IBM, Motorola, Japan firms agree on common mask for x-ray lithography

05/01/1997

IBM, Motorola, Japan firms agree on common mask for x-ray lithography

In an effort to standardize x-ray mask technologies and streamline their development, IBM, Motorola, and four Japanese firms have agreed to use silicon carbide (SiC) and tantalum as mask membranes and absorber materials.

The new unofficial standard, which succeeds the use of a silicon substrate and gold absorber, answers some of the unresolved issues in maskmaking and brings x-ray lithography one step closer to mainstream production. Pat Hughes, IBM`s mask technology development manager based in Burlington, VT, said that after more than a year of collaboration, the two US companies met recently in Vermont with Toshiba, NEC, Mitsubishi, and Nippon Telegraph and Telephone (NTT). All agreed in principle, he said, to set SiC and tantalum as the standards, though some Japanese companies are still using a tungsten absorber.

Both IBM and Motorola, which have formally collaborated on x-ray lithography development since 1994, will begin using the silicon carbide-tantalum combination in test production this month. IBM plans to use the new materials on test masks for 1-Gbit test sites, with features of 0.18 ?m and below. Similarly, Motorola will work with the new materials on advanced logic test sites, Hughes said.

The two companies have the only substantial industrial x-ray research program in the US. Most research is conducted at IBM`s Advanced Semiconductor Technology Center in East Fishkill, NY, but some work is performed at IBM`s Burlington, VT, location and at Motorola`s Austin, TX, site. The standardization is a significant move for the US companies, because they do not have the same level of experience with the new materials as their Japanese counterparts. Toshiba, NEC, Mitsubishi, and NTT have been exploring the use of SiC and refractory metals for the past two to three years. The four firms have been researching x-ray lithography at a synchrotron facility operated by NTT.

Until recently, IBM and Motorola had been using a silicon membrane and gold absorber in their research program. The more rigid SiC membrane offers better overlay performance than silicon membranes.

Perhaps an even bigger step for IBM and Motorola is the shift to tantalum. Hughes said IBM has not had experience with the refractory metal. But what makes it so enticing is that it permits the use of a subtractive etching process, one familiar to today`s wafer lines. "It brings forth a toolset that is in common use with wafer technology," Hughes said. "For IBM, that is a significant change." Tantalum has other promising qualities that make it more attractive than gold, which can contaminate a wafer line and has long been thought to lose its absorbing abilities below 100-nm feature sizes.

It remains unclear when x-ray will make it into full production. Hughes said neither IBM nor Motorola has current plans to bring x-ray lithography into manufacturing. But, in December, a Mitsubishi researcher said x-ray would "very likely" be used for second- and third-generation shrinks of the 256-Mbit DRAM generation.

"I really feel the key issue here is not technology, but cost," said Peter Heinz, president of Suss Advanced Lithography, the primary developer of x-ray wafer steppers. He points to a study compiled by several Japanese firms that shows x-ray lithography with a better cost-of-ownership than other advanced technologies such as 248-nm DUV and e-beam. A SEMATECH group was expected to release a report in April comparing costs of x-ray to other competing technologies.

But as, industry observers point out, x-ray will remain just off the horizon as long as optical lithography is useful. "What`s really holding it up is that there`s no need for it," said Franco Cerrina, director of the Center for X-ray Lithography at the University of Wisconsin. "The question is, when is optical going to stop?" - Christine Lunday, WaferNews