IBM will get some help pushing the envelope on chip production technology from the U.S. government.
IBM Microelectronics' Essex Junction, Vermont, plant announced it will receive $18 million in federal funding for the research and development of X-ray lithography chipmaking. The grant is part of the fiscal year 1998 Department of Defense appropriations bill now waiting on President Clinton's signature.
The funding confirms IBM's important role as a chipmaker. Besides taking part in X-ray lithography research, the Essex Junction plant produces the PowerPC microprocessor.
It also underlines growing interest in the search for a new manufacturing technology. Earlier this month, the Department of Energy headed up the launch of a government-industry consortium that will look into using using "extreme ultraviolet" waves, or EUV. $250 million will be devoted to the effort.
Semiconductor manufacturers rely on ultraviolet light to "mask" or map out circuit lines on silicon chips. The current process yields chips with lines .25 microns across, and will be used to make the next generation of chips, with lines as thin as .18 microns. Thinner lines allow more transistors to be placed on a single chip, yielding smaller and faster chips.
But ultraviolet is reaching its physical limitations because the mask lines can't be made thinner than the wavelength of the light making the images.
IBM has been experimenting with X-ray lithography as a next-generation manufacturing process for ten years, according to Jeff Couture, a spokesman for the company's Advanced Mask Facility in Essex Junction.
The federal government has been funding the facility for the last five years, since the Defense Department's Advanced Research Projects Agency (ARPA) first identified the technology as important to the national interest. The facility has received yearly funding roughly equal to its fiscal 1998 grant since then, Couture said.
Demonstration masks of .18 micron chips have been made in conjunction with a sister facility in East Fishkill, New York, Couture said, and production could begin early in the next century. IBM hopes to eventually develop a .10 micron process.
But the transition from demonstration to production masks for x-ray lithography or any of the competing technologies will be complicated, said Dataquest analyst Klaus Rinnen.
Great costs attend getting a research process to production. "It's a long way from printing one test pattern to making millions of patterns day in and day out," Rinnen said.
More important, currently no industry consensus exists about ultraviolet's successor, and research into X-ray, EUV, ion projection, and three other technologies continues apace because Sematech, a semiconductor industry consortium, is due to begin winnowing the field of candidates by the end of the year. "Everybody's trying to ensure there's enough information to make an educated decision," Rinnen noted.
Pointing to the recent announcement of a European consortium devoted to the ion projection process, Rinnen concluded, "A decision, a global decision, has to be made."