By Hank Hogan
Along with the usual New Year’s resolutions, the semiconductor industry has one that impacts contamination control and cleanrooms. The industry needs additional money, to the tune of almost $10 billion, for research and development over the next five years. R&D expenditures are climbing for chip companies, and they’re looking to share-and-share-alike with subsystem and component suppliers.
It’s outsourcing at work, according to Ron Leckie, president of the analyst firm Infrastructure Advisors (Saratoga, Calif.). “You shift the burden of manufacturing, but also the burden of R&D, to your suppliers,” he says.
In this case, however, the shifting is significant and uneven. According to an October 2005 study done by Infrastructure Advisors for the semiconductor industry trade group SEMI (San Jose, Calif.), the gap between what’s needed to keep up with Moore’s Law and industry funding could reach $5.8 billion by 2010. That shortfall swells to $9.3 billion if a transition from 300 mm to 450 mm wafers takes place. Another finding was that the R&D funding from component and subsystem supplies has climbed from under 5 percent of revenues 20 years ago to an average of about 15 percent now.
 Shown here is a 300 mm FOUP from Entegris that cleanly carries wafers through a fab belonging to one of the company’s customers. Photo courtesy of Entegris, Inc. |
This has manifested in a number of ways. At particle counter manufacturer Particle Measuring Systems, Inc. (PMS; Boulder, Colo.), there’s an ongoing drive for better sensitivity so that smaller particles can be spotted. That trend, according to PMS vice president of marketing John Mitchell, has gained momentum in the last few years. Consequently, there’s R&D spending to meet that need.
R&D expenditures have also gone up for another reason. End users increasingly prefer to rid themselves of separate particle monitors. “They’d rather have it integrated into the process tool,” says Mitchell. Accomplishing that can sometimes be relatively easy and other times difficult or even impossible. However, it’s never without some R&D cost.
At Entegris, Inc. (Chaska, Minn.), the increasing burden shows up in both the company’s wafer- and fluid-handling businesses. With regard to the latter, there’s a demand for increasing expertise and associated R&D expenditures in fluid chemistries, blending, metrology and contamination control.
According to Tracy Niebeling, marketing director for the company’s wafer-handling business, bearing the research and development costs associated with changes in wafer size isn’t anything new. The lesson learned from the move to 300 mm wafers is that shifting to a larger size can take quite a bit longer than anticipated and, so, the payback may be slower. For that reason, Niebeling notes, the industry is being cautious about any investment for 450 mm wafer handling.
But that reluctance doesn’t impact the other R&D driver, the shift to smaller and smaller feature sizes and more advanced process nodes. One outcome is an increased concern for contamination. “We have to be worried about particle generation at smaller and smaller particle sizes. At the nodes that we’re looking at now, we also have to be concerned with molecular contamination,” says Niebeling.
As for how the R&D shortfall can be eliminated, the SEMI-sponsored report has a number of suggestions, including maintaining R&D spending during downturns, reducing R&D costs, external funding by chip customers and governments, slowing the pace of innovation, and a back-to-the-future approach of partnering and collaboration among chip companies, subsystem suppliers, and others. In the end, however, any solution has to recognize a fundamental fact: “Money isn’t going to come from anywhere other than from revenues,” says Leckie.