backtop


Print E-mail del.icio.us 9 comment(s) - last by ViroMan.. on Aug 28 at 7:24 AM


Previous vapor deposition methods left a rough surface which was difficult to work with and destroyed fragile organic molecules attached to the silicon substrate on which it was deposited.  (Source: Coll Bau/NIST)

NIST's new organic molecular switch construction technique not only results in a plausable product, but a much smoother and friendlier gold surface.  (Source: Coll Bau/NIST)
A practical molecular switch may be in sight thanks to NIST research.

The inner world of semiconductors is constantly shrinking, producing faster microchips with more transistors and multiple cores in the same or an even smaller space than the previous generations. As a result, technology is starting to struggle to keep up with Moore's Law. Small breakthroughs are being made at a good clip by continuously improving technologies like optical computing and using new materials like graphene, but no one is sure where the next great advancement will come from.

With semiconductors firmly entrenched in 40 nanometer production fabs and shooting straight for the low 20s, there doesn't seem like there's very much room left to keep shrinking. However, before the leap to quantum multi-state or even electron spin computing, there still lies the frontier of single molecules. Using single molecules as the tiny switches in microchips has been an idea kicked around for several years, but as NIST materials scientist Mariona Coll Bau explains, trying to assemble such chips using the standard metal vapor deposition technique does not pan out well; "Imagine what hot steam would do to your arm. Evaporated metal is much hotter, and organic switching molecules are very fragile—they can’t stand the heat."

Rather than attaching the molecules to a silicon surface and then depositing gold vapor and destroying the molecules, the NIST and University of Maryland team reversed the process. They first deposit the gold onto a super-smooth non-stick surface. This process actually has another benefit aside of not wreaking havoc on the fragile molecules -- previous deposition techniques created a very rough metallic surface which could either destroy the molecules or cause them to not connect to the conductor at all.

After the gold has cooled, they apply simple overhead transparencies as a laminate. This allows them to peel the gold from the non-stick base while leaving the smooth metal surface intact.

They then attach the molecules to be used as the switches to the gold conductor's surface and flip the whole thing over onto a new silicon base. An imprinting machine is used to compress the layers together without crushing the switches.

And this technique may not be useful for just future microchip manufacturing. Coll Bau indicated that it could also be useful for applications like biosensors, where the electronic world meets the organic to perform certain tasks.



Comments     Threshold


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

Moore's law
By bighairycamel on 8/27/2009 10:04:29 AM , Rating: 1
I realize they are trying to keep up with Moore's law, but even he knew there would be a theoretical ceiling. Having single molecules act as transistors has to be the limit unless they manage to find a way for single atoms to hold a charge.

The tech sounds awesome but i can't help but think it's a little ironic that they are trying to keep up with Moore's law while simultaneously reaching it's limit.




RE: Moore's law
By mmcdonalataocdotgov on 8/27/2009 11:40:20 AM , Rating: 3
I think they are trying to keep up with Moore's law since it is primarily a statement of business practices, and not a physical law, per se.

Moore thought the limit would be reached with 65K transistors on a single wafer by 1975. And they have been exceeding this on dies until now. I am not sure what wafer size he had in mind then, or the number of dies per wafer (DPW) they could yield back then, but I think they have far exceeded what Moore thought was possible in 1965.

I would like to see larger dies, myself. This would reduce yield, but if you could design multiple asics across acres* of wafer material, I think you could get a super computer on a "chip," as it were.

*for example, not for comment. =)


RE: Moore's law
By Jaybus on 8/27/2009 1:12:07 PM , Rating: 2
quote:
I would like to see larger dies, myself. This would reduce yield, but if you could design multiple asics across acres* of wafer material, I think you could get a super computer on a "chip," as it were.


That is essetnailly what Intel's Larrabee is. Moore's Law refers to the trend over the past 40 odd years of transistor density to double every 2 years. Transistors can continue to shrink down to the atomic level, but as we have seen, thermodynamic limits are reached before size limits. So future increases in density will have to be used to expand the number of cores, 100% on-die memory, integrated i/o, etc.

My favorite Moore quote is
quote:
Moore's law has been the name given to everything that changes exponentially. I say, if Gore invented the Internet, then I invented the exponential.


RE: Moore's law
By exNihlio on 8/27/2009 11:43:29 AM , Rating: 2
That is when biological and quantum computing will come to fill the gap when we hit that ceiling.


"Let's face it, we're not changing the world. We're building a product that helps people buy more crap - and watch porn." -- Seagate CEO Bill Watkins

DailyTech Poll
Which web browser do you use on your primary personal machine? 






44 Comments












botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki