Photolithography in the limelight. From Electronic News.
sumnet.com
From this week's Semiconductor Manufacturing Feature
Advanced Lithography At Critical Juncture
By Chad Fasca and Dylan McGrath
Colorado Springs, Colo.--The lights go on. The smoke floats up. A tuxedoed announcer steps to the center of a canvas
square. He grabs a microphone that drops down from the ceiling and says, "Let's get ready to rumble." The crowd erupts.
That may seem like a stretch for Semicon/Southwest or a Sematech meeting, but weeks after the last Semicon booth is
reduced to plywood, the industry elite will meet for the semiconductor equipment incarnation of fight night--Sematech's
November workshop to discuss advanced lithography. For four days, five technologies and 100 of the industry's brightest
minds will descend upon Colorado Springs, Colo., to engage in an old-fashioned stump session. Some downplay the
workshop. Others consider it to be a potential watershed.
"This is a critical meeting for advanced lithography," said Karen Brown, Sematech's director of lithography. "We want to be
sure that the industry is fully informed."
The current titleholder--optical lithography--has been the mainstay in chip manufacturing since the 1960s and, according to
Sematech, is expected to remain so through at least 2003. But to continue delivering more computing power in smaller
devices, a new lithography technology will be required in manufacturing by the 2005-2006 timeframe. The process will
necessitate several steps: the development of a new exposure tool technology for volume manufacturing and the photoresist,
photomask and metrology technology to support it. Sematech expects this to cost more than $1 billion and require an
estimated eight to 10 years of R&D. With all this at stake, several contenders to replace optical lithography in the fab are
lining up to take a shot at the title: proximity X-ray, electron-beam projection, electron-beam direct write, extreme
ultraviolet and ion projection.
For the November session, Sematech has invited lithography experts and managers from semiconductor manufacturers,
exposure tool manufacturers and mask makers--not limiting the participants to Sematech members--to join lithography
R&D experts and university researchers from around the world. The invitation-only workshop, scheduled for Nov. 4-7, will
focus on the facts known about each technology and discuss in detail what hurdles must be cleared in order to attain the
desired manufacturing solution. After the four days, a few technologies should emerge as leaders of the pack. It is not
expected that one technology will be chosen.
"It is not that simple, that everybody goes (to Colorado) for a few days, raises their hands and makes a decision. That is not
what is going to happen. That is not what it is about," said Ms. Brown.
Meeting Of Braintrust
An assignee from IBM, Ms. Brown sees the November meeting as a collection of the most informed minds in the business.
The list of participants is being kept under wraps. Workshop participants will witness presentations from seasoned
advocates of each advanced lithography technology. However, the real focus of the four-day event is interaction among
workshop members and the various lithography interests. Members will be encouraged to offer up their own sizing of the
risks, rewards and problems inherent in each technology--basically, to help Sematech and the various lithography interests
"understand what needs to be focused on as the next step."
White papers on each technology have been distributed to the November workshop attendees. Ms. Brown expects
participants to have done their homework, although most already exhibit a great wealth of knowledge on the subject.
"Right now, there are five technologies, all at different stages of development; all of them have more unknowns than
knowns," said Ms. Brown.
She is reluctant to release the names of the companies and/or experts who will champion the various lithography projects
evaluated at the workshop. Some technologies have somewhat obvious champions. X-ray lithography has long been the
subject of intense research at IBM, with some Defense Advanced Research Projects Agency (DARPA) assistance. Ion
projection lithography, mainly under development in Europe, is believed to be the turf of Siemens and others. Lucent
Technologies, the AT&T spinoff, recently dropped its development of X-ray and extreme-ultraviolet (EUV) lithography in
favor of E-beam projection technology, its Scalpel program. Lucent has received some DARPA assistance as well. Armed
by its recent private consortium announcement, Intel is the logical choice to champion EUV lithography, although Advanced
Micro Devices' and Motorola's participation in the EUV Limited Liability Co. (EUV LLC) adds their names to mix. Intel
has been working with the Sandia, Lawrence Livermore and Lawrence Berkeley National Labs--collectively known as the
newly-formed Virtual National Lab--to bring about this technology. The final contender is E-beam direct write, which
comprises several different technologies also called multi-beam, micro-column or cell projection. In the U.S., Etec Systems
comes up as the name of a possible champion. E-beam direct write is also the subject of U.S. government research and
work done at Cornell University, according to sources.
For each technology, the November meeting will have one person that is the champion of that technology. The champion
has put together a team of development people who are experts on the technology. The November meeting also features
breakout sessions and a complex survey for "how they see cost of ownership, what they see as the biggest risks (and) what
they see as the next set-up," said Ms. Brown.
In This Corner, The Contenders
The technology foremost in recent memory is extreme-ultraviolet lithography. News regarding EUV broke recently (EN,
Sept. 15) when a private consortium called the EUV Limited Liability Co. was formed by Intel, AMD and Motorola,
among others. The consortium has joined forces with the Department of Energy's Virtual National Laboratory to launch a
research project into EUV lithography, with an expected private investment of $250 million.
EUV lithography has been around for several years. At one time, EUV lithography was known as "soft" X-ray lithography.
Its relative similarity to a competing technology, X-ray lithography, was a factor in the name change. Driving the technology
throughout its history have been the three DoE labs. At one point, EUV research was supported by the Strategic Defense
Initiative, the Cold War-era military program widely known as "Star Wars." As of September 1996, federal funding for this
particular program ceased with the end of the fiscal year, prompting Intel to step in. The group hopes to move today's
developmental EUV technology into production factories early next decade. "We don't have any guarantees," said an AMD
spokesperson, referring to competing technologies like X-ray lithography and electron-beam lithography, being pushed by
IBM and Lucent, respectively. "We just happen to feel that this is an extremely promising technology."
A curious addition to the EUV corner is Motorola. Motorola was participating in the X-ray Proximity Lithography
Collaboration with IBM, Lucent and Lockheed Martin. The industry group had been formed in 1994 with AT&T (which
became Lucent) and Loral (which was later acquired by Lockheed Martin) among its founding members. That agreement, a
three-year contract, reached its conclusion this September. According to Joe Mogab, manager of advance process
development at Motorola, the company is still considering X-ray as a candidate for its next-generation process and the
news of the contract's conclusion "should not be taken as a symbol." The company will continue to work with IBM
informally, though, until the end of the year, he said. The reasoning for Motorola allowing the contract to terminate (the other
parties also agreed to conclude the contract) was a feeling that the company understood firsthand the manufacturing hurdles
X-ray faces. Motorola feels it can return to X-ray, if that becomes the solution, given this firsthand knowledge. The
company plans to address EUV, and "at least one additional technology" in the same firsthand exploratory manner,
according to Mr. Mogab. He could not give any more details as to which technology or technologies outside of EUV
Motorola was planning to explore, although he hoped to have details available soon.
What's Needed For EUV
For EUV, several technology breakthroughs are said to be necessary to bring the lithography to its commercial fruition.
Creating the next generation of photoresists, mask technology and optics each provides a set of roadblocks for the
advancement of this form of lithography. What is especially challenging for the EUV camp is believed to be the need for
radical new materials and technology for the resists and masks. Several sources familiar with the research indicated that the
photomasks necessary for EUV will demand a completely new approach. It is believed that the EUV photoresist will
require properties not found in current photoresists, necessitating a very complex compound chemical.
For EUV, there seems to be definitely more unknowns than knowns, with some considering the technology as the
generation after the successor to optical lithography. As one industry source put it laconically, "Other than needing a mask, a
light source, a resist and equipment, they are ready to go." Still, industry veterans do not discount EUV lithography, although
some believe its promise lies in the generation after this next one.
The EUV LLC itself has been the topic of questions and industry discussion regarding the structure of the private
consortium, which involves several foreign lithography companies, and the timing of its announcement. Mainly, the fear is
that Intel is paving the way for Nikon to become its primary contractor once the technology is viable. Whether Intel tried to
upstage the Sematech workshop with the timing of its EUV LLC announcement has been the subject of some speculation.
Karen Brown did not agree with such talk. According to Ms. Brown, the decision by Intel to publicize the private
consortium was not interpreted that way by Sematech.
"I don't think (the formation of the EUV LLC) has anything to do with the November workshop or things not moving fast
enough. It is an 'and,' not an 'or,' " she said. "For these technologies to become a reality, people are going to have to invest
money in them. My understanding of what happened is that the technology development was going to stop because it was
no longer going to be funded (by the government) and Intel, among other companies, thought that this technology had
potential...so they decided to put money into it to show, just as IBM has supported X-ray and Lucent supported Scalpel,
somebody had to pick up the shield."
Applying The Scalpel
Lucent Technologies, another of the three main contenders vying to create the technology for the future of lithography, has
been engaged in research on its contender for the last 10 years. At one time, the company invested in all three
technologies--EUV, X-ray and E-beam. The company decided to place its bet on E-beam because the other two did not fit
the timetable that Lucent felt was needed by the technology roadmaps of the industry, according to Lloyd Harriott, head of
advanced lithography research at Lucent Technologies/Bell Labs. Last year, the company took a long, hard look at each
technology and decided to go with E-beam lithography, the technology behind its Scalpel program.
According to Lucent, Scalpel has successfully been used to print semiconductor features at 0.08-micron. Scalpel uses
high-energy electron beams to expose the mask pattern onto the wafer. Since the electrons' wavelength is much smaller than
that for ultraviolet light, it is possible to print much smaller linewidths, which permits the manufacture of smaller devices.
Although work on electron-beam lithography has been going on since the 1970s, each of the systems produced had certain
inherent limitations. Work on Scalpel E-beam lithography has been going on at Bell Labs since 1989, with the current
Scalpel system under development the past two years.
Since deciding on E-beam, Lucent has already begun teaming up with semiconductor equipment and materials companies.
In December 1996, Lucent and Integrated Solutions, Inc. (ISI) signed a contract for ISI to develop and sell manufacturing
systems based on the Scalpel technology to make submicron ICs. This August, Lucent and Photronics agreed to explore
manufacturing issues involving masks for use in the Scalpel program. According to one source, Lucent has made a similar
pact with Du Pont Photomasks, Inc. (DPI), as well. And although the company has made no formal announcement, Lucent
has hinted through previous releases that MCNC of North Carolina supplies blank wafers for the Scalpel program,
something a company spokesperson acknowledged. Not to be forgotten, Olin Microelectronic Materials, which has been a
resist supplier to Lucent, is said to have some involvement with the company in the Scalpel program.
According to Mr. Harriott, Lucent, like Intel, has been drawing together membership for its own E-beam consortium. The
company expects to announce the formation of it sometime soon.
"The Sematech decision would be a very strong endorsement for Scalpel technology; hopefully, that decision is based on
technical information and data, and we look forward to supplying that," said Mr. Harriott.
Rising Awareness
Whether or not Scalpel's star has been on the rise is up to speculation, but its profile has been lifted over the past year.
While she declined to speculate herself, Ms. Brown did note that "today, there is much more interest in ion projection and
much more interest in Scalpel than at the beginning of the year," she said. Up until this year, EUV and X-ray received most
of the consideration. She did not want to comment further on any of the technologies to be presented at the meeting. Asked
about the recent upswing in Scalpel's profile, Mr. Harriott said, it was "just a matter of education--spreading the word." He
cited Lucent's efforts over the past year to increase the technology's profile through presentations, private meetings, daylong
reviews and other forms of stumping. "It has always looked very good on paper; the difference is this year the results that it
has gotten. It is one thing to show it on paper, but to demonstrate it on wafers, as we have this year, that really gets people
interested," said Mr. Harriott.
According to Mr. Harriott, the decision isn't something that can be boiled down to placing a bet. "It is really more about
consensus building," he said. This process of consensus building should help set some directions in the industry for
equipment people to look at and aim their businesses in particular directions. But the decision does have some funding
ramifications that could cripple a technology in the long run.
"All of these technologies require quite a level of funding support. And if a couple of them don't receive the funding support,
it will be harder for them to maintain funding and they will slip away," said Mr. Harriott.
***CONTINUED IN NEXT POST*** |
