REGISTER | LOGIN
Breaking News
News & Analysis

Scientists Pursue Super-Fast Material

Crystal electrons 1,700x faster than silicon
3/16/2015 00:01 AM EDT
10 comments
More Related Links
View Comments: Newest First | Oldest First | Threaded View
Bert22306
User Rank
Author
Re: Timing & Characterization
Bert22306   3/18/2015 4:16:19 PM
NO RATINGS
Yes, Alex, I think you have it exactly right.

alex_m1
User Rank
Author
Re: Timing & Characterization
alex_m1   3/18/2015 12:08:36 AM
@Bert, thanks for chiming in. So electron mobilitiy just affects transistors, which is why it's important for rf circuits, but in digital circuits the problem is with wire capacity , no? and basically the only ways to speed this up is either by shortening wires(i.e. smaller features), or using better materials to decrease stray capacitance or byh decreasing drive current(but finding a way to manage the heat) ?  

Bert22306
User Rank
Author
Re: Timing & Characterization
Bert22306   3/17/2015 9:48:40 PM
NO RATINGS
Colin ,roughly speaking ,by how much sucha 1000x fold improvement in mobility will improve computing speeds(let's just consider the internal speed/throughput of a cpu) ?

I ain't Colin, but I would question why speed would increase at all. It's not electron mobility that determines how fast an electric pulse travels, right? It is the mutual josteling of one electron against the next one. The speed of an electric pulse traveling through wire is already about 2/3 of the speed of light, which is essentially identical to the speed of photons traveling through an optical fiber.

So my question is, will electron mobility change this appreciably? I think the answer is no, at least no for anything having to do with the speed of electric pulses traveling through metals.

R_Colin_Johnson
User Rank
Author
Re: Timing & Characterization
R_Colin_Johnson   3/17/2015 7:12:39 PM
NO RATINGS
Sounds like a trick question:) I'll Ask the authors what they are shooting for--but I'm on a plane now so might take a while.

alex_m1
User Rank
Author
Re: Timing & Characterization
alex_m1   3/17/2015 1:54:42 PM
NO RATINGS
Colin ,roughly speaking ,by how much sucha 1000x fold improvement in mobility will improve computing speeds(let's just consider the internal speed/throughput of a cpu) ?

JanineLove
User Rank
Author
Re: Timing & Characterization
JanineLove   3/17/2015 12:30:12 PM
NO RATINGS
Ah. OK, thanks!

R_Colin_Johnson
User Rank
Author
Re: Timing & Characterization
R_Colin_Johnson   3/17/2015 12:01:19 PM
NO RATINGS
JanineLove: The researchers say their next task is characterization at low temperatures to make sure they've got what they were aiming to get--a pure crystalline electron material like Wigner predicted. That done, the physicist will begin optimizing it for specific devices, but they'll have to partner/license to a big-name chip maker to put it together into commercial devices--5 years minimum (I would say).

JanineLove
User Rank
Author
Timing & Characterization
JanineLove   3/17/2015 10:34:11 AM
NO RATINGS
Any idea on how long until this will be to a point that semiconductor manufacturers can seriously look at it? Also, any concerns about characterization issues, do we have the tools to work with this material?

R_Colin_Johnson
User Rank
Author
Re: Faster than a Speeding Bullet. Not.
R_Colin_Johnson   3/16/2015 12:41:33 PM
BOMBOVA: thanks for the link. Yes, there are many other factors involved in the "speed" of a circuit than the electron mobility of its material. Electron mobility is important, but as your link points out it takes a lot more than a fast material to make a fast circuit! Thanks again for the perspective.

BOMBOVA
User Rank
Author
Faster than a Speeding Bullet. Not.
BOMBOVA   3/16/2015 12:10:48 PM
NO RATINGS
i am more a practioner,   i design, after the theory and physics have been discussed.  when i first encountered  ECL,    i thought,    my, that is fast,    i used this link for a refresher :> http://www.engin.brown.edu/faculty/daniels/ddzo/speed.html    

Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed