"Demonstrating this caliber of result in a practical environment means that as our collective client base moves to next-generation technology by using the 'gate-first' approach, they will continue to maintain a significant competitive advantage," said IBM Vice President Gary Patton.

Small, Fast and Efficient

Executives from IBM, Chartered Semiconductor Manufacturing, Freescale, Infineon Technologies, Samsung Electronics, STMicroelectronics and Toshiba believe that high-k/metal gate technology will be able to serve as the basis for achieving a long-sought improvement to the transistor. Foundry support for customers implementing chips featuring 32-nanometer circuitry will be made available in the third quarter, they said.

However, Gartner Vice President Dean Freeman notes that the switch to high-k/metal gate technology did not come easy. "The integration of the high-k dielectrics with silicon has been challenging, to say the least," he said, adding that technical issues led device manufacturers "to push out high-k materials just one more generation, until now."

The new material's use in a critical portion of the transistor -- where the primary on/off switching is controlled -- will enable chips to become smaller, faster and more power-efficient than previously thought possible. And industry observers say the new technology could soon move into many consumer devices, especially those with critical power-consumption and battery-life requirements.

"The alliance enabling sampling at this early date will allow companies to begin producing devices in 2009 for commercial consumption, which is about the same time Intel will be moving into production with 32nm," Freeman said.

The Biggest Challenge

The chip changes taking place in high-k dielectrics and metal gates are just the first of many technical innovations that Freeman expects, with the next major changes likely to occur at 22nm.

"Logic transistors will likely move to multi-gate technology at this time" and it is possible "we will see the same transition" for memory, Freeman said. "NAND flash will likely have a change to the tunnel dielectric and the floating gate technology at 22nm, possibly earlier, he added.

One of the biggest challenges, Freeman said, is that some of these solutions may last only a generation before a new material or structure is required to take the industry to the next level. "Each new transistor generation from 45nm on down will need a redesign," Freeman said. "It is possible to use the current hi-k materials through 32nm, and they could work at 22nm, but to do so the cell will need to be redesigned, so in this case a multi-gate structure."

The current high-k materials will eventually have a limitation, Freeman noted. "The question that the industry doesn't have an answer for yet is when," Freeman said.

The ultimate goal, Freeman explained, is to move electrons as fast as possible using the least amount of power, while generating the least amount of heat. "That is what the researchers are focusing on for the next generation of transistors," he said.