Science & Technology - Posted by Anne Ju-Cornell on Thursday, December 10, 2009 13:56 - 8 Comments 
(4 votes, average: 4.75 out of 5)
Loading ...
(4 votes, average: 4.75 out of 5) Loading ...Tiny transistor may become conductor king
A microscopic image of a small gallium nitride-based device, which can be easily scaled up for higher current handling capability. Below, a larger version of a gallium nitride transistor. Researchers at Cornell University believe the device could form the basis for the circuitry in products from laptops to hybrid vehicles to windmills to other power electronic systems. (Credit: Junxia Shi)
CORNELL (US)—A newly developed and extremely efficient transistor may soon replace silicon as the semiconductor of choice for power applications.
Share this article
Email
Republish this article
The text of this article by Futurity is licensed under a Creative Commons Attribution-No Derivatives License.
Researchers believe the device could form the basis for the circuitry in products from laptops to hybrid vehicles to windmills to other power electronic systems.
The patent-pending electrical switch is made from the compound gallium nitride, a material with unique electrical properties that Lester Eastman, the John Given Foundation Professor of Engineering at Cornell University, has been studying for more than a decade.
Details about the recent breakthrough were published in the journal Applied Physics Letters.
The new transistor’s on-resistance, or measure of resistance to electric current, is 10 to 20 times lower than today’s silicon-based power devices and has a high breakdown voltage.
High breakdown voltage is a measure of how much voltage can be applied across a material before it fails.
Researchers say the key to the device lies in gallium nitride’s low electrical resistance, causing less power loss to heat, and its ability to handle up to 3 million volts per square centimeter without electrical failure.
Silicon, a competing material, can handle only about 250,000 volts per square centimeter.
At the heart of improving electronics, Eastman says, is the ability to make devices that can switch electricity from high voltage to high current, which is a measurement of electrical applicability, while minimizing power loss.
“Power has to go from A to B in a machine with a high voltage transmission line to minimize power loss,” Eastman explains.
“Before now, there were no electronic devices that could handle both high current and the high voltage, but our device can do it.”
Eastman believes the transistors, which were made with Cornell nanofabrication equipment, might one day power everything from hybrid electric vehicles to Navy destroyers.
In fact, the U.S. Navy first funded Cornell’s research into gallium nitride transistors more than 10 years ago and is a major funder of Eastman’s research today.
In next-generation electrical devices, “you want to have the power that’s coming out to be not much less than the power that’s going in,” Eastman says. “This is the best material we know of that can do this conversion without loss of energy.”
The device was developed by Junxia Shi, a graduate student in Eastman’s lab. The New Jersey-based company Velox and Motorola spinoff Freescale have also helped fund the research, with the hope of producing the devices at an industrial scale.
Cornell University news: www.news.cornell.edu
Please wait
8 Comments
Marco
Andy
@Marco: a patent only gives the patent holder the right to prevent others from practicing the invention. The patent holder may choose not to exercise that right, may grant a license for no cost, or may charge a licensing fee. Assuming they charge a fee, a reasonable expectation, history has shown that this is not an impediment to further R&D. The semicon industry is crowded with patents on every aspect of process technology and device architecture, and I haven’t noticed any letup in the pace of development. Cross-licensing arrangements are commonplace, and are built into the economics of R&D.
emc2
Go Navy. You have to wonder, if it is this useful, whether there will be enough of it (gallium nitride)
GaN has been studied for many years, and is currently the semiconductor crystal that creates blue LEDs. Nitronex sells GaN on silicon wafers for HEMT and LDMOS apps. Like all compound semiconductors, GaN single-crystals are more difficult and expensive to work with compared to silicon, and so will be limited in use to niche applications such as discrete power amplifiers, LEDs, and money-is-no-object military/aerospace ICs (the few applications where Si cannot get the job done). As a comparison, recall that GaAs ICs remain niche chips after 25 years of work, though Skyworks and others have made nice businesses out of expoiting such niches.
Tektrix
@emc2
Gallium is extracted from bauxite. (see, Mikolajczak, C, Availability of Indium and Gallium, Sept. 2009 http://www.indium.com/_dynamo/download.php?docid=552). Gallium is present in bauxite at around 50 ppm (http://www.mii.org/Minerals/photogallium.html). The reserves of bauxite are huge at ~27,000,000,000 metric tones (http://en.wikipedia.org/wiki/Bauxite). Currently, only about 10% of the available gallium is extracted from bauxite. Would seem there is a lot of it available.
Gray Gaffer
I’m a little confused about the “3 million Volts per square centimeter”. I have not seen this metric before. I can believe a metric of a voltage gradient of 3×10^6 V/cm i.e. vertically through the junction material, but it does not seem right to me that surface area should affect the breakdown voltage.
I would also like some indication of scale in the operating voltage vs max Ic current, and also capacitance effects related to switching speed and swing amplitude. There’s no point raising the operating voltage beyond the power consumption of today’s standard Si based technologies, which is why they are heading to less than 1 volt operating after having started out at 5 volts, when the power consumption is so closely related to operating speed x operating voltage x node capacitance.
Of course, this is moot if the targeted system for the technology is green power and charging control for vehicles, which needs high voltage and high power tolerance far more than it needs speed and density.
wjv
Mr. Gaffer:
the volts per area measurement given here is a rather standard metric for semiconductor materials because they are almost always manufactured in disks/slides (i.e. flat and with small uniform thicknesses). It is more typical to give resistance per unit area (a.k.a. sheet resistance) however these are proportionately related via ohm’s law so its preference really.
Sheet resistance is simply resistivity with the thickness dimension neglected for simplicity. This metric is convenient because it turns an extensive property (resistance) of a material into an intensive property (sheet resistance or resistivity) so that one can compare apples to apples with various semiconductor materials.
Read more at: http://en.wikipedia.org/wiki/Sheet_resistance
Buk
I wonder if this material functions as well after an electro magnetic pulse? Might be one of may reasons the military might be funding it. Maybe it doesn’t “fry” as easily as the current semi conductors under those conditions.
I’m confused, how is patenting this going to make the design be adopted faster if it really is the replacement for silicon and the future of electrical devices?