 g
s
Z Z0 r 2mE Z , EZ d , Z (,W )
 
G
= 2 ~ 2 1 , W Eg W
0
s
Z Z0 r
mE Z E Z
 

= (,2W ) 2 G Z
1 , W , E d ,W
0 ~2 g
3=2
Z , EZ

r
2 mE 1 , W E
Z0

G g
= (,2W ) ~2 3=2



"
0
r 3=2 3=2 #
W mE Z E E


= 4
, 3 2 G 0
1, W , E Z
, 1, E Z
(2)
~2 g g

where

1 , ZW0 , EEZ = 1 , Eg E, EZ , EEZ

g g g
E g , EZ E g , EZ
= E , E g g
(3)
Hence

2mEG 1 , EZ 3=2
r

 = 4W
 

3 ~2 Eg
r

 3 4 W 2 mE G

3 E
1, 2 E Z


(4)
~2 g
provided that EZ =EG  1. Substituting the last expression into the WKB
result for the transmission coecient gives

T (EZ ) = e,2
r

, 3 2mE
4W G 2W r 2mE
G
2 EZ
e| ~ 2
} Eg ~
= {z
e
T0
= T0eEZ =E0 (5)
where we have dened

5
 s

E0 = 2EWg 2mE ~2
g
s
~2 Eg2 1
= 2mEg 2W
r
1 2 Eg
= 2W ~2m (6)
The expression for T (EZ ) suggests that the larger the energy EZ , the larger
the probability for transmission, which should be expected.
The next task is to calculate the current under forward bias conditions, for
which we use the previously derived expression

Figure 5: Forward bias conditions.

J = JL!R , ZJR!L
em  1 Z 1

= , 22 ~3 dEZ T (EZ ) dEt [fL (EZ + Et ) , fR (EZ + Et + eVa )]

0 0
(7)
where

E = EZ + Et
= EZ1 + Et , Ecn (8)
Using the approximation (valid at T ! 0);

fL(EZ + Et ) , fR (EZ + Et + eVa ) =

6
 
@f 
= f (EZ + Et ) , f (EZ + Et ) , eVa @E =

EZ +Et
= ,eVa(EZ + Et , EF ) (9)
we get for the forward current
 Z 1 Z 1
em
J = , 22 ~3 (,eVa )T0 dEt dEZ eEZ =E0 (EZ + Et , EF )

%
0 0
0
max
= e2 m Va T0 Z Et dE e(EF , Et )=E0
t
22 ~3 0
E max
e 2 m Va T0  t
= 22 ~3 (,E0 )e , E t =E 0 


0
e2 m Va T0 E0 h max i
= , 22 ~3 e,Et =E0 , 1 (10)
where
Etmax = EV P + Ecn , eVa (11)
Therefore

e,Etmax=E0 , 1 = e,(EV P +Ecn,eVa )=E0 , 1

 1 , EV P + EEcn , eVa , 1
. .
(12)
0
i.e.,
2 
J = , e m2V2a~T30E0 , EV P + EEcn , eVa
 

0
!
e 2 m  V a T0 E + Ecn}
V P {z
= 22 ~3 |
, eVa
Et
e 2 m T0 Et  eVa


= 22 ~3 Va 1 , E (13)
t
The maximum transverse energy is obtained using
,Ecno + eVa + E?max = EV P (14)

E?max = EV P + Ecn , eVa (15)