Skip

QuIET is the New Loud
November 28, 2006 10:14 AM   Subscribe

University of Arizona physicists have discovered how to turn single molecules into working transistors. The research could result in much smaller, more powerful computers and other devices with the ability to process many more channels of high-resolution audio and video than current products can manage. The abstract is available in PDF.
posted by terrapin (17 comments total)

 
My favorite part was the last paragraph:

"We're not futurists at all and can't predict it, but imagine that you could make an artificial intelligence, that you could have this little submarine that goes inside somebody's arteries and capillaries to repair them," Stafford said.

Honestly. The man fantasizes about creating artificially intelligent nano-subs into human circulation? Has he not seen T2 or The Matrix? We don't want that. We don't want artificially intelligent nano subs in our bodies do we?
posted by Mentallo The Brain God at 10:38 AM on November 28, 2006


What is this, Slashdot? "Impossibly Obscure Scientific Discovery Announced"
posted by Brian James at 10:38 AM on November 28, 2006


1. This was a homework problem in my undergraduate honors physics class, in 1995. I don't think I solved the problem, but several other undergrads (freshmen, mostly) solved it.

2. It sounds like they didn't actually make the device. They just speculate that if everything works as expected, the device will work as expected.

3. Yawn.
posted by b1tr0t at 10:40 AM on November 28, 2006


Also, this has nothing to do with audio, sound and video. Quantum Transistors are pure binary devices. Either you get an electron wave-packet out the end of it, or you don't. You don't get the same smooth scaling funciton that an analog transistor produces, so this wouldn't be useful for making an analog audio amp.
posted by b1tr0t at 10:42 AM on November 28, 2006


We don't want that. We don't want artificially intelligent nano subs in our bodies do we?

We don't?

Crap. How do I get these things out?

I liked them at first. It was neat them torpedo-ing cancer cells and all. But the constant tiny little klaxon horns blaring! And the "DIVE! DIVE! DIVE!" at all hours! Then I really got worried when they started singing "It's a long way to Tipperary" in German.

I tried miniature depth charges to get rid of them but then they played dead laying in my lower GI tract all quiet— jettisoning all sorts of garbage to fool me. Shitting diesel and tiny cafeteria trays is no party let me tell you.
posted by tkchrist at 11:23 AM on November 28, 2006 [2 favorites]


We don't want artificially intelligent nano subs in our bodies do we?

If they are crewed by the likes of Raquel Welch, abso - uh, well, maybe not. Or maybe. I'm torn.

Now we can expect Intel floating-point bugs to turn into Intel fissionable bugs. Low recovery rate from that one.
posted by CynicalKnight at 11:25 AM on November 28, 2006


More single molecule nonsense. Furthermore, a fucking theory paper, published in Nano Lett. Aromatic annulenes have been around for about, ooh, I don't know about 80 years.

THEY HAVEN'T DISCOVERED ANYTHING.

What they've done is say the symmetry of an existing molecule under theoretical conditions can act as a transistor if a bunch of disturbingly simplistic modelling assumptions about lead-aromatic bonds work and you can make devices with this kind of physical and electronic symmetry.

Please stop believing press releases put out by university PR departments. This is PR.

Here's a quick credibility test:

Replace in your press release:

"Theory about XXX " with "If we could all just feel the love",
and

"the possibility exists" with "then the spawn of gullible imbecilles like myself"

and

"of solving problem XXX" with "might have the ability to stop war and we'd live happily forever fucking after in clover with bunny rabbits and puppies".

And then try and post it. Gaaaah!
posted by lalochezia at 1:25 PM on November 28, 2006


This is fascinating research. I recognize that picture, too (the one of the Benzene ring) - the prof in my semiconductors course has it splashed on the course website.

However, this research is far from the only research going on in single-molecular transistors, and they haven't solved any major problems; it's just an incremental improvement, which is really only interesting to device physicists and electrical engineers who can see the implications. This 'breakthrough' seems to be a university press-release for PR purposes; the article ismostly fluff and daydreaming.

Still nanoelectronics is a fascinating topic and could make a decent post if more comprehensively surveyed.
posted by PercussivePaul at 1:27 PM on November 28, 2006


Also, this has nothing to do with audio, sound and video.

I first read about this in Electronic Musician and they seem to think it does have audio applications. The article (not online) states that while other researchers have succeeded in attaching two leads to such molecules, the UA team expects three-lead devices to be "just around the corner." And that three-lead devices will allow one lead to act as a valve for current flowing between the other two.

However, I am neither an audio engineer nor a quantum physicist.
posted by terrapin at 3:21 PM on November 28, 2006


Thanks to comments, did not rtf press release.
posted by creeptick at 3:27 PM on November 28, 2006


terrapin, that's pure marketing spin.

It is true that three-lead devices will allow one lead to act as a valve for current flowing between the other two- this is basically how a transistor works. We know how to make transistors quite well out of silicon and metal - there are several hundred million of them inside every Pentium, for example, and one prominent computer science once calculated that transistors cost less per unit than printed characters in a New York Times newspaper. They are the basic building block of any piece of electronics, including audio visual equipment. For audio applications that require analog signal processing, precision is vital. Transistors are carefully designed to produce a good frequency response. A preqrequisite for this is a mature manufacturing process.

There are a few reasons for investigating single molecule transistors rather than stickign with our traditional methods. The most obvious is their amazingly small size, which could help us increase the amount of transistors in one chip and thus increase what we can do with one device - imagine a super-cellphone with unlimited processing power and storage and you get the idea. Also because the technology is such a radically different approach to the problem of electrically manipulating information, some of the problems that currently affect conventional systems, such as prohibitively high power density, could potentially be sidestepped completely (though I'm less certain of this point).

However all of the work with molecular electronics is strictly experimental. For example, some of the most successful devices to date do not work at temparatures larger than 3 or 4 Kelvin (-270 C). There is nothing approaching a mature manufacturing process and there is no reason to expect anything like this within the next 5 years. Talking about audio applications or nanobots or anything like this is pure daydreaming.

There is a very good reason for such PR - it generates buzz, which generates funding, which is how research survives. It's one of the ugly but unavoidable parts of academia.
posted by PercussivePaul at 4:13 PM on November 28, 2006


without having rtf links, I'll just chime in with PercussivePaul and say that the University of Arizona certainly is shameless when it comes to research dollars. at the expense, of course, of educating its students.
posted by carsonb at 4:26 PM on November 28, 2006


Let me clarify: the article is mostly PR. I read the article and I even skimmed the paper they wrote and I think the research itself does seem to be quite significant. It does have the potential to be a breakthrough. It's just not relevant outside the field (yet).
posted by PercussivePaul at 4:39 PM on November 28, 2006


hey, neat. i was just at a conference a couple weeks ago with 2 of the guys in the picture (had lunch with the young guy), and stafford gave a talk about their research.

anyway, re lalochezia:

there is already a great deal of work out there (both theoretical and experimental) on two-junction molecular devices that have diode-like current-voltage behavior.

a guy at yale named mark reed is quite famous in this area for measuring conductivity across a single molecule, back in 2001. that may not seem like long ago but the world of research is really quite fast-moving and these techniques are now well established.

these devices are usually similar to what's shown in this article, in that they are a single molecule containing one or more benzene groups, strung across a narrow gap in a gold film called a break junction; to achieve (for example) diode-like behavior the molecule is generally designed to be asymmetric in some way.

i can't imagine there's much more to making a three-terminal break junction; it may have already been done but i don't know. alternatively one could make a regular two-terminal break junction and then bring another contact in close with a mems-like manipulator or AFM tip or something. it doesn't sound particularly difficult.

beyond that an implementation of this is just synthetic organic chemistry to make the molecule. one ethanethiol group and a couple of extra sulfur atoms on the right spots doesn't sound particularly difficult to me, and it's probably already in beilstein or something.

anyway i don't know that there's anything "disturbingly simplistic" about their assumptions (in fact, i can't even tell what said assumptions are; can you? most of these devices are modeled using a method called density functional theory which works very well in predicting the behavior of two-terminal molecular devices.) maybe you'd like to justify this statement?

i think dismissing what is fairly solid theory as "they haven't discovered ANYTHING" is a little presumptuous. by that logic, einstein didn't discover anything either.

the real issue here is turning a molecule that acts like a switch into a device that's useful; this means assembling the molecules in some kind of meaningful way and coming up with some sort of scheme for addressing them in order to do computations. however a three-terminal molecular device is nothing to dismiss and certainly a big step forward in this field.
posted by sergeant sandwich at 5:10 PM on November 28, 2006


also, re PercussivePaul:

all of the work with molecular electronics is strictly experimental. For example, some of the most successful devices to date do not work at temparatures larger than 3 or 4 Kelvin (-270 C)

i'm not sure where you've gotten this impression from, but it's simply not true. most molecular electronics i've come across work at room temperature (and indeed, sometimes in ambient conditions!) though i do agree that anything resembling a commercializable technology isn't here yet.

also, sorry, i was wrong, the mark reed paper was in science from 1997, not 2001 - pdf here if you've got access.
posted by sergeant sandwich at 5:20 PM on November 28, 2006


Hmmm... perhaps that figure was referring to a specific class of device. It came up during an undergrad class a while back but I don't remember the specifics. I'm clearly wrong. Consider it withdrawn.

I could, however, spout off a few *correct* reasons why molecular transistor manufacturing will not be commericalizable anytime soon and I could talk a good deal about the incredible sophistication of traditional electronic manufacturing processes - but why bother since you agree with the point anyway.
posted by PercussivePaul at 5:38 PM on November 28, 2006


OK. So I went a bit OTT in my pre-caffienated rage at PR nonsense.

I hereby recant my complete dismissal.

My rant refers to the fact that they've taken a fairly well known class of molecules, and modelled some reasonably unsurprising behavior using DFT. (if memory serves, some derivative of DFT actually). DFT incidentally does not do well linking real-world bulk metals to single molecule organics. This is not a discovery with a capital D.

einstein is a straw man. Please: I know theories are discoveries. This one, to paraphrase: I know general relativity , and sir, you are no general relativity.

You're both right: the issue of single molecule work comes down to a) organization of components and b) device stability and robustness and c) cost. None of which have been addressed by this paper (not that it was the aim of the paper to do so).

Apologies for derangement: some of us dread people hyperbolising about unimpressive science, as it makes selling our small, non-miraculous, hard-won advances that much harder to do with a straight face and not gving in to the juggernaught of "it must be huge or it doesn't lead".
posted by lalochezia at 11:15 AM on December 3, 2006


« Older This Film is Not Yet Rated   |   Whatchoo Talking 'Bout, Willis? Newer »


This thread has been archived and is closed to new comments



Post