Editor's Page

300-mm pacesetters

The evolution to 300-mm wafers, what SEMI President Stan Myers calls "one of the most costly technological transitions in industrial history," is accelerating faster than anticipated. Major elements of the tool set are either being evaluated by the I300I and SELETE consortia or will be submitted for appraisal by year's end. With panelists from SEMI, the two international consortia, chipmakers, wafer manufacturers, and equipment companies, last month's well-attended 300-mm symposium held in Grenoble, France (in conjunction with Semicon Europa), offered a good progress report.

The coming change is driven more by economics than technology, as George Lee, SEMI's 300-mm initiative director, pointed out. Productivity factors include as many as 2.7 times more die per wafer compared to the 200-mm size, improved perimeter efficiencies, higher output per wafer starts, lower per-die cost (as much as 40% better than 200 mm, according to Lee), and fewer fabs needed to meet demand. Since the magnitude of the 300-mm transition precludes setting too many standards after tools are in full production, standardization activities are being aggressively pursued. Lee noted that standards are already in place for wafers, carriers, and interfaces.

Many pieces of 300-mm equipment have become available faster than had been forecasted as recently as the middle of last year. Komiya-san from the Japanese SELETE consortium said 35 pieces of 300-mm equipment would be evaluated inside the group's Yokohama facility by the end of 1997, with 70 to 80 tools scheduled for appraisal by the end of 1998. He also disclosed that 15 analysis and measurement systems are already in place at the SELETE fab. Frank Robertson, general manager of the multinational I300I consortium, said he expects at least 46 equipment demonstrations to take place this year at the group's Austin, TX, fab, with a full tool set in place by late 1998. He mentioned that 300-mm wafers from seven suppliers are being characterized and that his group expects to have one of the key process tools, a 248-nm lithography unit, in place this fall. Robertson also revealed that, based on input from the consortium's 13 member companies, he and his colleagues are finalizing plans to extend I300I's 18-month mandate.

Although three companies had officially aired plans for 300-mm facilities at press time, off-the-record head nodding confirmed imminent announcements by several other key players. Bob Doering of Texas Instruments (one of the companies which has publicized plans for such a plant) underscored some benefits mentioned by Lee, adding that utility, gas, and chemical costs for 300-mm fabs should be about 25-40% higher than those for 200-mm factories, figures that track below previous estimates. Equipment and silicon supplier panelists from Applied Materials, Wacker Siltronic, Jenoptik, and Steag all emphasized the need for less-customized, more-standardized tools and for increased sharing of precompetitive technology between chipmakers and vendors to keep the lid on costs that could spiral to as much as $30 billion industrywide. The supplier representatives also cited automation, tool integration, process optimization, and, in the case of Wacker, large-volume crystal growth as major challenges.

With industry insiders saying at least nine 300-mm pilot lines should be functional by the end of the decade, the semiconductor sector is adding its own fuel to the already roaring bonfire of premillennial expectations.

TOM CHEYNEY

EDITOR

tom.cheyney@cancom.com