Can Calcium Fluoride Handle New Demand?

Daniel C. McCarthy

Although the demand for semiconductor lithography equipment decreased in the past year, system manufacturers struggled to advance their technology closer to the 193-nm and 157-nm nodes. For calcium fluoride (CaF₂) suppliers, that climate put the emphasis on performance and quality, rather than on yields and quantity. With signs of a possible recovery later this year, however, that could change, and the availability of CaF₂ could again become an issue.

The slowdown in the semiconductor industry influenced capacity buys of CaF₂ lens materials, but not technology buys, because lithographers needed the material primarily for developmental, not commercial, steppers. If the market rebounds later this year, so will demand pressures on CaF₂ supply. Courtesy of Schott Lithotec AG.

Lithography tools at the 193-nm node use calcium fluoride for a small number of lens elements, but the percentage will rise as numerical apertures increase. That will draw from the supply pool needed for exposure lenses in 157-nm development tools, which rely entirely on calcium fluoride material. Demand for these lenses will increase significantly in 2003 and 2004 as the industry's road map approaches the 157-nm lithography node.

The supply chain for CaF₂ materials has evolved to keep up with rising demand. In 1999, both Corning Inc. and Schott Lithotec AG significantly expanded their CaF₂production and, in 2001, nearly doubled their output again: Corning at its facility in Canton, N.Y., and Schott at its factory in Jena, Germany. That same year, Japanese toolmakers Canon and Nikon announced plans to build production facilities for CaF₂.

Corning's Canton expansion is only in its initial phase, according to Gitimoy Kar, director of the company's fluoride crystals business in Corning, N.Y. "There will be two more over the course of two years, each modifying the technology based on lessons learned from 157 development and 193 manufacturing," he said.

Konrad Knapp, vice president for corporate business development at Schott Lithotec, suggested a similar time frame for his company's next expansion. Schott links the timing to the release of high-volume 157-nm tools, which Knapp said would very likely happen in the next 12 to 18 months.

Suppliers won't be idle in the meantime. Some of the lessons Kar mentioned stem from tight specifications for crystal-axis orientation in CaF₂ lenses. Controlling crystal orientation is linked to controlling the material's intrinsic birefringence, and, ultimately, that is an optical design issue. But its ripples will be felt throughout the production cycle, from lens blank manufacturing and lens polishing to coating and mounting.

The learning curve for these problems might help explain why providers like Schott see some hesitation in commitments for long-term supply agreements. Or maybe that hesitation is due to uncertainty in the market. Either way, the short- to midterm demand (and supply) looks very good, Knapp said.