Single Molecule Transistor A Reality

Petersko writes "A team from the University of Alberta has proven for the first time that a single molecule can switch electrical currents off and on, a puzzle that scientists worldwide have been trying to crack for decades. The finding could revolutionize the field of electronics, providing a leap ahead for everything from computers to batteries to medical equipment."

Comments

How long until it's usable?

[LordKazan]

That's freaking sweet, but how long until we see this filtering into usable technology?

Re:How long until it's usable?

[rbarreira]

From TFA:

While a computer using this new technology is at least a decade away

...

Re:How long until it's usable?

[lcnxw]

I would predict that now that it has been determined to be possible it will be a while before big companies pick up this technology and begin testing with it. TFA didn't mention the cost of manufacturing products using this technology, or the reliability and durability of the technology. at such a small scale if a molecule was shifted even a very small amount, it could render the device unusable. Also, I don't know of any methods of mass producing these types of nano-technology. It could be quite a while (10 years or more) before we see any application of this technology in consumer devices.

Re:How long until it's usable?

[grammar fascist]

While a computer using this new technology is at least a decade away

It's always "a decade away" for these people, just like it's always 30 years for the AI people. Estimates like that seem less an informed guess and more an expression of confidence.

By the way, your sig:

Vader:You're either with me, or my enemy/Bush:You're either with or against us/Obi-Wan:Only the Sith deal in absolutes

So you caught Lucas's sorry attempt at political commentary?

Yoda: Do, or do not. There is no try.

Huh.

Re:How long until it's usable?

[QuantumG]

It's funny ya know. There once was a time where if you were a researcher at a university and you discovered something like this you'd actually go form a company and make a billion dollar industry. Now people have so given up on turning research into products that they just publish their results and hope one of the big boys takes their work and turns it into something useful.

Re:How long until it's usable?

[HillaryWBush]

So you caught Lucas's sorry attempt at political commentary?

Will anyone catch yours?

Re:How long until it's usable?

[Doc Ruby]

AI has not been proven experimentally in a lab, needing only industrialization engineering to make the science a reality.

BTW, just because Lucas' simple philosophy is understood by most before we finish elementary school doesn't mean that George Bush ever understood it. Or that it's wrong, even though Bush might say different. But of course, since the facts are biased against Bush, he needs your help to prop up his hideous propaganda.

But at least you wear your worldview on your sleeve (or your .sig):

"In mathematics you don't understand things. You just get used to them. -- John von Neumann"

"In politics, grammar fascist doesn't understand things. GF just gets used to them.." - Doc Ruby

Re:How long until it's usable?

[Nogami_Saeko]

Or in this case, by RTFA, you find out that "Wolkow and the University of Alberta have filed for a U.S. patent on the technology.".

So universities that actually don't want to screw-over their researchers, and want researchers to actually hang around, rather than leaving to keep their intellectual property will do a joint-patent of some sort.

N.

Other sources

[waynegoode]

Also read about it at the University of Alberta website and in the Press release

Re:Other sources

[Husgaard]

Or have a look at the home page of the researcher.

but how fast is it?

[LiquidEric]

It seems if it were a C8H10N4O2 molecule it would switch much faster.

Re:but how fast is it?

[the_atomsmith]

Yes, but would the chip get the jitters after a while? C10H15N would make it switch even faster although side-effects may include irrational behaviour and violence against other molecules.

The article in Nature

[waynegoode]

Also see the article in Nature.

Go Bears!

[SpaceAdmiral]

Whooo! Yeah! Rock on! Go UofA!!! Oh, wait, is my Alumni pride supposed to be limited to sporting events? Also, UofC sucks.

Really slow device

[Husgaard]

The researcher admits to that: "It takes us on the order of minutes to change conditions that make current go or not, so for any computer technology, this thing is today impractical."

Still I think this is very interesting news. This is very early research. The speed will probably be improved, and the smaller dimensions of single-molecyle transistors can give space for more hardware to compensate for the speed.

Re:Really slow device

[TopSpin]

The speed will probably be improved...

That statement doesn't mean that the device is slow. It only says that it takes the researchers a long time to establish the necessary conditions. The odds are that the device, having vanishingly small mass, can switch at very high frequency. Imagine if you were asked to operate an ordinary light switch using the passenger side rear corner of a dump truck while blindfolded. This is analogous to what it's like to manipulate individual molecules with an STM.

Re:Really slow device

[M. Baranczak]

Imagine if you were asked to operate an ordinary light switch using the passenger side rear corner of a dump truck while blindfolded.

Sounds like a great drinking game.

If you don't know...

[Ghoser777]

I believe that's the molecular structure for caffeine.

Kind of mediocre article

[ZorbaTHut]

So efficient is this potential new technology, said Wolkow, that "the question now about the battery life in your laptop would go away. Your battery today would run your computer all week or all month instead of three to four hours."

Of course, by the time we *can* build CPUs with this technology, we'll be able to build the equivalent of your current laptop into a watch or a cellphone - and the new generation of molecular-CPU laptops will be the same size, massively more powerful, and run for three to four hours. Doh.

Not only that, but because the microelectronics could eventually be made out of molecules, some computer parts could be biodegradable since molecules can be broken down into small bits.

"Made out of molecules"? What do you think they're made out of now? Rainbows and unicorns?

That said, this is damn cool. Miniturization is unstoppable! (At least until these molecular transistors become used in everything - I'm not quite sure where we'll go from there.)

Re:Kind of mediocre article

[ghutchis]

There are a couple avenues open, such as using light instead of electrical current. Electrons move around very slowly compared to c.

Meh. Photonics is really hard to do on the scale of a chip. You could multiplex easier, but remember that with current CPUs, the wires are much smaller than the wavelength of visible or infrared light.

Plus, electrons aren't traveling very far on a chip. So even if you get the photons working, you're not saving a lot of time per gate.

Lots of work, little advantage = little reason to switch to photonic computing.

Now if you're talking networking or telecommunications, then you've got something. Bet we'll see all-optical switching at some point.

Re:Kind of mediocre article

[Jafafa Hots]

How is a current-tech battery supposed to power a molecule-transistor laptop for a month when they won't even hold their charge that long sitting unused on a shelf?

Homegrown IC boards

[FlynnMP3]

I can't help but think that in 20 years or so that we'll be able to use the latest design inkjet printers, pop special 'inks' in them, load in IC plans into Photoshop IC, hit print and 2 hours later have usable expansion cards for our computers (sans power connectors). Need more memory or replace bad sticks? Just download the plans for your particular machine and away you go. Wonder how much that is going to cost for the IC plans? Or better yet, what about those evil _hacker_ people who design and release IC plans onto the Internet for free?!? Are they nuts? (sarcasm)

It's coming. Oh boy, then manufacturer's are going to be fighting for their way of survival just like the **AA of today. Fun fun fun!

From TFA

[Hockney Twang]

"It's the biggest thing that we've ever done."

Holy shit, if this is the biggest thing, I can't wait to see their other work.

MoneyMoneyMoney

[1967mustangman]

I can see the Dollar Signs rolling infront of the eyes of the board of trustees. They are going to make a mint off of this!!!!!!! Cha-Ching

Cool, but NOT Revolutionary

[ghutchis]

I work in the field of molecular electronics -- I'm sorry, but this doesn't sounds "revolutionary."

It's hard to comment before I've read the article, but there are a lot of other, very reliable single-molecule transistor experiments. In 2002, Nature called it a "discovery of the year." (Sorry, can't find the URL right now.)

There have been pretty good single-molecule transistor measurements in other groups since then.

Granted, if they're able to image the single molecular wire, that's a solid advance over other techniques. But it's hardly the solution to a 20-year old puzzle.

(By the way, it's more like 30 years since it was shown how a molecule could function as a switch. The first paper on the subject was published in 1973.)

-Geoff

Rainbows and Unicorns

[ghutchis]

"Made out of molecules"? What do you think they're made out of now? Rainbows and unicorns?

Chips aren't made out of molecules. Current semiconductors are made out of various forms of silicon crystal.

That's a lattice -- there aren't individual "silicon molecules" anywhere in there.

Just FYI.

yep, single-molecule transistors made previously

[jedibo29]

I work on sort-of-related stuff in condensed matter physics, and I have skimmed through some papers about the subject, and I had the same thought as the parent (though I was too lazy to post) - that it's cool, but not quite "revolutionary" and that other groups have come up with single-molecule transistors before.

Here's a link describing what two groups published in Nature back in 2002 about single molecule transistors (maybe what the parent post was referring to):

http://www.sciencedaily.com/releases/2002/06/02061 3073522.htm/

I briefly scanned through the Nature "News and Views" that introduces and supplies background for this recent paper (from today's Nature) and it looks like the main innovation is that the group used a different technique from earlier molecule transistors. According to that Nature article, previous groups relied on metalmoleculemetal types of molecular junctions. Apparently, these junctions have extra geometric complexity due to the metal-molecule coordination that can complicate the charge transport measurements. This group avoided the "geometric uncertainty" by using a junction on a semiconductor (silicon) instead of a metal. They apparently added a carbon atom with an unpaired electron as a 'dangling bond' on the surface of the silicon electrode to form the molecule-electrode interface. The interface is then a covalent bond, instead of the more complicated coordination bond from other molecular junctions. Anyway, that's why I gathered from a quick reading of the Nature summary article, but I don't know about the subject in much depth.

There's definitely a lot of potential with this stuff, though.

Re:I WONDER

[mol_transistor]

interesting discussion. our paper actually goes well beyond the HP work mentioned above. in fact, if you read further on the HP work, you will see that they eventually discovered, as many suspoected, that they had no (active) molecules in their device. the rotoxane, while a brilliant achievement of chemical synthesis, played no role in the observed current voltage spectra. so detailed characterization is very important. in the work we published today, for the first time it is possible to say definitively that one molecule is characterised - not more or less than one. we know where every atom is and we know the congifuration of the molecule. the other key to this work is that we have managed to make two electrodes serve as three. that is a crucial step as there isn't room to have three fine probes converge on a volume the size of a molecule lke styrene. the key is in the ability to charge one single silicon surface atom, with one single electron, near the one single molecule. the result is like a conventional field effect transistor. it is a robust effect that works at room T - unlike say the break junction work at cornell. i hope that helps answer some of the questions.