ECE 340
Lecture 26 : Reverse Bias
Breakdown
Class Outline:
• Avalanche and Zener Processes
Things you should know when you leave…
Key Questions
•  What is the Zener effect?
•  What is avalanching?
•  When do these occur?
•  What do they depend on?
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
Most of the preceding analysis dealt with forward bias, what about the
reverse bias case?
We can use the same equations and analysis to determine the reverse bias
behavior…
Set V = -Vr which biases the p-side negatively with respect to the n-side and
examine the relationship for the excess hole concentration…
Vr >> kbT/q
• For large reverse bias, the minority carrier concentration goes to zero.
• Minority carrier concentration equations still given by previously derived equations.
• Depletion of minority carriers extends one diffusion length on either side of the
junctions.
• Referred to as minority carrier extraction.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
What is happening physically to
the carriers…
• Carriers are being swept down the
barrier at the junction to the other
side.
• They are not being replaced by an
opposing diffusion of carriers.
• Reverse bias saturation occurs
because of drift of carriers down
the barrier
• But the rate of drift depends on the rate of minority
carriers arrive by diffusion from the neutral material
supplied by thermal generation.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
And the quasi-Fermi levels move again…
• Fn moves farther away from EC towards EV because in reverse bias we have fewer
carriers than in equilibrium.
• Quasi-Fermi levels here go inside the bands but we need to remember that Fp
is a measure of the hole concentration and is correlated with EV and not EC.
• This just tells us we have very few holes (smaller than in equilibrium).
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
We already know what happens when we apply a small
reverse bias…
Forward bias lowers the barrier and results in a
forward current.
• Small reverse bias
causes a reverse
saturation current
caused by EHPs
being swept across
the space charge
region.
• At large enough
reverse bias, a very
large current begins
to flow.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
Is the reverse bias breakdown
destructive?
• The reverse bias regime is not
destructive so long as the current is
limited.
• Under these circumstances, reverse bias
is no more destructive than forward bias.
The resistance is chosen so as to make the
current sufficiently small to keep from
harming the device.
Failure to do so can also cause significant
heating to the p-n junction.
M.J. Gilbert
Damage is not necessarily due to
mechanisms specific to reverse
bias.
Similar effects can be present
with too much forward bias.
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
The first of two breakdown mechanisms is zener breakdown…
The Zener effect is the breakdown mechanism if the
reverse bias required to force breakdown occurs at low
voltages.
Consider the heavily doped p-n junction shown to the left
and then apply a reverse bias to the junction.
Reverse bias brings the conduction
band very close to the valence band.
This brings many occupied states on
the p-side into energetic alignment
with vacant states on the n-side.
Electrons tunnel from the valence
band to the conduction band giving
rise to a reverse current.
This is the Zener effect.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
So what do we need to cause Zener breakdown?
The basic requirements to drive a tunneling current are:
• A large number of electrons
• A large number of holes
• Separated by a narrow barrier of finite height.
Keep in mind…
• Tunneling depends heavily on the barrier width so we want
to keep the junction sharp and doping high.
Zener effect
• This will ensure that the transition region W extends only
a very short distance from each side of the junction.
• Failure to attain high doping or sharp junctions will result
in no tunneling current.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
What else is important about the Zener
process?
• As the reverse bias is increased the
distance between the bands decreases.
• This is due to higher electric fields
increasing the slopes of the bands.
• We assume that the transition region,
W, does not increase with bias which is
valid for high doping and low voltage.
When does it happen…
What does the current look like…
• Occurs in Si for fields ~ 106 V/cm
• Must have high impurity
concentrations
Average electric field in junction
M.J. Gilbert
• Occurs in general for reverse biases
of less than 4Eg/q.
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
For devices with large breakdown voltages or devices that are lightly
doped, the major breakdown mechanism is avalanche breakdown.
In a lightly doped junction the tunneling is
negligible due to the large distance between
bands.
Instead, breakdown is caused by impact
ionization of the host atoms by energetic
carriers.
Normal lattice scattering can cause the creation
of EHPs if the carrier being scattered has
sufficient energy.
These single event scattering interactions are
results in carrier multiplication.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
But we are not dealing with a single collision…
When the electron scatters it creates an EHP.
Each of these carriers then has a chance to be
accelerated and collide with the lattice creating a
new EHP.
This process may continue and is referred to as
avalanche breakdown.
Let’s make an approximate analysis of the physics…
• A carrier (either electron or hole) has some probability of having an ionizing
collision with the lattice while being accelerated through the transition region, W.
• For nin electrons entering from the p-side, there will be Pnin ionizing collisions and
an EHP generated for each collision.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
The process continues…
After Pnin collisions by the primary electrons, we have the primary plus the
secondary electrons, nin (1 + P).
After a collision, each EHP moves a distance of W within the transition region. If
the pair is created at the center, the electron will drift W/2 to n and the hole will
drift –W/2 to p.
For ninP secondary electrons there will be (ninP) P collisions creating ninP2 tertiary
pairs.
After n collisions we have:
Assume no recombination and equal probabilities of ionizing collisions, then the
electron multiplication is…
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
If the ionization probability approaches unity then the carrier
multiplication becomes infinite!
So then what limits the current?
External circuit
But this was waaaaaay too easy…
• We assumed no recombination and equal ionization probabilities.
• We expect the probability to increase with increasing electric field so the
multiplication should depend on the reverse bias…
N is between 3-6
depending on the
material.
We can determine an
empirical relation…
In general:
• Critical voltage for breakdown increases with bang gap.
• Peak field in W increases with increased doping, thus Vbr decreases as doping
increases.
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
Can we make any quantitative statements about when this occurs?
Let’s consider a step-junction and evaluate the maximum
electric field in the junction as we did when we looked into pn
junction electrostatics…
But we already know something about the fields and
potentials…
when
Now square the
field dependence:
But this field should have no doping dependence!
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
In summary…
Single collision
Primary, secondary
and tertiary
collisions
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
For reference, let’s look at some actual devices…
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
For reference, let’s look at some actual devices…
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
For reference, let’s look at some actual devices…
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11
Avalanche and Zener Processes
For reference, let’s look at some actual devices…
M.J. Gilbert
ECE 340 – Lecture 26
10/21/11