Researchers compete to speed up chips

British and U.S. researchers say they can squeeze more components on to microchips.

2 min read
New technologies developed by research teams in both the United States and the United Kingdom may be leading to radical advancement in microchip technology by piling a substantially higher number of components on to a single chip.

Researchers at Bell Labs, the research and development arm of Lucent Technologies, today announced that they have developed an electron beam system that would be able to place four times as many transistors on a chip of the same size as today's smallest chip.

The more components a single chip squeezes in, the more it can do and the faster it can run. Manufacturers now use visible light rays, which have a relatively long wavelength and therefore result in larger components.

According to Dick Muldoon, a Bell Labs spokesperson, the new technology would mean that a chip that today holds 16 million bits of information would hold 8 billion bits of information in the same amount of space.

Scientists at Leicester University, however, say they have developed a X-ray lens that is able to place one hundred times the number of components, according to a report in New Scientist magazine.

The university teamed with Massachusetts-based Nova Scientific to devise an economically feasible way to use x-rays to etch transistors on a chip. The lens is still in development, but even after it's finished there will be one more barrier to the use of X-rays for microchip manufacturing: researchers must still develop a substance that will react properly to X-ray exposure in place of the silicon that now covers the chip.

Although Bell Labs' electron beam can etch fewer components, it has the advantage of working with silicon wafers.

Manufacturers are rapidly approaching the limits of optical lithography, as today's chip-making process is called. Whichever technology wins out, the pair of announcements gives hope that "Moore's law still has legs," Muldoon says, referring to an 1964 observation Intel co-founder Gordon that said that the amount of data stored on a given amount of silicon had roughly doubled every year since the technology was invented.

Muldoon projects, however, that the new electron beam technology won't appear until about 2010, and X-ray technology might be further out. Not so much because the technologies aren't workable, he added, but because manufacturers don't want to pay to upgrade their factories until existing processes are exhausted.

"The economics of the industry are such that manufacturers want to get as much out of their equipment investment for the current and each succeeding technology as they can. We're about four generations out from seeing chips that use this [electron beam] process," he cautions.

Related story:
Toshiba quadruples chip memory