Is Basic Research the Government's Responsibility?
Peter Singer, Editor-in-Chief -- Semiconductor International, 10/1/2000
 |
However, according to a report disseminated by the Semiconductor Industry Association (SIA), the situation is not as rosy as it seems. One concern is that federal R&D spending is decreasing as a percentage of the total, from 50% in 1980 to 27% today. The conclusion is that less basic research will be conducted, since it historically has been the government that has funded such research. Another concern is that life sciences are being funded at a much higher rate than physical sciences and engineering, which form the cornerstone of semiconductor manufacturing know-how. In FY2000, for example, the National Science Foundation gets a 5.2% increase over 1999, while the National Institutes of Health gets a huge 14% boost. The end result of this uneven funding is that the attraction of these fields for university researchers declines; new discoveries in the more neglected fields shrink, and graduate students drift to other fields. So not only does less research get done, but fewer people enter the workstream ready to take on semiconductor manufacturing challenges.
The good news is that semiconductor companies are investing heavily in R&D, with internal investments of 13% of sales per year, well above the national average of approximately 5% per year for all industry. There also has been tremendous participation in cooperative R&D ventures, including the Semiconductor Research Corp., SEMATECH and the Focus Center Research Program (which also receives 25% federal funding support). The only problems are that these R&D investments are a function of sales, and therefore cyclical, and that most industry-sponsored R&D tends to focus on short-range issues.
But do we need more basic research? It's debatable whether the semiconductor industry's long-term success will depend that much on what researchers are learning by fiddling around with quantum dots and the like. Proponents say we need a better understanding of the physics of atomic-scale materials and how to synthesize novel materials. They point to the "sea of red" in the International Technology Roadmap for Semiconductors as if that is proof-positive that, for the first time, the industry is facing technical challenges for which there are no known solutions.
I don't believe it. The industry has always seemed to have difficulty predicting how it would overcome problems five years or so down the road. (Does nobody remember the seemingly impossible challenges in breaking the 1 µm barrier?) We understand plenty about the fundamentals of science; it's applying those fundamentals in a complex production environment that's difficult - and that classifies as short-term research.
There is a very real, looming problem of an inadequate supply of individuals with the advanced scientific and technological education to meet the semiconductor industry's needs. If the best way to produce those individuals is to give them government money and let them play "blue sky" in a lab for a few years, so be it. I would rather see that than tomorrow's engineers drift into the life sciences. But I think an even better solution is for the government to give the semiconductor industry some R&D tax credits and regulate how those credits are used to even out the industry's cyclical nature, and for semiconductor companies to use the credits to sponsor more students in engineering studies - and maybe even let a few of them fiddle with those quantum dots.
