**PART 2** Photolithography in the limelight. From Electronic News.
sumnet.com
Japanese Interest
Electron-beam lithography has caught the eye of several Japanese interests as well. In September, NEC reported its
development of an operational metal oxide semiconductor (MOS) transistor, with an effective gate length of 14 nanometers.
The company claimed that the successful operation of the transistor in experiments meant that the first step toward
developing 10-terabit (10 trillion bits) memory devices had taken place. The new transistor, an electrically variable shallow
junction metal-oxide semiconductor field effect transistor (EJ-MOSFET), utilized a two-gate construction. By using a
calixarene resist, developed by NEC, with what the company termed "extremely" advanced electron-beam lithography.
NEC was able to realize what it claims to be "the world's first" 14nm polysilicon gate EJ-MOSFET, using an E-beam
system and special resist of its own design, according to a company spokesperson. Further information on the enabling
technologies could not be gathered by press time.
IBM has been working on its contender, X-ray lithography (XRL), since the early 1980s. Although X-ray was the subject
of exploratory research at IBM's Brookhaven facility throughout the 1980s (the company demonstrated proof of concept
there in 1988), the XRL program picked up steam when the company committed to build a facility, the Advanced
Lithography Facility (ALF), in 1987. Work on ALF started in the second half of 1988 and was completed in 1990. Since
then, IBM has installed a synchrotron ring from Oxford Instruments, two Karl Suss (now Suss Advanced Lithography)
X-ray steppers and one SVG Lithography Systems (SVGL) X-ray stepper, and has routinely exposed wafers at the facility.
According to Michael Polcari, director of silicon technology at IBM Research, and George Gomba, senior manager for
lithography technology development, IBM's X-ray fits a range of 180 to 150 nanometers. The company has gone further,
though. "We've built circuits at 100 nanometers," said Mr. Gomba. Mr. Polcari added, " We've made functional devices
with 80-nanometer X-ray lithography."
Instead of an optical source like a laser, XRL uses X-rays provided by a synchrotron as its source to write the wafers.
Another source for XRL, a point source, is not being researched by IBM for various technical concerns. IBM feels this
technology will replace optical lithography because it has more knowns than competing technologies.
"All of the technologies have pluses and minuses, but the reason why we believe in X-ray lithography is that we have
demonstrated proof of lithography, and I think none of the other technologies have done that," said Mr. Polcari.
The company believes that it has demonstrated that 1x X-ray litho is viable and feels that it has advanced X-ray lithography
to a point years ahead of the other next-generation contenders.
X-Ray's Established Base
"The other technologies are where X-ray was 10 years ago. X-ray has an existing and an evolving infrastructure; the other
technologies don't have that. (X-ray) has a large established base of worldwide development experiences (and) end-users
that have used X-ray and understand it," said Mr. Polcari, who ticked off a list of other companies familiar with XRL,
including Motorola, Mitsubishi, NEC and Toshiba.
"Nobody has touched the other technologies that have been proposed. The tools aren't available and won't be available for
five or 10 years, most likely," said Mr. Polcari. "X-ray is a very mature technology relative to the other technologies."
It is precisely this maturity that has been a cause for concern among some. Responding to these questions about the speed
at which XRL technology has come along, Mr. Polcari pointed out that work on deep-ultraviolet lithography started in the
early 1980s and just now has become an industry standard. "It just takes a long time to bring this technologies to fruition.
Innovations in this industry don't happen in a year or two; they take a long time to become a standard."
The most compelling argument for XRL, according to Messrs. Polcari and Gomba, is that the technology has been proven
and roadmapped. Its technological hurdles for production and extendibility are understood, and solutions have been
identified.
"There is no doubt a lot of fundamental engineering work has to be accomplished (for XRL). All of the technologies have a
lot of work ahead of them. We don't have a total solution today, but (XRL does) have major elements of the infrastructure
in place," Mr. Polcari said.
"I look at ours as a path of engineering and evolution, more than of invention," added Mr. Gomba, referring to competing
technologies like E-beam and EUV. Mr. Gomba will represent IBM at Sematech's November workshop.
Asked about the status of the X-ray Proximity Lithography Collaboration, Mr. Polcari said that the three-year contract
regarding this work concluded this September. Whether Motorola would return to X-ray lithography given their involvement
in the EUV LLC was something neither Mr. Polcari nor Mr. Gomba wished to speculate on.
"I think all the parties agree that it was a successful endeavor, and that we accomplished what we set out to accomplish,"
said Mr. Polcari.
Prior to the completion of the collaboration contract, Lucent bowed out when it ceased its work in X-ray and EUV
lithography in favor of Scalpel.
Asked if IBM would form its own consortium like the one recently formed by Intel, AMD and Motorola, an IBM
spokesperson noted that consortiums generally form at an exploratory phase in a given technology. Though a consortium of
X-ray interests is not in the works, IBM is talking to other industry members about working together on the technology.
"We certainly have discussions with other people working on this technology; there are certain aspects that we will continue
working on," said Mr. Polcari.
The Stepper Makers
In the field of X-ray lithography, IBM is not alone. Suss Advanced Lithography (SAL) is said to have built 14 X-ray
lithography steppers during the last 10 years. Canon is said to have made one. SVGL is said to have made at least one.
Nikon is said to be looking into it, although the company is not forthcoming with any details. A Silicon Valley start-up in
X-ray lithography equipment and masks, Micronix Corp., went out of business in the 1980s. Another U.S. X-ray litho
house, Hampshire Instruments, closed its doors in 1993. In all, an industry source said the number of X-ray steppers built to
date numbers in the 20s.
SAL has been at work on X-ray systems since 1982. Formerly a part of Karl Suss, the Vermont-based company is now
privately held and has been since 1994. Two SAL X-ray steppers from its XRS Series 2000 (200mm) systems are located
at IBM.
"Right now, any X-ray stepper that we make can print 0.07 lines. Alignment and overlay on the best is good enough for
0.13-micron devices. It doesn't do any good to make a fine line if you can't align the layers under that. And that is true of
anybody," said Ted Bettes, VP of marketing at SAL.
The company will send its president, Peter Hines, to the Sematech workshop in November.
According to SAL, X-ray, as opposed to EUV and E-beam, is ready with real resists, masks and steppers right now. And
"for the most part, the stepper you would get today would be extendable to 0.1-micron without buying a new stepper," said
Mr. Bettes, who added that with upgrades of software, hardware and mask design, any SAL X-ray system could reach
0.1-micron.
But Mr. Bettes hesitated to predict that X-ray would hold the title of choice for a long time. "We may make an EUV
stepper some day," he said.
SAL's Fourth Generation
But SAL is already in its fourth generation of X-ray stepper, with the forthcoming 3000 series receiving some interest from
prospective buyers. SAL has most recently supplied Mitsubishi with one stepper and NTT with two machines, with a
synchrotron light source. SAL will deliver a point-source stepper as part of a DARPA contract awarded to Sanders, a
Lockheed Martin company.
At this point, X-ray lithography has two major sources for synchrotrons, Oxford Instruments in England and Sumitomo in
Japan. The synchrotrons serve up to 20 steppers. SAL is working with another company on developing point-source
technology. The point source would serve one stepper exclusively. Mr. Bettes noted that SAL's X-ray systems can use
either a synchrotron or a point source.
As far as other existing technology that applies to X-ray, Mr. Bettes added that the 248nm resist is the same thing that you
would use for X-ray. He also noted that X-ray masks are available; both IBM and NTT make X-ray masks. Mitsubishi
also makes X-ray masks for its own program. "So it is not an unknown technology. You won't find anyone making an EUV
mask. You will find some people trying to make 193(nm), though," said Mr. Bettes.
With so few X-ray steppers out, SAL has survived because "we are a very small company and we don't need a lot to
support us. A system or two a year, that is all we need--that is not our business plan--but that is all we need," said Mr.
Bettes.
In the area of E-beam direct write lithography, Etec Systems is currently working on two technologies, multi-source and
micro-column. Multi-source is similar to the laser pattern generator architecture currently in use in Etec's two laser beam
product families, ALTA and CORE. Its objective is to extend the resolution of those tools within the system architecture.
The focus is on developing photo electron emitters that convert light into electrons. The advantage to this technology, of
course, is that Etec already has the system architecture available, particularly in the writing strategy. The challenge, on the
other hand, lies in understanding the space charge effects of E-beam and how these may limit resolution and throughput.
Because the beams are in such close proximity, it is anticipated that space charge effects will manifest themselves early.
Micro-column technology is based on the idea of scaling down the beam delivery system. It refers to the design and
fabrication of electron-optical focused electron-beam devices with column lengths of several millimeters, as opposed to
current technologies, which are much larger. This technology was originally developed by IBM. In November 1996, Etec
acquired the intellectual property and equipment associated with it, and eventually hired much of the IBM team responsible
for it, including its leader, Philip Chang. Etec is developing this technology for mask writing, with the idea that it may be
possible to further develop it into a direct-write technology.
Mark Gesley, Etec's VP of technology, will serve at the Sematech workshop as the champion of E-beam direct-write
technology. He characterized the attitudes of other conference members toward direct write as cautious. "I think there is
some interest," he said. "I think there is some skepticism also, given that E-beam direct write has been touted for 20 years
as being a lithography that might be adopted in semiconductor manufacturing. The one major challenge has been the low
throughput. As long as the incumbent wafer lithography has a method that can put out 20, 40 or 60 wafers per hour,
E-beam will not be adopted."
Lithography Concerns Take A Ringside Seat
Ed Grady, VP and GM of KLA-Tencor's Rapid division, has been working with both the EUV and Scalpel technologies
for the past seven years. He also has specific knowledge of E-beam and X-ray lithography. He said the Sematech
conference, while providing a helpful forum for discussion of future lithography technologies, will not ultimately result in a
solution to the lithography question.
"We have to agree that ultimately the decision on which technology is going to win is going to be based on market forces
and economics," Mr. Grady said. "This decision will be made, not by a forum, not by a group of technical professionals, but
by business economics."
"It doesn't surprise me that they have not yet reached a consensus," said Mark Bigelow of ASM Lithography. "It may not
yet be time to throw all the chips into one basket. There's probably more research and cost studies that need to be done."
Art Zafiropoulo, CEO of Ultratech Stepper, agreed, noting, "None of these technologies are gimmes. They all have unique
problems to be resolved."
Phil Ware, Canon USA director of technical marketing, said he does not expect the Sematech conference to reach a
consensus in November, either. "It doesn't seem likely they will be able to narrow down their focus to a single technology,"
Mr. Ware said. "It seems more likely to me that they will simply narrow down the focus to two or three technologies. There
are still too many major proponents of each leading technology. Among Canon customers, we have a strong advocate for
each leading candidate."
Mr. Grady agreed that it is possible that one technology may never clearly emerge, and that a combination of several of the
leading candidates may be used. He said KLA-Tencor appears to be in good shape regardless of which technology is
ultimately chosen. The company has head starts on products in all the leading candidate areas, he said.
Mr. Ware said Canon does not want to limit its options either, at this point. He said Canon has been a participant in the
Sematech workshops leading up to the November conference and that Canon's research division continues to pursue all top
technologies. "We've been a player for several years in 1x X-ray lithography," he said. "We would certainly have an
advantage if that technology were picked. We also have a strong start in E-beam. I would say we don't have a favorite at
this point. We are seriously considering all the options." Canon is also listed among those mentioned in the EUV LLC.
Mr. Bigelow said ASML would be keeping an eye out for the results of Sematech's November conference, but he is not
expecting any earth-shattering decisions. "We'll be paying attention," he said. "But we're obviously going to keep our options
open. Sometimes, reading the future is difficult."
Sematech's Role Applauded
"I applaud Sematech for doing this," Mr. Ware said. "They are taking a needed leadership role. (But) if they don't achieve a
consensus, then the same problem faces all vendors. Do we spread our resources three ways, when only one will ultimately
be chosen?"
"The infrastructure can't afford to invest in the final solution until we know that it is the final solution," Mr. Grady said. "No
company can work on four mask technologies at the same time. It's just cost-prohibitive. You can leverage whatever
development you do so that it can be applied to all the technologies. That's something KLA-Tencor will do."
Mr. Grady said he expects KLA-Tencor to receive some Sematech funding for the development of future technologies.
"That's the real value of what Sematech does," he said. "Sematech can accelerate the rate of development by providing
funding, which takes some of the risk off companies and allows the technology to be developed sooner. I would expect that
KLA-Tencor will be working with Sematech and member companies going forward on requirements for masks, whatever
technology is chosen."
Mr. Bigelow believes it's possible that more than one of today's leading next-generation lithography technologies may have
its moment in the sun. "I think all three leading candidates may play out in time," Mr. Bigelow said, "meaning EUV, X-ray
and Scalpel may all have their moments in history. Ten years from now, there may be requirements for lithography that we
can't predict today."
Mr. Grady said while it's hard to predict a timetable for when next-generation lithography technology will emerge, he
expects the industry to continue pushing optical lithography until it is no longer cost-effective. In fact, he said, he wouldn't be
surprised if optical lithography technology was pushed all the way down to 100nm, well past the point where most industry
observers believe it can go. "I think there are ways that people will come up that could push the technology that far," he
said. "Economics will drive it. Lower cost per function is the driving force in the industry."
John Wiesner, Nikon Precision's senior VP of technology, said he expects the conference to come up with a prioritized list
of technologies from the standpoint of which will be the next-generation lithography technology and which may be used at
another time down the road.
"Nikon is participating with interest and with vigor," Mr. Wiesner said. "I don't think the mission is to work up development
programs, but rather to prioritize the initial prospect of these technologies. Since each of these technologies is terribly
expensive, it is not likely that the world can fund the development of all these competing technologies. My guess is that
eventually one or two competing technologies will emerge as the winners."
"I don't think it's proper for Sematech or anyone else to unilaterally pick winners and losers at this point," said Mr. Wiesner.
And according to Ms. Brown, Sematech will not.
"The manufacturers and the equipment suppliers will make the so-called 'Decision of the Century.' Sematech is helping them
by providing a forum where data can be discussed and the process can move forward," said Ms. Brown. "It is not
Sematech's timeline. It's one that we agree with around the world, throughout the semiconductor industry."
Regardless of how companies posture themselves, the Sematech workshop in November is shaping up as pivotal for the
industry, even if it simply places more hard data in the hands of the industry's leaders.
"Maybe that makes the job more difficult, but I think that makes it better for the industry," said Ms. Brown.
As far as the past track record at Sematech, which industry members recalled with little doubt as to Sematech's ability to
move the industry forward, the consortium's director of lithography said, "The track record with Sematech is excellent." Ms.
Brown, a part of the organization since 1994, did add that Sematech has changed its tactics in advancing technology within
the semiconductor industry, though. The consortium no longer tries to work beyond the timeline set for emerging
technologies. "I think that was part of the problem with GCA; I-line was ahead of the time and wasn't readily needed," said
Ms. Brown, referring to Sematech's prior backing of GCA, a company that failed in 1993 after receiving millions of dollars
in R&D funding from Sematech.
Sematech is certainly handling things differently this time; in part, the circumstances have necessitated it. The future of
lithography might not be described so much as a move as a jump. With that in mind, industry groups like Sematech will
approach the issue with caution--looking before they leap.
"(The lithography question) is a big deal this time, because when we went from G-line to I-line to 248 (nanometers), it was
kind of a straight path. But now, we have five very different technologies on the table, and we don't know which one is
going to be the solution and which one is going to be on the factory floor," said Ms. Brown. Come November, the industry
may have a clearer picture. |
