Issue



Dual routes for R&D may be best


03/01/2004







With research dollars being cut back by both the US government and private companies (see "Will the US lose its lead in research?" Solid State Technology, January 2004, p. 12), it would be wise to tune the funding model to gain maximum effectiveness. Finding what works best was one of the issues probed by MIT's Commission on Industrial Productivity, a group I worked with in the late 1980s. The Commission studied eight major industries, including computers/semiconductors/copiers, to compare how they operated in countries around the world in comparison to the US.

Among its findings was that Germany appeared to have devised a particularly effective approach to splitting basic and applied R&D in universities. While Max Planck Institutes concentrate on basic research, Fraunhofer Institutes do more applied work, with each one specializing in a particular industrial sector or a few sectors. Thus, for example, there are Fraunhofer Institutes doing advanced work in areas such as MEMS and electron projection lithography in our industry. One finding of the MIT study was that when machine tools are ordered in the US, the specifications tend to be only a few pages long, while in Germany they may cover hundreds of pages, mainly because of the in-depth training engineers get through participating in specialized R&D efforts there. By separating out the Planck Institutes, fundamental studies of sciences like physics and chemistry can continue without being swallowed up by the Fraunhofer Institutes, which tend to gain more funding from industry looking for more immediate solutions to practical problems.

The direction of R&D has shifted in the US over time. During World War II, universities were enlisted to help in research on technologies for the war effort. Some formed organizations like Stanford Research Institute and MIT's Radiation Lab and Lincoln Labs, which became major R&D centers. This continued during the economic buildup that followed, with SRI doing pioneering work under Doug Engelbart (a giant of our time who deserves much more recognition) leading to the PC revolution and many other developments that are still unfolding, and MIT's RadLab doing innovative work in radar and inertial guidance. Lincoln Labs was instrumental in development of 193nm lithography and remains a major contributor in this and many other fields.

Eventually there was a backlash in universities, with many trying to return to more basic research. Princeton, for example, set up its Institute for Advanced Studies and attracted Albert Einstein, among others. Mathematicians, especially, glory in developing new fields that seem to have no practical application, but often turn out to be crucial in forging new technologies. Allowing brilliant people to explore their interests is often a worthwhile investment, even though lab administrators might chafe at the expense. Long ago at Bell Labs, a researcher told me he loved working there because his boss allowed him to use the instruments to explore his own interests after hours and on weekends. He was studying the way oil films create color patterns. Later, these seemingly useless effects were harnessed elsewhere for a type of color video projection. Benoit Mandelbrot was thought to be off track, claiming that there were fractional dimensions and exploring obscure mathematics, but the result was fractal geometry, which has made strong contributions to realistic computer-generated images.

What I liked about the MIT work was that we were studying what other countries were doing that might be superior to the way we do it in the US. Japan, in particular, is outstanding at benchmarking the best practices it can find anywhere in the world, and then not just adopting them, but trying to improve on them. The US could learn a lot from other countries, as our industry has been discovering now that consortia once devoted to boosting individual economies now work collaboratively on advanced semiconductor processes.

Germany has devised an elegant and useful model for supporting both fundamental and selectively applied R&D. The US should consider doing something similar to ensure that what remains of our basic research is not swallowed up by the thirst for industrial support.

Click here to enlarge image

Robert Haavind
Editor in Chief