Commentary: The EUV-LLC and the systems integration question

01/01/1998

Commentary: The EUV-LLC and the systems integration question

Peter N. Dunn, WaferNews

After a half-year or more of searching for partners to develop extreme UV lithography, Intel Corp. has signed up Motorola and Advanced Micro Devices to take part in the private EUV Limited Liability Corp. (EUV-LLC), which will spend about $250 million over three years to fund development work at three Department of Energy (DoE) laboratories.

The deal is a major step forward for the challenging EUV technology, which is competing with several other post-optical methods. It`s also a very appropriate expenditure for the chipmakers involved - they`re the three biggest US manufacturers of logic devices, and have an obligation (both out of self-interest and noblesse oblige) to plow some of their profits back into infrastructure development.

However, no system integrator has yet been chosen to build an actual exposure tool, and US-based stepper companies have expressed concern about the integration contract going to a non-US company. Ultratech Stepper president Art Zafiropoulo has also proposed the establishment of a second limited liability corporation for the specific purpose of developing a prototype tool; the US-based group would then finance volume manufacturing with a public stock offering.

Ultratech and SVG Lithography (SVGL) are the two US companies most likely to be capable of integrating the various technologies into production-worthy tools. Ultratech, in particular, was a participant in the first phase of EUV development, and has a corporate strategy of re-entering the critical lithography market in the EUV era. While only US companies are welcome in the EUV-LLC, the consortium is willing to transfer production of the exposure systems to whomever commits to build them in the US and deliver them initially only to EUV-LLC members.

Energy secretary Federico Pe?a commented, "We want US equipment manufacturers to be able to participate fully in this venture. When licenses are provided by EUV-LLC, they will be provided on an equal basis so that all companies can participate. To defend that, any international company will be required in the first two years of the program to actually have the manufacturing done in the US. I have invited representatives of the US companies to our offices in Washington, to talk to them about how, perhaps, we can work with them in a creative way to make sure that they are able to fully benefit from these opportunities."

Zafiropoulo stated, "I believe that if EUV-LLC goes to foreign companies for system integration, it will mark the beginning of the erosion and ultimate destruction of the lithography business in this country, and potentially the rest of the semiconductor equipment business."

Bob Richardson, corporate VP of business development and marketing at Silicon Valley Group, concurred that "It would be upsetting" if an offshore company were given the integration contract. He added, "We definitely want to participate, though we don`t know what the terms would be. There`s been no agreement put forward on how they want to work." He pointed out that SVGL is "a key supplier to Intel," and noted, "I think we have more to bring to this than anyone. We will certainly be plenty big enough to supply the tools by the time the market evolves." Richardson said that he discussed possible collaboration with Zafiropoulo, and commented, "We need more specifics, but we`re certainly interested in looking at it. The mechanics are always difficult."

Moral obligations aside, the three chipmakers just want to be able to buy litho tools with the right capabilities at the right time, from a vendor they trust. They don`t care if that vendor is based in the US, Japan, or Mars as long as the tools are delivered on schedule and meet spec, and the service people show up promptly.

If the three chipmakers were working on their own, they would simply award their development contracts to the companies that gave them the warmest, fuzziest feeling. But they`re not working on their own - they`re teaming up with a group of three DoE national laboratories, and reaping the benefit of research that has been partly funded for several years by taxpayer money. As a result, there are a bunch of different interested parties coming into play. In the end, however, this might not make much difference.

On one level, the DoE`s involvement in EUV is simply aimed at paying the light bills, and showing that the national labs, which have a truly remarkable collection of minds and hardware, can do work that the private sector finds valuable and will fund. On another level, however, it`s darn handy that the sector in question is the semiconductor industry. Not only is it economically strategic and a hot button for politicians, but it will also provide the ever-more-powerful computing horsepower that the DoE will use to simulate nuclear explosions. This will enable the DoE to continue developing its weapons of mass death and destruction, in spite of the expected ban on actual testing of nuclear devices.

But then there`s the knotty question of infrastructure, and the contractors who have been plugging away on this wildly difficult EUV stuff - all of whom are US-based. Most of them should do okay under the new plan, extending their development programs and gaining knowledge that will come in very handy in the event that EUV emerges as the technology of choice after 193 nm. But when it comes to choosing a company to perform the systems integration work, the EUV partnership is clearly leaving open the possibility of awarding the contract to someone other than Ultratech, which has been involved in EUV for years, or SVGL. The contract might end up going to one of the three large, well-established stepper companies who fly Japanese or Dutch flags.

This would be a cold shot, and quite possibly a long-term shot in the foot. (didn`t Sematech`s efforts to support SVGL`s viability come from a desire to keep at least one US litho company viable?) But underlying American commerce is the principle that money confers the right to make choices. Intel, Motorola, and AMD are putting up the money, and their priorities will come first.

Those priorities include a minimization of risk. So Art Zafiropoulo`s intriguing idea of a US lithography consortium will likely not fly. SVGL would seem like the best domestic bet, but they`re still working on a demonstration of their ability to ramp manufacturing. The low-risk path runs through Japan or Holland; it wouldn`t be surprising to see the EUV group follow it. Some storminess and recrimination from Congress and the Pentagon would undoubtedly follow, but the real power in this situation lies with the big chipmakers, and they will use it to maximize their chances of having the tools they need when they need them.

Peter N. Dunn is editor and publisher of WaferNews, the leading weekly source of news and commentary for the semiconductor manufacturing industry. For more information about WaferNews or a free sample issue, contact Judy Simers at 603/891-9177, fax 603/891-0574, or e-mail [email protected].