In a breakthrough that may lead to high-rise processors, Lucent Technology's Bell Labs claims to have found a way to build a chip's components vertically instead of horizontally.
Dubbed one of the "Holy Grails" of the semiconductor industry, the ability to manufacture chips vertically may give current chip-making technology a new lease on life, Bell Labs said.
The chip industry is widely believed to be approaching the physical limits of how small chips can be manufactured.
Thanks to the laws of physics, the ability to make ever-smaller processors using current photolithographic techniques will likely hit a brick wall in the first decades of the next century, industry watchers say.
As a result, some of the world's top research labs are looking at building processors chemically at the molecular level, an endeavor that is only just now getting started but may take decades to mature.
However, new manufacturing techniques like Bell Labs' that allow components to be built on top of each other may give today's semiconductor manufacturing technology some breathing space.
Co-developed by Bell Labs researchers Jack Hergenrother and Don Monroe, the new design allows a chip's components -- the transistors -- to be built vertically instead of horizontally, a technique that allows many more to be crammed onto a chip.
The breakthrough may eventually lead to 3-D chips, where components are stacked on top of each other. However, the researchers warned that such designs are a long way away.
"We think there might be ways to use the technology," said Monroe. "Eventually we might be able to make a kind of two-and-a-half dimensional chip. But at the moment we don't quite know how."
As well as building transistors upward instead of outward -- a research goal pursued for the last 25 years -- Hergenrother and Monroe figured out how to create two gates per transistor instead of one, a development that may double the processing power of future chips, they said.
The design should also make it easier to use thinner insulating materials, a problem currently the subject of intense research in the industry.
Furthermore, the manufacturing process makes use of current techniques and shouldn't be prohibitively expensive, the team said. They estimated that the technology could appear in less than a decade.
"We're building on the strengths of the silicon semiconductor industry," Monroe said.
"You have to understand it's still at the research stage and there are critical questions to answer," added Hergenrother. "We're projecting five to ten years."
Using Hergenrother and Monroe's techniques, the transistor's gates measure just 50 nanometers, which is about 2,000 times smaller than a human hair.
Future refinements should shrink transistor gates to less than 30 nanometers, the researchers said. By contrast, today's transistors measure about 180 nanometers.
Keith Diefendorff, editor in chief of the Microprocessor Report, said the technology may work in the lab but the real test is making it work in a chip factory.
"If it's real and you can manufacture it, it might provide the kind of breakthrough we need to get through the 0.1 micron barrier," he said. "They have a good reputation on these things. The question is implementing it in volume manufacturing."
Hergenrother and Monroe will present their findings at next month's IEEE International Electron Devices Meeting in Washington, DC.
