You probably wouldn't find your cellphone or laptop computer too useful if you had to keep it as cold as liquid nitrogen, which checks in at a little less than minus 320 degrees.
That's why word that University of Illinois professors Milton Feng, Nick Holonyak and colleagues have run their new laser-emitting transistor at room temperature is big news as far as getting the device, which could radically change microchip technology and capability, to market is concerned.
"Until you're sitting at room temperature, you're not really ready to make a hard device that someone's going to sell to someone else," Holonyak said recently. "You're not going to have it at Best Buy and all that."
Feng, Holonyak and postdoctoral researchers Gabriel Walter and Richard Chan report on the room temperature operation of their transistor in Monday's edition of the journal Applied Physics Letters. It appeared in the online version last week.
Holonyak said getting the device to run at room temperature was more a matter of coming up with the right recipe – combining what they know about lasers and about fabricating transistors, for instance – than finding a secret ingredient, a single breakthrough.
"We knew we would get to room temperature," he said. "But what we couldn't tell you back a year ago is would it happen in two years or five years."
Holonyak, a UI electrical and computer engineering and physics professor, and Feng, an electrical and computer engineering professor, emphasized that much work remains before the transistor is ready for prime time.
"There's still a lot of experiments," said Feng, who's developed the world's fastest conventional transistor at 604 gigahertz. "We learn a lot every day."
But the researchers expect big things to happen with the new device, probably beginning with high-speed, high-capacity communications uses.
Holonyak, inventor of the first practical light-emitting diode, characterized the transistor as the first major change in transistor technology, the linchpin of the electronic age, in 50 years.
"Not much has happened fundamentally in a long time, and this is fundamental," he said. "The transistor has now become more than a transistor."
He should know. His mentor was John Bardeen, the late UI professor who won the first of his two Nobel Prizes for the invention of the transistor at Bell Labs in 1947.
Holonyak, who was involved in inventing the first semiconductor lasers in the 1960s, devices integral to CD and DVD players, noted that it took eight years to get them to room temperature, versus the year it's taken with the laser-emitting transistor.
It took two more years to make a reliable semiconductor laser and another decade before they began to be widely used.
"You can't have everything you want right away," said Holonyak, whose light-emitting diodes light up the displays of electronic devices all around us. "You have to give it time and work it into that situation. It's (the laser-emitting transistor) still in its infancy. It's still primitive."
Nonetheless, he and Feng both said they think the laser-emitting transistor could progress faster than some past breakthroughs.
Transistors today are "two-port" devices. A charge goes in and an electrical signal comes out. The laser-emitting transistor, on the other hand, is a three-port device. A charge goes in and two signals come out, electrical and light, which could have a variety of uses.
Moreover, laser light is "coherent." It can be stimulated as needed, finely focused, controlled and used in signaling, as in fiber-optic communications, one potential place for the UI transistor.
Laser transistors also could be used in "signaling" between the tiny components of integrated circuits, that is to say computer chips. Combined with Feng's work on faster transistors in general, the dual output may hold promise for increasingly powerful chips capable of handling more data faster, for a long time to come.
Feng said high-capacity data transfer is one thing that interests the Defense Advanced Research Projects Agency, which is funding the laser-emitting transistor's development.
The idea also has advantages over other possibilities, such as carbon nanotubes, for overcoming the limitations chip makers are facing with current technology.
That's because it can be more easily integrated into existing silicon-based processes for manufacturing and employing integrated circuits, Holonyak said.