By Phil LoPiccolo, Editor-in-Chief, Solid State Technology
During the panel discussion on “Solutions to the R&D Challenge,” held on the final day of The ConFab, Simon Yang, senior VP and CTO of Chartered Semiconductor Manufacturing, called for greater support for an industry-wide “open ecosystem of R&D” to deal with the impending R&D funding crisis.
Yang pointed to several factors that have been converging in the past five years to create a perfect storm in semiconductor manufacturing R&D. At the same time that technology scaling is approaching physical limits and profit margins for technology and manufacturing companies are shrinking, R&D operation costs are escalating out of control.
On the technology front, “Moore’s Law will have to bend,” predicted Yang. The reason is that gate-dielectric thickness scaling, voltage scaling, and interconnect RC scaling have all virtually stopped. Also, while he believes that critical dimension reduction will continue for the short term, he worries that it, too, will come to a halt after 193nm lithography. “One of the fundamental conditions that enabled Moore’s Law in the last 40 years was that silicon technology could scale without much innovation,” he said. “But we all agree that this condition no longer exists today.”
In terms of shrinking profit margins and rising R&D costs, Yang referred to a familiar chart from VLSI Research showing that as worldwide chipmaking revenues will grow at 6.5% annually through 2020, R&D costs will rise at a much faster annual rate of 12.2%. Adding to the chorus calling for greater collaboration to address the dilemma, Yang cited a number of R&D successes that have been achieved recently through industry partnerships, including Chartered’s Technology Alliance with IBM, Samsung, and Infineon to develop 90nm, 65nm, and 45nm technology. Without this alliance, he said, the Chartered technology roadmap would not have been able to keep pace with the ITRS roadmap at these technology nodes.
With respect to process technology, specific benefits of joint R&D efforts include roadmap leadership, consistency in delivery, the ability to use common rules and shared models, and enhanced innovation and economic efficiency, Yang explained.
The ability to share the cost burden with the alliance partners was also a key to success for Chartered. In fact, while the company’s annual R&D spending doubled from about US$50 million in 1998 to just over $100 million in 2005, R&D expenditures at some other foundries soared by as much as eightfold during the same period, from the same initial levels to some $400 million, according to Yang.
In the last few years, Chartered has been able to close the technology gap while keeping costs down, but this is not enough, Yang said. “To realize the full benefits of collaboration, it’s also necessary to enhance your business.” It was important, he said, that Chartered was also able to form a business arrangement with its partners that included common customer support as well as a combined manufacturing support system.
The advantages partnerships can bring in terms of enhancing manufacturing capability, Yang explained, include the ability to provide virtual capacity through reciprocal manufacturing, fab-to-fab technology transfer, common “copy exact” processes, and a worldwide manufacturing presence. What this approach offers customers, he said, are manufacturing design solutions that work, comprehensive technologies, multi-fab solutions, and an “open ecosystem” of third-part design, intellectual property, and tools.
Yang noted that help is needed from equipment suppliers to support these joint R&D efforts. For example, he said, they should provide single-wafer processing tools whenever possible, enable within-wafer splits, and offer low-cost multi-purpose platforms for research activities.
Finally, taking a leadership role in R&D is essential for success, advised Yang, and doing so requires a number of initiatives. The successful participant will create win-win partnerships, foster an innovative environment, develop the knowledge and sensitivity to spot winning solution paths, and enable world-class execution. — P.L.