Potential 100 GHz Carbon Nanotube Molecular Computer

leb writes: "Researchers at Harvard have developed "a concept for molecular electronics exploiting carbon nanotubes as both molecular device elements and molecular wires for reading and writing information. Each device element is based on a suspended, crossed nanotube geometry that leads to bistable, electrostatically switchable ON/OFF states. The device elements are naturally addressable in large arrays by the carbon nanotube molecular wires making up the devices. These reversible, bistable device elements could be used to construct nonvolatile random access memory and logic function tables at an integration level approaching 1012 elements per square centimeter and an element operation frequency in excess of 100 gigahertz." Unfortunately to read the entire article in Science Online you need a subscription, but you can still take a look at the abstract."

Comments

Available via NewsNow

[bstadil]

I know this is off topic (sorry) , but anyone knows why this Slashdot posting has been available via NewsNow for a few hours but does not show up on the main /. site. Seems kind of silly we need to go somewhere else to read /. the latest /. postings.

Re:Available via NewsNow

[MostlyHarmless]

This article was posted in the science section and deemed "not important enough" for the main page. You can still access it through older stuff or the Science slashbox. This probably either was because of its slightly vaporous quality or because of the required subscription.

--

Scientific American June 2000

[mr.ska]

Scientific American just had an article much like this, which you can read here without a subscription. They did mention Buckytubes, but didn't focus on them specifically.

Re:They had better hurry.

[Anonymous Coward]

No link or anything, but this is what I do for a living. Speed used to be limited by transistors, so it would double everytime we scaled the dimensions by a factor of 2. However, the intrinsic carrier transport delays of transistors have become so short by now that the RC delay of the interconnects between devices (which gets even worse with higher numbers of devices/chip) is controlling the overall circuit speed. Copper reduces the R a little and low-K dielectrics reduce the C a little but these are only 1-time boosts. Some people are looking at vertical integration (i.e. not all the transistors on one plane on the surface of a wafer) which will shorten the interconnects but this is basically going to be a show-stopper sooner or later. - Derek

They had better hurry.

[Black Parrot]

If conventional architectures are able to keep up with the standard "double in speed every 18 months", it will only take them about 10 years to hit the 100GHz mark too.

Will this technology ripen within 10 years?

--

Re:Problems with cooling.

[still_nfi]

Notice the word "reversible". Equates to 0 power consumption. Neat little concept that one.

No subscription necessary

[f.money]

I just clicked on the download a PDF reprint, and got the text of the article. No login or anything. Jon

Re:No subscription necessary

[ephraim1]

Some educational institutions have what amount to bulk subscriptions when accessed from within their domain, so that many articles can be accessed without an 'individual' subscription. That may be true of other articles or other domains as well. Even as an academic subscriber through my institution I find the occasional full feature article that needs a 'full individual' subscription for download. Thankfully, not this one. :)

Re:But when?

[driftingwalrus]

> Sounds like a bunch of researchers are in need of > more funding to me, so they try to make promises > of something that probably will never happen.

You have to realize that these technologies take a lot of time and effort to develop, things do come of them but you don't always hear reports that mention the original announcement. Consider the blue laser, I remember reading about the development of the blue laser in Scientific American back in 1994. Wasn't until much later that the technology finally saw it's full use in DVDs.

Problems with cooling.

[Valar]

The problems with useing this kind of thing as memory on a fairly low scale device (i.e. not supercomputer) is that a) they are very fragile, and b) they require tremendous amounts of cooling for the parts to remain suspended. If you tried this kind of system in a computer without a liquid gas cooler or some similar device, it would cease to function.

Re:But when?

[saider]

I thought DVDs used visible red diodes. This was an improvement over CDs which used infrared. Most of the improvement in capacity of the DVD comes not from the shorter wavelength, but from the improvements in controlling and focusing the laser. Old late 70's technology was pretty sloppy at this. The size of the features on the CD had to be very large so that the primitive devices could reasonably discern the features.

But you are right about time to market. Once you develop a technology, you have to figure out how to mass produce it. IBM can move individual atoms around, but until they mass produce the equipment to do that, nobody is going to be building things atom by atom.

Re:But when?

[drinkypoo]

Actually, they both use visible red; CDs use 780nm, and DVDs use 635-650nm. (http://www.microserve.net/~tpetchy/DVD .shtml)

The blue laser diode developed by Shuji Nakamura of Nichia Chemical Industries Limited, Tokushima, Japan, can be made for wavelengths between 390nm and 440nm.

Shuji says that "In 1991, the 3M company announced the first II-VI-based blue laser diode. Since that time, II-VI has been very popular for the development of blue laser diodes. However, the degradation of II-VI-based blue laser diodes and LEDs was so large that nobody succeeded in commercializing II-VI-based blue laser diodes and LEDs. Presently, the lifetime of II-VI based green laser diodes (Sony) is only 100 hours under CW operation at room temperature. The wavelength is still green, not blue." His "InGaN MQW laser diode" has a peak wavelength of 404.3nm.

He goes on to say "The main application is digital versatile disk (DVD) for optical data storage. Currently, with a red laser diode you can store 4.7 gigabytes/side on these newly developed compact discs. With a blue laser you should be able to store 15 gigabytes/side, three times as much. Other applications include laser printers and laser full color displays."

Finally, a comparison stolen from CMP techweb:

I'm only going to include the wavelength factor and not compression or improved tracking. They have it wrong, anyway, saying you can only get 7 albums on a normal CD, when we all know it's 11 or greater (Using mp3 compression.)

860nm (Near infrared) is capable of handling one album on a CD.
635nm (Red) can handle two.
430nm (Blue) should be good for four albums on a single CD.
350nm (Ultraviolet) will theoretically let us stash 9 albums on a CD, just by virtue of wavelength alone.

This of course all assumes that we're going to be able to make CDs that have the holes punched in them at the right frequencies; I suspect we'll move on to some more realistic technology, perhaps extra-fine-grade CD-R or CD-RW (Probably CD-R for distribution anyway.)