IBM Corp. Monday announced progress in the development of a new type of transistor in what can only be seen as the
latest salvo in a shoot-out with rival Intel Corp. over the creation of evolutionary chips geared to bolster
efficiency.
Like Intel’s announcement of last week, IBM claims
to be changing prototypical chip architecture to make headway in improving the efficiency of computer chips, which will eventually
drive down costs.
Unveiled at Big Blue’s microelectronics division in East Fishkill, NY, the “double-gate” transistor can carry twice the electrical
current and possibly operate at up to twice the speed of today’s transistors, which perhaps is not surprising. What is a big deal,
IBM said, is that the processor can be “reduced” in size. The outfit said this is important because “dimensions of the traditional
transistor are reaching the limits imposed by the law of physics,” which will put a damper on revving up performance gains.
Interestingly, this is a similar benefit to what Intel revealed last week with its TeraHertz
transistor threatening to slow the gains in performance that are required for high-speed communications, information systems and
consumer electronics. All of which is to say that the chipmakers are running to preserve Moore’s Law, which states that the number
of transistors on a chip would double every 18 months. Still, pretty much every chipmaker agrees that time is ticking for the extinction of the plain silicon chip, anticipating that transistors will run out of room on wafers in the next five years.
The breakthrough was made with new device designs and materials such as Silicon-On-Insulator (SOI), a material that IBM manufactured that is becoming an industry standard. SOI uses crystal silicon and silicon oxide for integrated circuits (ICs) and microchips. An SOI microchip is generally a third faster than metal-oxide semiconductor (CMOS)-based chips and power consumption is reduced by about 80 percent, which makes them great for mobile devices.
So, how does IBM’s new double-gate schema work? Within transistors, a “gate” feature controls the electrical flow through the transistor, but as transistors continue to shrink, it becomes more difficult for a single gate to control switching. In IBM’s double-gate transistor, the channel is surrounded by two gates, doubling control of the current
The double gate transistor works as a sort of on/off switch as the performance of the chip depends largely on the ability of its transistors to switch on and off quickly and completely, to ease power consumption.
Bijan Davari, vice president of semiconductor development at IBM Microelectronics, said SOI will be essential in changing chip design.
“Other than getting smaller, the basic transistor has largely gone unchanged for decades, but it has now been shrunk nearly to a
point where it will cease to function,” Davari said. “Fortunately, our wealth of experience using SOI allows us to change the basic
design of the transistor, permitting further shrinkage and other improvements.”
To be sure, the double gate architecture was no snap to forge. Big Blue said it made many attempts to make it work to the point where it could successfully manufacture it for sale because of electrical leakages, poor energy flow and taxing energy demands.
In any case, both IBM’s and Intel’s game of processor one-upmanship has been going on for years and both firms will be on hand to show their breakthroughs this week at the IEEE International Electronic Device Manufactures (IEDM) conference in Washington, D.C. Other chipmakers, such as Motorola Inc. and AMD Corp.
will also be on hand to unveil a .10 micron CMOS technology and a .15 micron CMOS chip, respectively.