backtop


Print 34 comment(s) - last by Autisticgramma.. on Dec 11 at 3:28 PM

Researchers see new device as path to sub-10 nm circuits

Employing an ultrathin dielectric composed of a 4-nanometer-thick layer of lanthanum aluminate with an ultrathin, 0.5-nanometer layer of aluminum oxide, Purdue University's nanowire transistor of indium-gallium-arsenide (IGA) reaches an important milestone of a 20 nm gate size.

Currently Intel Corp. (INTC) uses a 22 nm process for its Ivy Bridge silicon-based transistors.

The new IGA transistor, like Intel's fin shaped 3D transistors, employs a three-dimensional gate design, but it takes it even a step farther, creating a bizarre stackable design of triple-tapering nanowires that looks like a tiny pine tree.  

Peide "Peter" Ye, a professor of electrical and computer engineering at Purdue University, has an interesting name for his new device -- "the 4D transistor".  He comments, "A one-story house can hold so many people, but more floors, more people, and it's the same thing with transistors.  Stacking them results in more current and much faster operation for high-speed computing. This adds a whole new dimension, so I call them 4-D."

4D Transistor
A series of "4D" transistors [Image Source: Purdue]

He says the superior electron mobility of the new transistor allowed the novel design, and may allow even more ambitious successors.

The new work was published in a pair of papers [PDF] to be presented at the International Electron Devices Meeting on Dec. 8-12 in San Francisco.

Currently the silicon chipmaking industry is in an uncertain state.  14 nm chips are expected for 2015, while researchers hope to shrink to 10 nm by 2018.  But past 14 nm, leakage in current "high K" dielectrics will become to severe for the transistor to operate; hence to stay on course for 2018 researchers must race to discover new dielectrics.

Squeezing past 10 nm will be even trickier, as it's pushing the boundaries of the already strained optical lithography techniques.  Advanced techniques like self-assembly or mechanical manipulation of atoms may prove crucial at features sizes below 10 nm.

Sources: Purdue, Eurekalert



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

The scale of these always amazes me
By Solandri on 12/6/2012 7:09:17 PM , Rating: 2
That 40 nm line you see in picture (b) is about 170 silicon atoms wide.




RE: The scale of these always amazes me
By JKflipflop98 on 12/6/2012 9:30:25 PM , Rating: 2
Pfft weaksauce. I see smaller stuff than that every day.


RE: The scale of these always amazes me
By MadMan007 on 12/6/2012 10:28:40 PM , Rating: 5
But what's in your pants has no semiconductor applications.


By GreenChile on 12/6/2012 11:17:15 PM , Rating: 2
BURN!


By Alexvrb on 12/6/2012 11:19:08 PM , Rating: 5
Well... the article did say that mechanical manipulation of atoms may prove crucial past the 10nm scale.


RE: The scale of these always amazes me
By AnnihilatorX on 12/7/2012 8:42:55 AM , Rating: 2
The transistor in question is a indium-gallium-arsenide semiconductor, not silicon. I know what you mean though.

InGaAs semiconductors has been touted and touted to be much better than silicon in terms of performance, yet we never see it come mainstream, instead they are still specialised in high power / high frequency radio/communication circuits. Now graphene is touted to be the next gen. I am not sure if this research would see any real world use.


RE: The scale of these always amazes me
By Sivar on 12/8/2012 12:38:56 PM , Rating: 3
I work in the semiconductor field.

Gallium Arsenide is not likely to ever take over silicon other than for fringe products. While it has some advantages:

1) It is highly toxic.
2) It is rarer and much more expensive.
3) Silicon's thermal conductivity is almost 3x that of GaAs. It doesn't matter if you have smaller, faster transistors if you can't pack them even close to as tightly as with silicon.


By superstition on 12/9/2012 4:16:29 PM , Rating: 1
Are people going to eat the chips?

I bet there's less arsenic in a 20nm chip made this way than in a carton of eggs laid by chickens fed roxarsone feed.

Guess where that pretty pinkish color inside chicken at Indian restaurants comes from.

The expense doesn't seem as critical given the fact that the gallium indium arsenide is only being used in one layer.

"It doesn't matter if you have smaller, faster transistors if you can't pack them even close to as tightly as with silicon."

Well, this is a design that uses more than just gallium arsenide. It is conceivable that additional materials might be used that would improve conductivity. Building transistors at the atomic level provides a lot of options for dealing with the drawbacks.


By Autisticgramma on 12/11/2012 3:28:10 PM , Rating: 2
Additionally, significant Indium deposits are quite rare. (see commercially viable) The majority of indium captured as a derivative of zine manufacture, and is used for touch screens, because it can be made clear, and hold a charge.

It isn't anywhere near as available as silicon.

http://en.wikipedia.org/wiki/Indium

Cool stuff, just not revolutionary.


Surprised...
By EricMartello on 12/7/2012 12:52:56 AM , Rating: 3
...that Jason called it a "christmas tree" shape and not a "holiday tree" shape. LOL




RE: Surprised...
By inperfectdarkness on 12/7/2012 2:39:24 AM , Rating: 2
PC fail. "holiday" is a euphamism for "holy day" which in itself is un-PC. The word you're looking for is "Seasonal Tree."

/trollface


RE: Surprised...
By Jeffk464 on 12/7/2012 12:56:33 PM , Rating: 2
uhm, how about pine tree.


RE: Surprised...
By Sivar on 12/8/2012 12:39:57 PM , Rating: 3
That would be showing favoritism towards Stanford.


RE: Surprised...
By xthetenth on 12/7/2012 11:34:39 AM , Rating: 2
Haha war on Christmas. Nobody cares about going after Christmas except fox news. Speaking as an atheist who feels rather strongly about evolution denial like Jason does, Christmas is pretty awesome. Get together with family, show your appreciation for each other and generally have a good time. Some people do celebrate other stuff, and recognizing that is good, but there is absolutely no reason to go after the holiday itself and especially one of the co-opted pagan symbols that are pretty non religious by this point.


RE: Surprised...
By Jeffk464 on 12/7/2012 12:58:14 PM , Rating: 2
Just about the whole holiday is a co-opted pagan celebration.


RE: Surprised...
By lagomorpha on 12/7/2012 4:16:23 PM , Rating: 2
Personally I think Saturnalia sounds like is was a lot more fun before it became christianized. Gag gifts, social positions reversed, chaos...


RE: Surprised...
By Concillian on 12/8/2012 2:36:46 PM , Rating: 2
At my daughter's kindergarten class, they have a "Unity tree"

Bleh. I'm not Christian, but go ahead tell me Merry Christmas. I won't be offended, and I will wish you one right back.

People (on both sides of the issue) need to lighten up about the whole thing.


RE: Surprised...
By lagomorpha on 12/8/2012 5:52:08 PM , Rating: 2
quote:
People (on both sides of the issue) need to lighten up about the whole thing.


Have you ever met a single person that actually cared about calling it Christmas or have you only met people who were overly concerned some other hypothetical person would be offended?


RE: Surprised...
By superstition on 12/9/2012 4:18:31 PM , Rating: 2
Some religious types are very concerned. My mother, for instance, would get upset if my sister or I wrote X Mas on anything.


RE: Surprised...
By Lord_Conrad on 12/10/2012 5:59:17 AM , Rating: 2
I care very much about keeping Christ in Christmas. Most people today who call themselves "Politically Correct" wouldn't dare say anything against Islam or Judaism, but have no problems trampling all over Christianity. I see this all the time in network TV.


Nope
By Hector2 on 12/6/2012 7:00:23 PM , Rating: 5
Still 3D




RE: Nope
By gfxBill on 12/6/2012 9:38:08 PM , Rating: 2
But electrons require time to do any work, so... 4D!


RE: Nope
By Strunf on 12/7/2012 7:38:06 AM , Rating: 2
By that definition then transistors have been 4D since the very beginning...


RE: Nope
By Jeffk464 on 12/7/2012 12:54:54 PM , Rating: 2
Yeah, that's what I was thinking.


yule tree transistor
By Silver2k7 on 12/6/2012 6:24:46 PM , Rating: 2
Its interessting news and nice timing for pine tree shaped yule transistors :P

But the fourth dimension. Isn't that supposed to be time? The 4D in the name feels wrong.




RE: yule tree transistor
By sixteenornumber on 12/6/2012 6:58:51 PM , Rating: 2
quote:
The 4D in the name feels wrong.


I couldn't agree with you more. Adding another field to the already 3d transistor and calling it 4d sort of makes sense but it's horrible at the same time. This is along the same lines as describing computers with the word "cyber" NO!


RE: yule tree transistor
By inighthawki on 12/6/2012 9:42:17 PM , Rating: 2
quote:
But the fourth dimension. Isn't that supposed to be time?

Depends on which theories you subscribe to. Some people call time the 4th dimension and wrap it all up nicely together as space-time, some theories don't.


RE: yule tree transistor
By SPOOFE on 12/7/2012 7:10:23 PM , Rating: 2
Time is a dimension, just as length/width/height are dimensions, but a "dimension" is just an aspect of a thing, usually but not exclusively in reference to some measure of that thing. "How hot" in relation to hot sauce is a dimension, or "body temperature" another dimension if you're some criminologist or something and trying to determine a time of death.

L/W/H and T are just extremely common dimensions as it pertains to gauging and observing our existence, and "three-dimensional" is the way we describe the sort of space we apparently seem to inhabit (with time as "the fourth" to allow for perpetuity of that space).


RE: yule tree transistor
By Concillian on 12/8/2012 2:51:25 PM , Rating: 2
quote:
But the fourth dimension. Isn't that supposed to be time? The 4D in the name feels wrong.


I suspect reading the actual paper and being more familiar with the area of research would result in context that would suggest they are not talking about the traditional spatial relation "dimensions".

As an example, "dimensions" are unlimited in matrix mathematics. It's only when specifically discussing spatial dimensions that you are limited to 3D.


200 nm
By knightmike on 12/6/2012 11:19:37 PM , Rating: 2
According to the picture, these things are 200 nm wide.




RE: 200 nm
By hamstern on 12/7/2012 5:10:40 AM , Rating: 2
200 nm... but evil.
They look like Daleks. Another case of narrow sighted scientist thinking they invented something for the 'greater good'.


RE: 200 nm
By superstition on 12/9/2012 4:20:29 PM , Rating: 2
No. If you notice, that 200nm picture shows more than one of them. So, they don't need to be 200nm, individually.


Uhm
By EnzoFX on 12/6/12, Rating: -1
RE: Uhm
By ppardee on 12/10/2012 3:25:52 PM , Rating: 1
More like Bee Hive shaped, really.


"I f***ing cannot play Halo 2 multiplayer. I cannot do it." -- Bungie Technical Lead Chris Butcher














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki