The Rocky road to 300 mm
06/01/1997
The rocky road to 300 mm
Stretching ahead of us is a road plotted out for 300-mm wafer development. We are not sure exactly which route it will take, how wide it will be, or where it will eventually end. It will be "under construction" for quite a while with detour and "left lane closed ahead" signs and maybe a few "travel at your own risk" indicators. And although it will be a new road, potholes will likely develop quickly. Some may not take the new road immediately, but use side streets to avoid the risks of a highway still under construction.
But those vying to reach the destination first are showing an intrepid spirit. They are already engaged in the battle of press releases, with each company announcing plans to begin construction of 300-mm fabs. SEMI predicts nine 300-mm pilot lines by 1999. Competition is building between US and Asian companies, each wanting to be the first to achieve the yield efficiencies and anticipated revenue largess from producing and marketing chips from 300-mm wafers.
Many obstacles, however, obstruct the highway to megawafer heaven. One is the battle of standards, e.g. Japan`s open cassette concept vs. I300I`s closed environment. Another is readiness of available wafers, materials, and equipment for development activities to run true pilot lines before a full production fab is built (a variation of the chicken-and-egg problem). Added uncertainties are the unpredictability of the demand curve for chips in the next 10 years, and the readiness and ability of the technology to accommodate the feature-size reduction juggernaut simultaneously with retooling to 300 mm.
During the recent SEMICON/Europa conference in Geneva, an associated 300-mm conference was held in Grenoble, sponsored by the Grenoble Chamber of Commerce and Industry. At this session, industry leaders, including SEMI, Texas Instruments, SGS Thomson, and Steag, participated in a 300-mm Q&A period.
The conference was laced with optimism, but many of the problems mentioned above arose. It was pointed out that with more dice/wafer, fewer actual fabs will be needed to make more types of chips. What this means in terms of materials, equipment, and unit price remains to be seen. Attendees were also anxious about timing - whether metrology, wafer availability, and the lithography tools will be ready when needed.
What becomes clear is that the semiconductor business is dramatically different from many other types of production processes. More types of potholes will pop up along our road than with other, more conventional disciplines. Semiconductor manufacturing is unique in its disparity of processes that take us from the pure silicon ingot to the last test after packaging.
Since chipmakers are reluctant to fund 300-mm development, the larger companies will cope better, while many of the smaller manufacturers that supply critical elements to the production line will struggle to afford a 10-15% R&D budget. For 300-mm market entry, this percentage may even have to be increased. All of this, unfortunately, occurs at a time when many of these businesses are still trying to rebound from the recent slump.
No doubt the road will eventually be well paved, but some may find the toll too high. The road also may not quite lead them to the place they`d expected or take them there as fast as they`d hoped.
Wherever the road may lead, Solid State Technology and our weekly sister publication Wafer News newsletter will follow the travails of this international journey and take you along for an informative ride.