AT THE HELM: Bob Helms (inset), the new president of International Sematech, impressed the board with his technical and business know-how.

Sure, the technical challenges of running one of the world's most prestigious semiconductor manufacturing research consortiums are daunting. The position presents a raft of bedeviling issues in the front end alone: next-generation lithography, 300-mm, e-diagnostics, and new materials, to name a few.

But the business side can be even more daunting. Just ask the outgoing president and CEO of International Sematech. "The job is tougher on the diplomatic front, not on the technical front," insists Mark Melliar-Smith, who is voluntarily leaving the Austin, TX–based consortium he has headed since 1997. "While the technical challenges are tough, I think it's more the fact that you're trying to build a consensus among 13 companies who represent half of the world's production of semiconductors."

Add to that task working with equipment suppliers and organizational counterparts in Japan and other countries, Melliar-Smith notes, and you're often walking a tightrope. "All have slightly different cultures and different perspectives on life; coalescing to get synergy is a tough job."

Finding someone to take on that tough job took the consortium's executive search efforts through a list of "about 25 names," Melliar-Smith says. The candidate selected, Bob Helms, combines "both the technical and business background" required, said the outgoing president in a joint telephone news conference on the day after the May announcement.

Based on the stated requirements, the appropriately named successor appears well qualified to take the helm of an organization that has changed considerably since its inception in 1987. Called simply Sematech, the consortium began life with a mission to strengthen a U.S. semiconductor industry then threatened by Japanese competition. During Melliar-Smith's tenure, Sematech accepted its first members from outside the United States and changed its name after joining with its international counterpart in 1999.

"I've got tremendous respect for Kofi Annan at the United Nations," Melliar-Smith joked during the call-in news conference with industry reporters. The departing executive will remain as CEO until phase one of a transition period ends on September 1. The 54-year-old Helms will then have both hands on the tiller as he assumes complete management authority of the consortium and its annual budget of approximately $150 million.

Helms, who becomes president on July 1, has been corporate vice president and director of silicon technology research at Texas Instruments in Dallas since 1999. TI recruited him from Stanford University, where he earned master's and doctoral degrees in electrical engineering. Joining TI in 1997, he served as director of the company's Components and Materials Research Center.

The native of Wichita was a professor in Stanford's electrical engineering department from 1976 to 2000. At the university, Helms oversaw research into atomic-level surface phenomena as part of an overall focus on semiconductor processing and new materials. He followed this work with research on surface preparation, flexible manufacturing, and environmentally friendly processes.

Helms also served as vice chair of International Sematech's executive steering council for three years. On an extended leave of absence from TI, the new executive becomes the fourth president in the consortium's history.

The fact that Helms's résumé straddles both the boardroom and the classroom attracted Sematech's board of directors. "One of the things that drew the board to me as a good candidate is the fact that I have seen a good fraction of the different parts of the industry," Helms related. He noted that he spent the last four years at TI as an R&D leader at a time when the chipmaker was undergoing "an enormous amount of change" that helped make TI "into the company it is today.

Helms believes his academic experiences have prepared him for the unique requirements of running an organization such as Sematech. "I spent most of my career at Stanford. I was thinking that Mark's comments on balancing the needs of the industry remind me somewhat of my silicon technology research at Stanford. Research tends to be funded by consortia and groups of companies in some cases. [A result of this is that] managers need to understand and balance their requirements."

Peppered with questions about his plans for Sematech during the news conference, Helms indicated he would initially take a steady-as-she-goes approach, building on his predecessor's work before setting his own agenda. "I look at this as a great opportunity to come in and really work off of some key accomplishments in the past. From my perspective we will undoubtedly take on new directions." Largely, though, the consortium's activities "will be very much in line with the overall charter."

A questioner noted that Sematech had changed dramatically during Melliar-Smith's tenure, as chipmakers based outside of the United States were given full access to programs. Will that trend continue, with Japanese companies joining and equipment suppliers becoming full members?

Helms replied that Sematech would continue to focus on the needs of its 13 member companies "that you might say are our customers." Certainly, he added, equipment providers are other important "stakeholders." He preceded his reply, however, by pointing out the accelerating pace of new-technology introductions in areas such as lithography, new materials, and even new design architectures.

"The cost of those developments, as well as the speed at which the industry is trying to ramp up new technologies, is really requiring a much more international view—or global cooperation—than at any time in its history," he continued. "In many ways I think Sematech under Mark's leadership saw this need earlier than others and started to drive Sematech in that direction. There's certainly no turning back from that. Global R&D is the only way to get us to where we need to be."

Melliar-Smith emphasized that Sematech has no plans in the near future to admit Japanese members. He pointed out that the organization works closely with Selete, the Japanese consortium, and is conducting a joint research program with IMEC, the university-affiliated research consortium in Belgium, on new front-end process gate stack technology. Hitachi recently decided to take part in the joint endeavor, he said.

"I think you're beginning to see things mix up more and more," Melliar-Smith observed. "We're looking to build on that [direction] in the future."

Asked how Sematech will accommodate an industry trend toward greater foundry use, Helms noted that he views TSMC as one of its "key members." Both he and Melliar-Smith insisted the needs of foundries and the rest of the membership are similar. Helms added that Sematech shouldn't restrict membership to integrated device manufacturers alone. "There's no interest on our part to keep foundries out. We think [their presence] just strengthens the consortium."

Given his extensive background in semiconductor processing, Helms identified transistor gate insulators, next-generation lithography (NGL) issues, and photomasks as a few of the major challenges facing Sematech and the industry as a whole. He cited the consortium's partnership program with the Semiconductor Research Corp. (SRC) to fund research into new materials, mentioning in particular a recently established joint program with SRC to develop high-k gate insulators. "Sematech has and will continue to take the lead in driving progress in that area," he insisted.

With chipmakers drumming up less business than in previous years, sales of wafer fab materials are expected to slow considerably in 2001. The semiconductor industry downturn will take a toll in particular on sales of process chemicals, photoresist, and silicon wafers, according to a forecast presented in April by SEMI at its annual Semicon Europa trade show. Fab materials sales overall will grow from the $16 billion mark in 2000 to $17 billion in 2001, a growth rate of only 6.4%. SEMI based the overall prediction for 2001 on a 3 to 5% increase in chip sales. Photomasks fare well in the forecast, which foresees an 8% increase for the category. In addition, the market for other substrates shows relatively healthy 6.9% growth, as alternatives such as gallium arsenide begin to get a foothold. Projected to increase by 50% to $870 million in 2001, the new-materials category includes products such as low-k dielectric films and CMP slurries.

Addressing NGL issues, Helms said, "We have some enormous challenges right in front of us that need to be solved in the next year or so." He noted that Sematech's focus "has been largely on 193-nm technology in the last few years, going on to 157-nm." Getting a mature 193-nm process and expediting progress on photoresists, photomasks, "and the like" will be high on the priority list. "We view masks in particular as one of the key challenges we have today going into the future. Even without NGL [issues], we have a lot on our plate.

"With respect to NGL," Helms continued, "I think it's very clear that over the last three to six months the movement has been toward extreme-UV technology." He also regards EPL as a competitor "and a potential player, especially to specific segments of the industry." Direct-write lithography "is a little further out" on the schedule, Melliar-Smith added, because of the need to develop a pixel rate that will allow acceptable wafer throughput.

Helms acknowledged that 157-nm lithography faces yield problems caused by the calcium fluoride crystals used in the systems. "It depends on who you are speaking to about how difficult it was going to be in the first place," he said of the transition to 157-nm lithography. "As we see the industry come on line with 193-nm, it's clear that calcium fluoride is a key part of the supply chain that will need to be dealt with. The 157-nm developmental schedule we have is indeed very aggressive, but so far we're staying right on track. Calcium fluoride has to be dealt with, and that is driving the design method for final tools. We're guardedly optimistic about the schedules that the suppliers are putting on the table."

Melliar-Smith added that the industry is also having problems with soft pellicles but has made much progress developing resists. A "mature 70-nm patterning process" with proper CD control and a pattern etched into the "underlying material" of the whole wafer should be ready by 2005, he asserted, pointing out, "It's a heck of a lot more than just resist pictures."

Helms perhaps provided further insight into his overall philosophy during his keynote address in mid-April to an audience of ESH professionals at the Semiconductor Safety Association conference in New Orleans. The semiconductor industry, he said, faced distinct challenges from revenue growth and productivity. Chipmakers historically have experienced annualized growth rates of 15 to 20%, said Helms, adding, "I see nothing to prevent that rate from continuing through the second half of the decade."

Citing basic economic theory, the new Sematech president told the audience that a business can make more money by either increasing demand for its current products or reducing the cost of making the product. One way chipmakers can create more demand, obviously, is to increase the functionality of their products, he said. Under Moore's Law, costs drop by dint of the fact that chipmakers are putting more die on each wafer.

During the speech he pointed out that TI has been able to increase its ratio of die per wafer by 800%. The Texas-based chipmaker uses the term "technology entitlement" to describe a three-part strategy to increase performance, reduce power consumption, and increase functionality, Helms related. An example of more functionality is an "embedded, on-board SRAM device." Capacity requirements, he said, "are doubling every 18 to 24 months, with a comparable reduction in cost per bit."

Commenting on the speed of technological and manufacturing changes, Helms joked, "You almost need a roadmap to see where the roadmap is going."

Helms stressed that semiconductor manufacturers must see improvements in both capital productivity and R&D productivity. "We need smaller, high-productivity 300-mm fabs and significantly higher process tool utilization," he asserted.

During the joint telephone news conference, the incoming president was asked whether anything had happened in the month between the New Orleans keynote and the Sematech announcement to temper his optimistic belief in the industry's positive rate of growth. Admitting that "we're now going somewhat in the wrong direction," Helms replied, "I certainly see no reason that it can't continue" based on "where the end customers are and where the technology is going." Indicators showing "signs of life" for the industry include PC inventories and, more in TI's ballpark, inventory levels in the wireless technology segment.

Melliar-Smith said he has been planning the transition for year and added that his decision was based on a desire to ensure that Sematech's management direction remains fresh. "It makes sense in a consortium to change management every five to seven years," he insisted. The departing executive said he will not return to Lucent Technologies nor "instantly look for an 80-hour-a-week job running an organization." He is working with foundation boards in the Austin community and will eventually pursue "personal career interests."

Did Melliar-Smith have any advice for his successor? Well, yes, he did. "Whatever you've done in the past doesn't matter, running a consortium in the semiconductor industry is one hell of a ride. The one thing that's probably the greatest challenge of all is to be able to juggle the competing issues and needs of the industry around the world. Building that sort of consensus has been quite a challenge. And I'm sure Bob is up to that challenge."