What's Wrong with Today's Research
The semiconductor industry has a need for fundamental research -- it's time the world's universities took up the challenge.
Peter Singer, Editor-in-Chief -- Semiconductor International, 6/1/1998
There has always been a gap between the kind of manufacturing-oriented research that the semiconductor industry needs and the type of research that university researchers are able to deliver. As the industry moves to implement new technologies at an ever faster pace, this gap has widened into a gulf.
For some types of research, this gulf will never be bridged. University researchers simply don't have access to billion-dollar fabs, so they will never be able to effectively study issues relating to materials integration, factory flow, etc.
But there are plenty of areas where the average graduate researcher could make a significant contribution to the semiconductor industry -- and through the semiconductor industry to the world in general. Areas where the fundamentals of science -- chemistry, physics, electricity, math, mechanics, etc. -- can be applied to understand a practical situation.
One good example where the semiconductor industry desperately needs some help is with copper electroplating. As discussed in this issue's cover story, a major push is under way to switch from aluminum interconnects now used to copper, and the bulk of the copper will be deposited with electroplating. Unfortunately, few have studied the sophisticated interplay of electricity and chemistry involved when electroplating copper into deep, 0.18 µm structures. This is a situation tailor-made for a university research project -- no complex apparatus is required, just a tank, a solution of copper electrolytes and a couple electrodes.
Of course, there are many other examples that could easily be studied at the university level: particle transport in various mediums, algorithms that can speed mechanical movement, tests to accurately measure a film's moisture content, etc. Even something as seemingly mundane as the vacuum pumping characteristics of various bolt designs could wind up providing some real data that could be immediately put to good use.
To be sure, some efforts are under way to help students and professors understand what the semiconductor industry needs in terms of research. The Semiconductor Research Corp. (SRC), in particular, has done an excellent job in this regard (check out its web site at www.src.org). But a quick perusal of the technical journals where university papers are most often published turns up thousands of papers on subjects that are of absolutely no value to the semiconductor industry -- or anyone else for that matter.
Bridging the university-industry research gap will require work from both sides. The semiconductor industry needs to more vigorously promote an agenda of items, such as copper electroplating, that can be researched in a university environment. And university researchers need to be willing to take on some less glamorous, yet more practical research subjects. If that can occur, the world might finally benefit from what has been a largely wasted resource of talent and fresh young minds.
