Supercold Semiconductor

  • Thread starter Drakkith
  • Start date
  • #1
Drakkith
Staff Emeritus
Science Advisor
20,717
4,423

Main Question or Discussion Point

I just finished reading Absolutely Small, a book on quantum physics. In it the author says that when semiconductors get very cold the electrons no longer have enough energy to jump the band gap, and therefor they do not work.

I have also watched a video of a guy who poured liquid nitrogen onto his processor to overclock it to something like 6 GHz. I was wondering how this is possible? Is it that the processor never gets that cold due to the overclocking?
 

Answers and Replies

  • #2
ZapperZ
Staff Emeritus
Science Advisor
Education Advisor
Insights Author
35,467
4,230
I just finished reading Absolutely Small, a book on quantum physics. In it the author says that when semiconductors get very cold the electrons no longer have enough energy to jump the band gap, and therefor they do not work.

I have also watched a video of a guy who poured liquid nitrogen onto his processor to overclock it to something like 6 GHz. I was wondering how this is possible? Is it that the processor never gets that cold due to the overclocking?
First of all, liquid nitrogen is not "supercold". In fact, it is downright hot in many applications (liquid helium will BOIL when it comes in contact with liquid nitrogen). So at this temperature, one hasn't reached the state yet where the semiconductors (if the band gap is small enough) become an insulator.

Secondly, cooling a circuit typically reduces resistance and improve conductivity (at least in the "wiring"). How it improves on the processor speed, someone with more knowledge on computer engineering will have to chime in.

Zz.
 
  • #3
Drakkith
Staff Emeritus
Science Advisor
20,717
4,423
First of all, liquid nitrogen is not "supercold". In fact, it is downright hot in many applications (liquid helium will BOIL when it comes in contact with liquid nitrogen). So at this temperature, one hasn't reached the state yet where the semiconductors (if the band gap is small enough) become an insulator.

Secondly, cooling a circuit typically reduces resistance and improve conductivity (at least in the "wiring"). How it improves on the processor speed, someone with more knowledge on computer engineering will have to chime in.

Zz.
That's pretty much what I was thinking, that the liquid nitrogen wasn't quite cold enough or that the processor wasn't getting cold enough. Thanks!
 
  • #4
131
0
A few things:

CMOS processors use doped semiconductors. The intrinsic carriers (those that are thermally excited from the valence band to the conduction band) are only a tiny portion of carriers. The carriers excited from the donor levels are much closer to the conduction band (or acceptors closer to the valence band). These carriers will freeze out at much lower temperature than 77K.

Low temperature will increase the carrier mobility in the semiconductor, which enables faster speeds.

And like ZapperZ says, the lower temperature will reduce the resistance in the metalization layers, which is where a lot of the heat is generated.

But I think that the main benefit of LN2 in this case is simply better cooling to remove excess heat.

EDIT: actually, I'll revise my first paragraph a little : in MOSFETs, the channel is formed by inversion and the Fermi level is near or above the conduction band edge so even at lower temperature, the channel would still have a significant amount of carriers.
 
Last edited:

Related Threads for: Supercold Semiconductor

  • Last Post
Replies
4
Views
3K
  • Last Post
Replies
2
Views
2K
  • Last Post
Replies
3
Views
1K
  • Last Post
Replies
1
Views
3K
  • Last Post
Replies
1
Views
7K
  • Last Post
Replies
1
Views
3K
  • Last Post
Replies
5
Views
4K
  • Last Post
Replies
3
Views
3K
Top