Understanding Drift Current in a pn junction

[lonechicken]

I'm having trouble fully conceptualizing drift current- what I always hear is that it is caused by minority carriers within one diffusion length of the junction being swept across by the electric field that is generated by the imbalance caused by diffusion of electrons.

Here is what I am wondering:

1)Everyone says "within one diffusion length"- does this mean anywhere in the diffusion region including the neutral region for as far a distance as one diffusion length into the neutral region, or does it just mean WITHIN the neutral region, in an area up to one diffusion length away from the junction.
In other words, are the minority carriers generated only in the neutral region close to the depletion region or are they also generated within the depletion region?

2) Also, what is responsible for this thermal generation?

3)Finally, is the leakage current in a diode this drift current?

Thanks!

 

[lonechicken]

Any answers? This would really mean a world to me...

 

[es1]

1) Do you mean the length of the depletion region or the diffusion length? There are way more minority carriers at the boundary so usually just those are considered. But there will be some small amount generated within that will also get swept across.
2) Unless it is being held at 0K the semiconductor will have some temperature due to its environment. Even if it is just room temperature.
3) Technically there is also some leakage due to contamination in actual diodes. But this drift current is a large part of it and in a perfect classic PN diode, I think it would be all of it.

 

[lonechicken]

Thank you very much for your reply! It cleared a huge amount up for me and I am eternally grateful...

Just one more question- if I understood your reply to 1) correctly you say there are many more minority carriers at the boundary than within the depletion region- why is this? Shouldn't there be plenty of minority carriers entering the depletion region due to diffusion?
Aren't the electrons which travel from the n region to the p depletion region (and the holes which travel from the p region to the n depletion region) minority carriers or am I missing something?

Thanks again for your reply, in any case..

 

[es1]

So I am assuming the PN junction is reverse biased.

I think I might be wrong about the boundary though. I was thinking there would be more at the junction because the minority carries from the p-side would be at the junction and there would be basically none in the depletion region by definition (that is how the region gets it name).

But I am probably wrong because I do remember that drift current does not vary with reverse voltage so there is very like the same number of minority carriers everywhere.
(You can see it in figure 2 of this datasheet http://www.onsemi.com/pub_link/Collateral/1N4001-D.PDF)

 

[es1]

I refreshed my memory by reading online a little bit. I think I can answer 1) better now and I think "everyone" was referring to diffusion length, not depletion region width.

Here is a concise definition of diffusion length:
http://pvcdrom.pveducation.org/SEMICON/LN.HTM
"A related parameter, the "minority carrier diffusion length" is the average distance a carrier can move from point of generation until it recombines."

So a minority carrier that is farther away from the depletion region than a diffusion length does not get across the junction because it gets recombined (on average) before it can do so therefore it cannot contribute to the drift current. That is why one should only considers minority carriers within a diffusion length.

I thought this was a pretty good link as well:
http://www.ece.utep.edu/courses/ee3329/ee3329/Studyguide/ToC/PNdiode/currents.html
The last two paragraphs are especially relevant.

 

[lonechicken]

That definition made things a bunch clearer for me and once again, thank you for your reply and your effort!

However, what I may have expressed badly is that I am wondering whether minority carriers generated within the depletion region contribute to the drift current, or only those not in the depletion region but a diffusion length away from it.

In other words, I understand why minority carriers further than a diffusion length away in the neutral regions do not contribute to drift. What I am wondering is if minority carriers in the depletion region contribute to the drift or only those in the neutral regions within a diffusion length of the depletion region.

Hope this is more clear, thanks again!

 

[es1]

No problem. I am having a good time refreshing my memory.

I think the short answer to your question is yes. Minority carriers in the depletion region contribute to the drift current.

However, there is nothing really special or unique about them as the question implies.

The only reason we even mention the depletion region is because the current in the diode is defined as the net charge passing that point so we have to stick with it. In the reversed bias case that charge is made up of only the drift current of minor carriers because they are the only ones that can pass through the potential barrier (see the second link to utep.edu I provided in the post above for details).

So minority carriers are getting thermally generated everywhere in the SI, including the depletion region, and are getting pushed along by the electric field (the one causing the reverse bias in the first place) and they are the only ones not getting blocked by the potential barrier of the depletion region. The only reason the current is small is there are just not that many of them out there, hence the name minority.

You should definitely double check my understanding with an instructor or TA if you can. I am a bit rusty with all this stuff.

 

[es1]

Also, maybe now it makes sense why increasing the reverse bias voltage does not increase the drift current.

You don't get more minority carriers with more voltage so more charge cannot pass through the depletion region (our reference point) so current has to remain the same. Increasing the temperature however can increase the drift (or leakage) current significantly.

Check out figure 2 in the 1n4001 datasheet I linked to for some empirical data.
Keep in mind these are real diodes and the y-axis is log.