March 10, 2004 10:43 AM PST
Can nanotech keep chipmaking up to speed?
As a field of study, nanotechnology has uncovered ways of creating and manipulating materials that could be applied to manufacturing chips such as processors and memory chips, said Juri Matisoo, vice president of technology at the Semiconductor Industry Association.
Matisoo, who was speaking at the Nano Science and Technology Institute's Nanotechnology 2004 trade show here, urged chipmakers to view the field as a potential source of new technology and to include it in their research and development as soon as possible.
That's because the Semiconductor Industry Association (SIA), a trade group that represents chipmakers around the world and maintains a 15-year chip industry outlook, is predicting that today's manufacturing techniques may stop finding advances after four--or possibly, five--more generations of chips, at some time around 2011.
Cutting-edge chips now come in sizes of about 90 nanometers. (A nanometer is a billionth of a meter.) Without manufacturing changes, it may prove next to impossible to shrink silicon-based chips below the 22-nanometer level--a size expected to make a debut in about seven years. That's because electrical forces become more difficult to circumvent as the size of the chip decreases, Matisoo said.
Economics will also weigh in, he said. Chip manufacturing plants, which now cost between $2 billion and $3 billion to build, will increase in price, following Rock's Law, which suggests that chip plant construction costs double every four years.
The chip industry is "coming to hard limits--not only technical limits, but economic limits. So we need to make changes," Matisoo said.
Those potential barriers have Matisoo and the SIA urging chipmakers to explore the use of nanotechnology to develop methods of increasing performance and of lowering manufacturing costs as well. Technically, companies like Intel are already in the nanotechnology business, because their products are built from sub-100-nanometer components. But with further investment, nanotechnology could let chipmakers manipulate materials on the atomic or molecular level and thus produce new properties.
But time is short. The SIA is advocating more-extensive nanotechnology research, right now, Matisoo said. He urged chipmakers to enter into research that explores the potential synergies between nanotechnology and traditional silicon-chip manufacturing.
"What one would like, in the ideal grand-challenge world, is to obtain real performance improvement and get rid of some of these problems," Matisoo said.
If research by some of the presenters at the Nanotechnology 2004 conference bears out, one way forward could be the creation of hybrid chips that use traditional semiconductors as a base with structures built through nanotechnology--such as transistors--on top.
Two potential sources of transistors for future chips are nanowires, tiny strings of silicon atoms that are being investigated by companies such as Nanosys, and carbon nanotubes, which are essentially tube-shaped lattices of carbon atoms. A number of companies, such as IBM, have been experimenting with carbon nanotubes.
"If we are to continue significant advances in information tech, we need something new," Matisoo said. "In some senses, we're already a little late in looking at new ways of doing things."