Zener breakdown vs Avalanche breakdown

[Bassalisk]

Hello,

I am familiar with both of terms that i speak of in title. But I cannot find a full answer, so I might as well ask the PhD'ers here. What is REALLY happening in Zener and Avalanche breakdown? I have read this
http://cnx.org/content/m1009/latest/

And yes I get that impact ionization thing etc. But still, how does Zener differ from Avalanche breakdown? Why is doping so important? How come it doesn't damage the diode? Why is current constant? Why would it be constant when u exceeded the depletion zone(reverse bias), current should be proportional to voltage? (more voltage, more energy, more charges pulled out, more current)

I am trying to get a full picture here. You may post links with detailed quantum mechanics, semiconductor theory. I am very eager to learn.

Thanks

 

[jsgruszynski]

Same phenomena except that the Zener breakdown process has a negative temperature coefficient so it's self-limiting thermally while avalanche has a positive temperature coefficient so it's not and it can be destructive because it causes self-heating over time which accelerates failure mechanisms (damage) of all sorts including but also going beyond damage caused by the avalanche current flow.

The difference occurs based on voltage. Above a certain value the temperature coefficient changes from negative to positive. The doping will have some effect on this but it's also the balance of thermal conduction vs. IR losses.

 

[Bassalisk]

Same phenomena except that the Zener breakdown process has a negative temperature coefficient so it's self-limiting thermally while avalanche has a positive temperature coefficient so it's not and it can be destructive because it causes self-heating over time which accelerates failure mechanisms (damage) of all sorts including but also going beyond damage caused by the avalanche current flow.

The difference occurs based on voltage. Above a certain value the temperature coefficient changes from negative to positive. The doping will have some effect on this but it's also the balance of thermal conduction vs. IR losses.

Thanks. Can you provide any link with more detail to it?

 

[Bassalisk]

Relevant to this post:

If I control the voltage at which Zener diode works by doping, (effectively what is going on is quantum tunneling in reverse bias) how come zener diode doesn't behave like Tunnel diode in forward bias?

Heavily doped p and n regions allow quantum tunneling to happen in reverse bias ergo I will have zener breakdown and not avalanche breakdown. But this is same for tunnel diode, what is the difference?

 

[pantaz]

This should have the level of detail you're requesting:
http://ecee.colorado.edu/~bart/book/book/contents.htm

This is more verbose (even though the site is kind of cheeky):
http://britneyspears.ac/physics/basics/basics.htm

 

[Bassalisk]

This should have the level of detail you're requesting:
http://ecee.colorado.edu/~bart/book/book/contents.htm

This is more verbose (even though the site is kind of cheeky):
http://britneyspears.ac/physics/basics/basics.htm

Thank you

 

[Bassalisk]

Relevant to this post:

If I control the voltage at which Zener diode works by doping, (effectively what is going on is quantum tunneling in reverse bias) how come zener diode doesn't behave like Tunnel diode in forward bias?

Heavily doped p and n regions allow quantum tunneling to happen in reverse bias ergo I will have zener breakdown and not avalanche breakdown. But this is same for tunnel diode, what is the difference?

 

[dlgoff]

Relevant to this post:

If I control the voltage at which Zener diode works by doping, (effectively what is going on is quantum tunneling in reverse bias) how come zener diode doesn't behave like Tunnel diode in forward bias?

Heavily doped p and n regions allow quantum tunneling to happen in reverse bias ergo I will have zener breakdown and not avalanche breakdown. But this is same for tunnel diode, what is the difference?

Check out the difference in their IV curves.

http://hyperphysics.phy-astr.gsu.edu/hbase/solids/zener.html"

 

[Bassalisk]

But that is the question. WHY don't they both behave same? Zener effect occurs because of Quantum Tunneling appearing. Very important to differentiate from avalanche breakdown.

Tunnel diode uses quantum tunneling not only in inverse but in forward too. Both diodes are made by highly doping regions. How come they behave different ?

 

[Studiot]

Tunnel diodes are much more heavily doped than voltage reference diodes.

I will see if I can post some Fermi diagrams later.

 

[dlgoff]

But that is the question. WHY don't they both behave same? Zener effect occurs because of Quantum Tunneling appearing. Very important to differentiate from avalanche breakdown.

Tunnel diode uses quantum tunneling not only in inverse but in forward too. Both diodes are made by highly doping regions. How come they behave different ?

Like Studiot said.

In the tunnel diode, the dopant concentration in the p and n layers are increased to the point where the reverse breakdown voltage becomes zero and the diode conducts in the reverse direction. However, when forward-biased, an odd effect occurs called “quantum mechanical tunnelling” which gives rise to a region where an increase in forward voltage is accompanied by a decrease in forward current.

http://en.wikipedia.org/wiki/Tunnel_diode"

 

[Bassalisk]

But in order to quantum tunneling to occur(in reverse bias) you need lots of doping, so that the depletion region is thin and fermi levels are in valence and conductance bands(degenerate semiconductors).
This is, how I learned Zener diodes are made. Lots of doping, so that you have Zener breakdown(through quantum tunneling, and not avalanche effect). Is there any limit in doping,when this quantum tunneling starts dominate in forward bias too?

 

[dlgoff]

The more you dope the semiconductor the more it becomes like a conductor. So I would think there is a limit where conduction begins and tunneling stops. IMO

 

[Studiot]

Some facts & figures

Normal PN junctions both materials doped at 10¹³ to 10¹⁷ impurity atoms per millilitre.

In tunnel diodes both Pand N material doped at 10¹⁹ to 10²⁰ impurity atoms per mL.

Voltage reference diodes are doped differentially.

Low reference voltages

N material 10¹⁹ donor atoms per mL

P material 10¹⁷ acceptor atoms per mL

High (50V) reference voltages

N material 10¹⁹ donor atoms per mL

P material 10¹⁵ acceptor atoms per mL

These different values change the relative positions of the valence & conduction bands and Fermi level. Note there are two Fermi different levels in the normal N and P material. These levels coalesce across the junction depletion zone to a single level.

 

[Bassalisk]

Some facts & figures

Normal PN junctions both materials doped at 10¹³ to 10¹⁷ impurity atoms per millilitre.

In tunnel diodes both Pand N material doped at 10¹⁹ to 10²⁰ impurity atoms per mL.

Voltage reference diodes are doped differentially.

Low reference voltages

N material 10¹⁹ donor atoms per mL

P material 10¹⁷ acceptor atoms per mL

High (50V) reference voltages

N material 10¹⁹ donor atoms per mL

P material 10¹⁵ acceptor atoms per mL

These different values change the relative positions of the valence & conduction bands and Fermi level. Note there are two Fermi different levels in the normal N and P material. These levels coalesce across the junction depletion zone to a single level.

Hmmmm to me, these are very subtle changes in doping, between tunnel and zener diodes.

So these small changes in doping can dictate how diode will behave in reverse bias? Just to confirm: If I have regular doping, regular diodes, they have avalanche breakdown, because depletion region is too wide for quantum tunneling to occour?

Zener, or as you said voltage reference diodes, have significant more doping involved. Depletion region is thinner, and fermi levels are moved to conductance/valence band. But still these diodes behave just like normal ones do in forward bias.


Tunnel diodes have degenerate doping, so much that the quantum tunneling appear both in reverse and in forward bias?

I understand how this quantum tunneling works, quiet well. Can you confirm/correct this?

Thank you

 

[paulfr]

Not mentioned simple fact

Zener breakdown is a result of very high field intensity [voltage over the junction]
which pulls the electrons right out of their shell.

Avalanche breakdown results from kinetic electrons colliding with atoms and knocking them out of their outer shells.

Zener process occurs first, then the Avalanche process.
Not sure if this can be reversed.

 

[Studiot]

Seeing as you have exams soon I will post what I can.

First installment


My fig1

we should start with the current voltage characteristic of a PN junction (diode).
Points to note are that it may be forward biased (C-D-E), reverse bised (A-B-C) or zero biased.

In Portions AB and DE of the curve the junction has broken down and is acting as a conductor. Note both sections look like the curve for a resistor.
AB exhibits avalanche breakdown at sufficient reverse voltage, depending upon the doping.
DE the forward bias is sufficient to boost electrons into the conduction band. this leaves holes in the valence band. Both contribute to conduction.



My Fig2
By increasing the doping of we can reduce the reverse bias point at which B occurs thus creating high value reference diodes.
Further increasing the doping introduces a small tunnel effect as fig 3 creating lower voltage zenere diodes. The next installment will show that the zener and tunnel effects are very similar.

MyFig3

The tunnel diode I - V curve can be seen to be made up from a normal diode characteristic plus the tunnel effect which only occurs over a small range of about 0 - 1 volt. Since the normal diode has almost no conduction in this region the tunnel effect dominates here
The doping is so high that the diode is in breakdown in both forward and revers biase so there is a resistor like line through the origin.
then there is the characteristic tunnel hump which occurs as the forward bias first brings the conduction and valence bands into alignment, then drives them out of alighment again.

It should be noted that only electrons are involved in tunneling. Holes play no part in this.

 

[Bassalisk]

Seeing as you have exams soon I will post what I can.

First installment My fig1

we should start with the current voltage characteristic of a PN junction (diode).
Points to note are that it may be forward biased (C-D-E), reverse bised (A-B-C) or zero biased.

In Portions AB and DE of the curve the junction has broken down and is acting as a conductor. Note both sections look like the curve for a resistor.
AB exhibits avalanche breakdown at sufficient reverse voltage, depending upon the doping.
DE the forward bias is sufficient to boost electrons into the conduction band. this leaves holes in the valence band. Both contribute to conduction.
My Fig2
By increasing the doping of we can reduce the reverse bias point at which B occurs thus creating high value reference diodes.
Further increasing the doping introduces a small tunnel effect as fig 3 creating lower voltage zenere diodes. The next installment will show that the zener and tunnel effects are very similar.

MyFig3

The tunnel diode I - V curve can be seen to be made up from a normal diode characteristic plus the tunnel effect which only occurs over a small range of about 0 - 1 volt. Since the normal diode has almost no conduction in this region the tunnel effect dominates here
The doping is so high that the diode is in breakdown in both forward and revers biase so there is a resistor like line through the origin.
then there is the characteristic tunnel hump which occurs as the forward bias first brings the conduction and valence bands into alignment, then drives them out of alighment again.

It should be noted that only electrons are involved in tunneling. Holes play no part in this.

@ Figure 2: Ultimately, increasing the doping, you move the point where avalanche effect is starting. Do voltage reference diodes work with avalanche or with tunnel effect in reverse bias, or in both?

At some point in doping, you made the depletion region so small that the tunnel effect can occur? Those are Low voltage reference diodes? Ok from you post that makes sense.

Isn't avalanche breakdown bad? Because it is not self limiting, and it can destroy the diode?

Sorry if i repeated some statements of yours, I am trying to be precise.

 

[Bassalisk]

And I must say THANK you again Studiot. I really have a good knowledge of Electronic elements, just I have a lot of gaps which result in trivial questions like this. I am a fast learner but, you know, I have to be supplied with a pretty damn good explanation. I don't just accept things.

Thank you for your time and effort for trying to explain things (that usually one would pay for) to a complete stranger.

I must say that my knowledge of electronic elements and circuits wouldn't be as good if I wasn't active on this forum.

Thank you all, especially Studiot. He is gem of this forum.

 

[Studiot]

Isn't avalanche breakdown bad? Because it is not self limiting, and it can destroy the diode?

That is why you employ a series current limiting resistor.
Hopefully the answers to the others will become clear. I am trying to get several pages in.

This attachment shows cross sections of an PN junction for normal doping and heavy doping and explains how tunneling can occur.

The energy bands are at slightly lower levels in N material than P.
The fermi levels are also different so there are two fermi levels, one in the bulk N and one in the bulk P, when they are separated.
The first effect of creating a junction is to create a single unified fermi level.
The second is to create a 'depletion zone' .




First sketch normally a doped PN junction

There is an energy gap that needs to be overcome between the highest valence level in the P and the lowest conduction level in the N. This is why the junction needs a forward bias to operate.
The 'depletion zone' is about 1000 nanometres wide.

Second sketch show what happens when heavy doping add carriers to the material.

Firstly the depletion zone squashes up by up to 100 times in size.
Secondly the energy levels are brought together so the forbidden zone disappears and the bands overlap.
The fermi level cuts both the valence band of the P material and the conduction band of the N material. This allows electrons to 'tunnel through'.

Note there are more conditions to be met which will appear in the next post.

 

[Studiot]

Now that you have (hopefully ) got hold of the idea of studying what happens as we increase the bias voltage across the device we can do this for the tunnel diode.

The previous fermi diagram was for zero bias, so we shall start with this.

As before there is overlap in the energy bands between the P and N material.

However there is a slight modification. It is not possible for an electron to cross unless there is a filled slot in the conduction band of the N at the same energy (ie opposite) an empty slot in the P. An electron cannot tunnel from a filled slot to another filled slot.

Fig1
At zero bias there are no such vacancies opposite the filled slots in the conduction band so no tunnel current flows.

Fig2
Applying a forward bias voltage slowly raises the conduction band. As this bias increases the filled slots in the conduction band come opposite empty ones in the valence band so electrons can tunnel across from N to P at the same energy level.

I have shown again the tunnel response I - V curve and marked points on it so we have moved from point A at zero bias up the slope to point B.

Fig3
We continue up the slope to the point of maximum match between filled conductionband slots and empty valence band ones. This corresponds to the peak tunnel current at point C.

Fig 4
As we continue to increse the bias the gap lifts further and the tunnel current decreases down the slope on the other side as the match decreases. Point D is typical.

Fig 5
Finally the filled conduction slots lift clear of match with the empty valence slots and the tunnel current ceases when we reach point E

Sorry it's not very pretty but it's late here and I was rushing.

Go well in your exams.

 

[Bassalisk]

Its no problem at all. I am still doing this problem on paper drawing diagrams myself trying to fully comprehend this. I must say even after all this months of diode analyze, there still some new things to me.

Thank you very much, I will post results of my thoughts asap.

 

[Bassalisk]

God bless you Studiot. I understood everything you wrote and drew from top to bottom. Man I feel so smart now :D THANK YOU Now zener effect makes sense, avalanche I understood far before this. Everything is in place now.