Transistor junction, what's your function now? Irish researchers at the Tyndall National Institute have fabricated the world's first junctionless transistor, a nanotech development that could change the way semiconductors are manufactured.
The challenge over past decades has been to keep up with Moore's Law by cramming more and more transistors into the limited real estate provided by silicon chip fabrication methods. But as future tech leans more heavily on smaller, lighter, more mobile devices with increased computing power, the imperative to slim down chip design while increasing efficiency has grown increasingly greater.
Unfortunately, existing transistor junctions -- two pieces of silicon with opposite polarities that allow the current to be switched on and off within the transistor -- aren't all that efficient. Current can leak from junctions, upping power consumption and causing overall inefficiency in devices that increases with the number of transistors. Junctions are also a major factor in driving up costs in the chip manufacturing process; as gateways for current, they are the key mechanisms in transistors, and manufacturing high quality junctions can quickly become very expensive.
The junctionless transistor circumvents the need for junctions by pumping current through a thin silicon wire just a few atoms in diameter. A component nicknamed the "wedding ring" regulates the flow of current by electrically "squeezing" the wire to stop the electron flow, much in the way you might crimp a drinking straw to stop liquid from moving through it. The architecture of the junctionless transistor is simple enough that it can be cheaply produced even at very small sizes, meaning that the tech could contribute to significantly cheaper transistors in future. And because the current is moving straight through a silicon wire, it leaks very little current, making these new structures a good deal more efficient.
Of course, crimping a silicon nanowire to control current in a lab setting and manufacturing these junction-free transistors en masse are two completely different things, and it remains to be seen whether the innovation can scale effectively across the chip industry. But if the technology does indeed pan out on a large scale, the breakthrough could lead to a paradigm shift in chip architecture.
Illustrating this article with a generic picture of a PlayStation processor implies a link to Sony and implies that the new transister might already be in a Sony product or might be much further along in development than it really is (e.g.: the actual transistor is still essentially a one-off in the lab - it probably looks nothing like the surface mount component shown). How about you show a picture of the actual item, the inventors' faces or at least something with a more direct link to the actual story?
'Sorry, 'pet peve of mine. I hate the "Dog saves master" kind of 'news' where the picture is a generic picture of a dog and not the actual dog mentioned in the article copy....
I predict either Intel or AMD will acquire this tech (buying the technology or licensing it) sooner rather than later. It'll be another race between the two rivals (albeit, one's Goliath and the other David). It might be a few years before we see junctionless processors from either camp, though.
But if this tech can be produced en mass with current processor technology, maybe we'll see Moore's Law back in effect and stronger than before.
The outcome is good for the consumer. Better for the environment, too, since less power is leaked so there isn't as high a demand on the old PSU.
Err. Has anyone heard about what is called a field-effect-transistor? Been around from the 60's.
They are fine but they are slow and they are noisy. They generate harmonics and they cause vast amounts of electromagnetic static.
Maybe I am wrong but this seems to be what this article is about.
Just my opinion. I could be wrong.
mgopher/mac mac/original coder of the game donkey kong for the trs 80 mod 1/inventor of the sensor to detect the breakup of molecular cohesion of a polymeric pressure vessel, under pressure, in a vacuum.
mac^2 I think that they classify FETs as having junctions --and maybe there is some rationale to that claim.
The description does sound awfully similar to MOSFETs, but without some more detailed explanation, it's hard to tell.
The source article says that the discovery was made in Ireland. They also said that the fabrication should be quite easy --- maybe with the help of some wee leprechauns :)
Why don't you read the actual paper in "Nature Nanotechnology" . And yes it is like a MOSFET but without any junctions. Spare the world your stereotyping of Ireland, it is really a high technology location and leprechauns are fictious creatures to cater for the tourist tastes of gullible Americans
Plus, Ireland makes the best cereal. I dont know how my day could start without the hearts, stars, horseshoes, clovers, blue moons, pots of gold, rainbows, and not to forget the red balloon.
If this works in a way it sucks, I know IBM is coming close on the light gate technology, but if this is cheap and comparably fast, that might be dropped, but the light gate would still be highly useful, because it would allow chips to be no where near each other and operate in tandem.
It was just a joke, no intent to disrespect the Irish.
As for the junctions, it's a matter of definition. FETs are generally regarded as unipolar, but it's open to interpretation. I did read the source article to this one, but not closely enough to find a link to any in depth paper.
In any event, it's not important. If it's a good idea and it works, hooray. If it doesn't it's still a good effort.
I'm sorry folks, but tech reached its apogee with the points ignition system, and its been in free fall ever since.
I don't think this will be enough of a parachute to keep tech from a nose-down landing. :)
The actual paper can be found at: www.nature.com/nnano/journal/vaop/ncurrent/fig_tab/nnano.2010.15_F5.html
Or by going to the Nature Nanotechnology website and searching for "Colinge" (the first author's last name).
Check out Fig 5 to get a quick idea of how the device works.
There are a lot of reasons why nothing has replaced Si MOSFETs (metal-oxide-semiconductor field-effect-transistors) for the last 50 years. The deciding factor will likely be manufacturing concerns, like throughput, yield and lifetime.
It could even be taken a step further and the note could be printed in the envelope or at least sent with the shipment
just connect 2 diodes together.
Here are some articles which might be of interest:
(the same as above but with a whole extra picture)
The Tyndall Institute is a research center which is part of University College Cork. Woo, my university made it onto PopSci! Watch out MIT!
isnt graphene what scientists should be working on? thats by far the most promising.