Michael Polcari, President and CEO, Sematech

Michael Polcari (Source: Sematech)

Michael Polcari became president and CEO of Sematech (Austin, Texas) in February 2003. He has led the launching of two Sematech subsidiaries, ATDF and the International Sematech Manufacturing Initiative (ISMI), and overseen the inauguration of the Advanced Materials Research Center (AMRC) in Austin, a collaboration involving Sematech and Texas universities. Previously, during his nearly 30-year stay at IBM, Polcari was vice president of procurement engineering for IBM Global Procurement in Somers, N.Y., and was responsible for IBM's engineering effort in procurement, including supplier quality management; previous to that, he held various positions of importance at the company. Polcari has a Ph.D. and M.S. in solid-state physics from the Stevens Institute of Technology. He conducted additional graduate work in physics at the University of Maryland, and earned a B.S. in physics from the University of Notre Dame . He served as chairman of the board of directors of the Semiconductor Research Corp. (SRC), and is a member of the American Physical Society, the Electrochemical Society, the Society of Photo-Optical instrumentation Engineers and the Institute of Electrical and Electronics Engineers.

SI: We interviewed you when you first took over Sematech. Since then, can you tell us what you regard as your prime accomplishments?

Polcari: We've had several ongoing challenges and successes. We have had some very good results stemming from our technical programs. We have an excellent advanced-gate technology program that is working on high-k metal gate problems. It has made very good progress in understanding the process implications of mobility degradation, and how to resolve some of the issues. We've been doing good work in film characterization and reliability. We've also managed to do quite a bit of work on immersion lithography, by running several workshops and helping to find answers for some of the critical questions that the industry has had. Now a number of people are focused on immersion lithography, and we have started an immersion litho tech center to focus on understanding the extendibility issues and how far the technology can be pushed... In conjunction with that, we also have our EUV program, with much of that work centered at Sematech's Albany facility, where we have made considerable progress in the reduction of defect levels in mask blanks. We have a microstepper for doing EUV exposures in place. This is now open to the membership and some suppliers, for them to come in and learn about EUV exposures in lithography.

SI: Have there been any changes in Sematech itself?

Polcari: Organizationally, we've changed some of the structure to allow more flexibility and enable us to focus on commercializing technology and accelerating the pace of the mainstreaming of innovation from either startups or equipment companies. These are things that we've always done, but we've been focused on improving the pace of these activities. Now the AMRC, which is a joint program we have with the state and university system of Texas, is another important part of our effort to accelerate the commercialization of the research and technology taking place and being produced in universities. We pull these results through and couple them very directly with industry, which greatly accelerates commercialization.

SI: Your fab is also online.

Polcari: Yes, it's open to the industry, and people can come and use it to try new kinds of ideas and take advantage of the semiconductor infrastructure that we have in place, using it to help take those ideas into the market. We have also opened our manufacturing program to members specifically interested in the technology's manufacturing and manufacturability aspects. We will be announcing some new members very soon.

SI: What would you say has been the most difficult aspect of your tenure, and what did you find out was easier than it seemed?

Polcari: Probably the most difficult thing of all is to ensure that we're focused on the correct technology direction. Certainly, our membership is of invaluable help here, and their guidance helps us to remain aligned with the common themes of the technology challenges that they're interested in.

SI: One of the reasons for Sematech's existence is the fact that carrying out any sort of major R&D effort today can be prohibitively expensive. Partnering has become a requisite for this kind of effort, but it isn't always easy — some have compared it to herding cats. How do you go about putting together your teams?

Polcari: (Laughing) Many people have used that analogy! Seriously, though, being able to do this has been Sematech's heritage, as well as one of its strong points. Over the 18 years or so of our existence, we've developed the techniques and processes that successfully enable us to carry it out. Don't misunderstand me, it is always a challenge, but we have been able to consistently meet it to just about everyone's satisfaction. In this sort of an effort, the worst hurdle is to identify where the areas of overlap lie, and what is most beneficial for all. Once you've accomplished this, it is mostly a matter of organization and going out to tackle the problems. We have a series of processes in place to identify these challenges — an advisory system. For instance, on a yearly basis we poll our members on what they think are the top 10 technology challenges we should address. We correlate the information, and then in a membership meeting we structure what our activities will be. We structure a plan aimed at solving the problems, and work with the membership to determine whether the projects specified are the right ones for the goals we wish to accomplish. This way it isn't difficult to find commonality in our efforts.

SI: Since we last talked, processes have become more complex and unforgiving, capital equipment is more expensive, and yields have become problematic. In view of all this and more, how do you view the future of Moore's Law?

Polcari: Most of the players are still committed to maintaining the pace. I think the two- to three-year nodal rate will continue. Regardless of appearances, we're finding that the changes we're making per node are not as aggressive as originally planned, and we continue to be able to move down the node. Much of the technology that gets worked on maybe comes at a node later than was originally thought, but it still comes in. Something we continue to focus on at Sematech is cost, and we'll emphasize this even more in the future. We'll look at the cost implications and manufacturability issues related to some of these technology changes and try to address them as we formulate our programs. I believe we will continue at the 18-month pace, and I don't think that this will change anytime soon.

SI: The industry seems considerably more conservative. Had we followed some of the projections of the ITRS, for instance, we'd be using ultralow-k.

Polcari: That's precisely what I meant. We tend to be aggressive on the material changes and slow down on some of the implications. But we're pretty much on track on the density change.

SI: Do you think we're too aggressive and running into some fundamental walls?

Polcari: In some cases, yes, there are limits of that kind. However, in a lot of cases, the issues of the features being introduced run into an unexpected level of difficulty. This, coupled to costs vs. payoff, results in some rethinking of options, and things may slow down a bit for a generation or so. This is what happened with low-k; there are ways to get around it by using more levels of wiring. It might not be the most effective solution, but it can be cost-effective at the time.

SI: Does it seem to you we'll be running out of workarounds around the 45 nm node?

Polcari: Possibly. There are workarounds and then there are workarounds. Yes, sometimes you do run out of room and have to do something really different. However, in many cases, new workarounds come about. Partly, this is what has enabled the roadmap; if you go back and read some predictions about what we would have to do at the 1 µm level, much of that didn't even come close to becoming real. In the end, whatever it is, we will do the cost-effective thing. If this is low-k, then we'll put it in; if it is an extra two levels of metal, we will do it. Yes, workarounds are getting more difficult to do, but we still have maneuvering room.

SI: Immersion lithography and the hybridization of circuits through nanotechnology are starting to be viewed as the panacea for many of the problems we'll encounter at the 45 nm node and beyond. How do you see this?

Polcari: We are strongly pursuing the extendibility of immersion; what the extendibility for 45 nm half-pitch will be. There are some critical things to be looked at, and this is an aspect we'll inspect with the tool we've received: What is the real extendibility, what are the issues that need to be solved? As far as hybridization with nanotechnology is concerned, I don't know if we'll see it at 45 nm, 32 nm or beyond. Personally, I think we'll have a silicon-based technology at least to 32 or 22 nm. I do not see things like carbon nanotubes coming in until 22 nm or beyond.

SI: With all the buyouts and mergers that we've witnessed over the past few years, which have resulted in a smaller base of companies, do you find that it is getting increasingly difficult to mount industry-wide R&D efforts?

Polcari: Since its beginning, the industry has been changing and consolidating, people have left, IDMs are now fabless, etc. Certainly, there are fewer players nowadays and therefore fewer companies and people to participate; however, when R&D is needed, the industry always manages to carry it out.

SI: Looking at $2B+ fabs, how do you view the fab as the technology gatekeeper, in the sense of resisting to mainstream technology that forces them to retool possibly one-third of their capital equipment every 18 months, before an ROI?

Polcari: It is all part of the cost/benefit trade-off: If the benefit warrants the cost trade-off, it will happen; if it doesn't, it won't. As long as we can continue to move down the productivity curve, we will do it and we will do it for the smallest cost possible. If new tools will help with this, the tendency will be to retool; otherwise, it won't. This is why a change such as the advanced gate stack won't happen until we really need it and the product's benefit is commensurate with the cost.

SI: There's talk of a downturn — again. Is there anything that we can do to ameliorate these industry cycles?

Polcari: I don't believe we can completely eliminate the cycles. By working more closely together, we can help alleviate some of the issues in downturns. One can reduce costs by sharing in innovation experimentation and development, and thereby continue to innovate during some of these downturns when dollars are in short supply.

SI: Now that you've had your hand on the Sematech tiller for some time, what are your plans — short and long-term — for the organization?

Polcari: We have established a good structure together, and enhanced our ability to speed up the commercialization of technology to manufacturability. We've also provided an environment and opportunity to foster innovation. Our role, now and in the future, is to accelerate the next technology revolution — whether it's in next-generation materials, lithography or manufacturing — with a focus on both innovation and manufacturability.