NASA has developed a new chip manufacturing process

OhioDude

Diamond Member
Apr 23, 2001
4,223
0
0
New manufacturing process developed by NASA uses carbon "nanotubes" instead of copper interconnects. The carbon supposedly doesn't degrade over time like copper and allows for smaller circuitry on the chip.

Posted this over in DC. Thought the folks over here might find it interesting.

Here's the entire press release.

:)

======================================

Michael Braukus
Headquarters, Washington April 14, 2003
(Phone: 202/358-1979)

John Bluck
Ames Research Center, Moffett Field, Calif.
(Phone: 650/604-5026)

RELEASE: 03-136

NASA IMPROVES COMPUTERS WITH TINY CARBON TUBES ON SILICON
CHIPS

The life of the silicon chip industry may last 10 or
more years longer, thanks to a new manufacturing process
developed by NASA scientists.

The novel method, announced in the April 14 issue of the
journal Applied Physics Letters, includes use of extremely
tiny carbon 'nanotubes' instead of copper conductors to
interconnect parts within integrated circuits (ICs). Carbon
nanotubes are measured in nanometers, much smaller than
today's components. A nanometer is roughly 10,000 times
smaller than the width of an average human hair. ICs are
very small groups of electronic components made on silicon
wafers.

"NASA needs high-performance computing in small packages for
future autonomous spacecraft," said Meyya Meyyappan,
director of the Center for Nanotechnology at NASA's Ames
Research Center (ARC), Moffett Field, Calif., coauthor of
the article. "The bottom line is that computer chips with
more layers and smaller components can do more for us. While
we are working on carbon nanotube-based chips for long-term
needs, we also are indirectly helping industry to keep
silicon-based computer chips in use as long as possible," he
said.

One advantage of using carbon nanotube interconnects within
integrated circuits is that these interconnects have the
ability to conduct very high currents, more than a million
amperes of current in a one square centimeter area without
any deterioration, which seems to be a problem with today's
copper interconnects," said Jun Li, lead scientist of the
team at ARC that developed the new process. "Also, there is
no need to create deep, narrow trenches on silicon wafers in
which to bury copper conductors, a step that also is
becoming a problem as components are made smaller and
smaller," Li added.

"Our process allows us to use the tiny carbon nanotubes to
replace copper to interconnect network layers on silicon
chips," Meyyappan said. "We think this new process may well
help to sustain the Moore's Law growth curve."

Moore's Law stemmed from an observation made by computer
chip pioneer Gordon Moore in 1964 that the number of
transistors in a given area of an IC had doubled every year
since its invention. Moore predicted the trend would
continue at a rate of about 18 months between doublings.
Continuing down this 'doubling' path is becoming
increasingly difficult, according to Meyyappan.

"Roadblocks exist in several common technologies such as
interconnects, lithography and others currently used to make
the chips," he said. "However, I think our new process could
be in use by industry for the next generation of ICs,
removing some of these roadblocks," Meyyappan added.

"Using the new process, manufacturers will be able to add
more cake-like layers of components to silicon chips to
increase computer capability," Li said. Because copper's
resistance to electricity flow increases greatly as the
metal's dimensions decrease, there is a limit to how small
copper conductors can be. In contrast, extremely tiny carbon
nanotubes can substitute for copper conductors in smaller
computer chip electronic configurations, because carbon
nanotube electrical resistance is not high.

The new process includes 'growing' microscopic, whisker-like
carbon nanotubes on the surface of a silicon wafer by means
of a chemical process. Researchers deposit a layer of silica
over the nanotubes grown on the chip to fill the spaces
between the tubes. Then the surface is polished flat.
Scientists can build more multiple, cake-like layers with
vertical carbon nanotube 'wires' that can interconnect
layers of electronics that make up the chip.

For more information about NASA nanotechnology on the
Internet, visit:

http://www.ipt.arc.nasa.gov

Sound files suitable for radio broadcast and publication-
size images are available on the Internet at:

http://amesnews.arc.nasa.gov/audio/nanoconnsound/nanoconnsound.html

http://amesnews.arc.nasa.gov/releases/2003/03images/nanoconn/nanoconnpix.html

For more information about NASA on the Internet, visit:

http://www.nasa.gov

-end-

* * *
 

dakels

Platinum Member
Nov 20, 2002
2,809
2
0
OK I am trying to make some sense out of this as the microprocessor/semiconductor noob that I am.
Now I understant somewhat the whole size thing and the ability to make smaller wafers, but the power allowance, does that mean a carbon nanotube generation of processors could handle more voltage? How does that equate to more processing power or clock speeds? Also because it's carbon based wouldn't it be more resiliant to higher temperatures then copper? So could we expect processors that can handle much higher operating temperatures then today's copper based chips? If the carbon based interconnects have less resistence then less resistence = less heat output as well doesn't it?
One last question, something in this R&D phase, even if it did turn out to be 100% effective, probably wouldn't affect the mass consumer market for decades right? Maybe at least 10-15 years? I know all this is hard to guess or estimate, I am talking in relative assumptions.
 

RyanM

Platinum Member
Feb 12, 2001
2,387
0
76
Carbon nanotubes have been the holy grail of many scientists for the last 20 years for several reasons.

They conduct electricity, are extremely rigid and durable, and can be arrayed into structures stronger than any material currently known to man.

In theory, a car whose frame was made out of carbon nanotubes would weigh in literally a thousand pounds lighter. Rockets for space shuttles would weigh in at 1/100th their current weight.

And in theory, you if arrayed properly, anything made out of carbon nanotubes could be a processor. Meaning, as you drive your carbon nanotube car to work, it could be crunching Dnet blocks or processing your MP3 playback.

If NASA has managed to find a way to precisely create and connect nanotubes for usage in microprocessors, this is a HUGE first step to the future of....everything, really.
 

f95toli

Golden Member
Nov 21, 2002
1,547
0
0
I looked at the paper and there is nothing in there about a "new fabrication process", the process they use is well known, they are only using it in a new way. The resistance of the interconnects is still about a magnitude higher than the theoretical value. It is nice work but as far as I can tell it is hardly ground-breaking.

As far as I know NASA is not a major player in the nanotube "business" (IBM is). Secondly the new-release is full of techno-bable about things that are aloready known and even the pictures on their website are not THAT interesting. If they really had made big breaktrough they would not have published in APL but in Nature or Science (no one really reads APL unless you are looking for something in particular).

The main problem with nanotubes is to get them to where you want them to be, right now the only way to do it is to scatter lots of them on surface and then hope that some of them will end up in the right place. There has been some progress in the field of "controlled growth" but we are far from a well-controlled process.

I think this is another case of a over-enthusiastic PR-deparment...
 

zephyrprime

Diamond Member
Feb 18, 2001
7,512
2
81
But is the resistance of nanotubes less than copper? That's the most important thing. I suppose that since the nanotubes are so small they might be able to exhibit ballistic conductance.

Carbon nanotubes have been the holy grail of many scientists for the last 20 years for several reasons.
That would be kinda difficult since c nanotubes were only discovered ten years ago.
http://www.sciencedaily.com/releases/2000/12/001201073219.htm
 

VirusTwin

Junior Member
Apr 16, 2003
1
0
0
I read an article a few months ago about the possibility of carbon nanotubes being used in future processors and it stated that the carbon nanotubes needed a "sticky end" to join them together to form transistors and gates. I wonder how they overcame this obsticle?
 

RyanM

Platinum Member
Feb 12, 2001
2,387
0
76
Originally posted by: zephyrprime
But is the resistance of nanotubes less than copper? That's the most important thing. I suppose that since the nanotubes are so small they might be able to exhibit ballistic conductance.

Carbon nanotubes have been the holy grail of many scientists for the last 20 years for several reasons.
That would be kinda difficult since c nanotubes were only discovered ten years ago.
http://www.sciencedaily.com/releases/2000/12/001201073219.htm

Hrm. Perhaps I'm confusing them with that other carbon nanostructure, the buckyball. I may have gotten their discovery dates mixed up.

Ah, yes, the buckyball was discovered in 1985. 18 years ago. Close enough. ;-)
 

digitalix

Member
Apr 18, 2003
31
0
0
Carbon nanotubes are just a whole new can of worms, so many potential practical uses, no economically feasible way to mass produce it yet and lots of money to be made :)
 

f95toli

Golden Member
Nov 21, 2002
1,547
0
0
I read that a new japanese factory will be able to produce 40 tons of C60 (Buckyballs) a year selling it at $4/gram.
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
all said is true but to address the awesome potential of teh catbon nanotube processor. it could feasable handle two or three times the voltage and still operate cooler at 100 times the speed. however to corectly structure carbon nanotubes into a computer processor would take precision so advanced that the process is a decade away from feasability. but to use them in the constuction of things like cars, or bikes or something like is very feasable if you don't expect them to do anything but if you want those carbon nanotube frames to do anything you are back to making a corbon nanotube processor, granted much bigger one which creats room for error in construction but still. and as for bucky ball being made in a factory in japan that is great but what the hell are gonna use them for besides research that will not be complete for two decades. it has been two decades since we found the buky ball and it will be two decades before we can utilize them.
 

f95toli

Golden Member
Nov 21, 2002
1,547
0
0
I was also wondering why they are building this factory. My guess is that they are going to use doped buckyballs in some sort of opto-electronics, but that is just a guess.
I figure if a large japanese company builds a factory they must have a good reason, right? Maybe they know something we don't.
 

MadRat

Lifer
Oct 14, 1999
11,977
294
126
Something about using nanotechnology and lacing buckyballs with precise amounts of medicine. Saw it in Discovery magazine that they have a future in medicine.
 

chemicalplay

Member
Jun 17, 2001
71
0
0
i saw something about Carbon 60 talk at a uni, c60 is produced in very small amounts and rolled into a tube, to create a much smaller, more conductive type of material (opposed to copper) which can decrease processor size, increase data transfer time and basically push forward microprocessor technology. Its all good, they just have to somehow mass produce this stuff. Cos it is made by passing a lazer thru uncut diamond fragments which produces a whole load of stuff, which must somehow be seperated easily from the c60.