I hope there is room for 2. I own both LU & INTC.
Posted at 12:11 a.m. PDT Monday, June 7, 1999
Intel, Lucent duel on advanced lithography
BY TOM QUINLAN
Mercury News Staff Writer
Chicago, not known as a center of high-tech innovation, could be where the future of semiconductor development is decided this week, as the chip industry evaluates the progress of critical dueling technologies.
The technology in question is advanced lithography -- the techniques required to etch ever-smaller patterns onto silicon wafers, the raw material for microprocessors and memory chips. The duel features two industry heavyweights -- Intel Corp. and Lucent Technologies -- each of whom has invested millions in exotic new lithography methods to make sure they can continue to produce smaller, cheaper and more powerful chips.
Intel is putting its muscle behind Extreme Ultra Violet technology. Lucent is pushing an alternative, Electron Beam-based technology. Each is anxious to demonstrate that its favored technique now has the upper hand.
At an industry conference in Chicago, both companies intend to offer proof that their technologies can create patterns on chips with a line width of less than .07 microns -- there's 25,400 microns to the inch -- compared to the .18 micron process that is state-of-the-art today. Both will showcase their ability to project complex diagrams onto silicon wafers at a cost competitive with today's lithography. And both will offer timelines that put their technologies in operation no later than 2005, just when chip makers are expected to require such advanced techniques.
The audience for the International Sematech Advanced Lithography Critical Review will be limited to 175 engineers. International Sematech is themultinational chip industry consortium that may decide to throw its weight behind EUV or Electron Beam later this year.
''There's some pressure building for (International) Sematech to choose one or the other,'' said Chuck Gwyn, Intel's Program Director for the EUV Limited Liability Corp., a consortium of semiconductor companies and the federal Department of Energy that are jointly developing the advanced lithography technology.
''(A decision) would help generate more industry support, and it requires a lot of infrastructure to make something like this succeed.''
If Intel can count its own consortium and the federal government in its camp, Lucent still has a compelling message on behalf of its own technology, known as Scalpel -- or scattering withangular-limitation projection electron-beam lithography. It promises to be ready to go in 2004, a year before the projected date for EUV. It recently gained support from equipment manufacturers such as ASM Lithography and Applied Materials Inc.
Lucent's trump card, however, may be the fact that a slightly different version of its electron beam technology is being championed by IBM and Nikon. Called Prevail, this technology too will be presented International Sematech's members.
''The idea really is to have different suppliers that develop one technology,'' said Lloyd Harriott, head of Advanced Lithography Research at Bell Labs, Lucent's research arm. ''We've already had discussions with Nikon to try and make sure that the technologies aren't so different that they require a different process.''
It's a strange competition.
Neither Intel nor Lucent expects to benefit economically or competitively if they win the race to develop the next generation of lithography equipment. And both are committed to providing hundreds of millions of dollars to develop these different technologies, regardless of which one gets the nod.
As long as one technology succeeds, Intel and Lucent -- along with the rest of the semiconductor industry -- will come out winners.
That doesn't make the competition any less fierce, however.
At stake is the ability of a $200 billion industry to continue delivering on Moore's Law -- an axiom first developed in 1965 that shows processing power will roughly double every 18 months to two years.
Since Intel co-founder Gordon Moore first outlined it, that 'law' has formed the technological and economic basis of the entire industry -- but without radically new technologies by the year 2005, that underpinning is in danger of collapsing.
''The fear is what if traditional technologies suddenly hit the wall, and can't be used to keep producing smaller chips and there's nothing to replace (that technology) with,'' noted Dan Hutcheson, president of the San Jose-based market research firm VLSI Research.
While both EUV and Electron Beam -- a third contending technology, based on X-rays, has now largely been abandoned by everyone except IBM -- have shown substantial progress so far, ''either one of these -- or both -- could lead to a technology cul-de-sac,'' he said.
That wouldn't mean an end to the semiconductor industry, but it would change the financial model they've been operating under since the 1980s.
''One thing about chips is they don't wear out,'' Hutcheson said. ''The idea that every two or three years you have the opportunity to buy something significantly more powerful would go away.''
The need to have a replacement technology in the wings is what fuels the competition. The need to choose a technology that proves successful is what makes International Sematech's decision so critical.
It is possible, of course, that International Sematech will decide not to choose between EUV and Electron Bean, at least for now. Since research began on those two techniques, the industry has refined and improved its traditional lithography processes -- a move which, in essence, buys more time before the switch to a new method must take place.
''People are a little more sophisticated today when it comes to evaluating traditional optical lithography technology,'' Harriott said. ''They aren't as willing to predict when that will hit the wall, if ever.''
Lucent itself recently announced that it is able to trace a line .08 microns wide using existing technology -- a line that is almost half the size of the light wave generated by today's ultraviolet lithography machines. And many observers, including Hutcheson, expect that existing technology will be used to etch chips at 0.1 micron, something that most experts called impossible as recently as two years ago.
''The demise of optical lithography has been predicted for a long time, and obviously hasn't come true yet,'' acknowledged Intel's Gwyn. ''But there's also the fact that at some point it becomes harder and more expensive to extend those technologies.''
|
