Lithography sector prepares for 0.25-micron
09/01/1997
Lithography sector prepares for 0.25-micron
While total stepper sales may have declined for the year, the lithography segment of the semiconductor equipment industry seemed especially vital at SEMICON/West `97, driven by preparations for the 0.25-micron era. Shipments of top-of-the line DUV steppers and step-and-scan systems were reported to number about 100 by Karen Brown, SEMATECH director of lithography, with deliveries tripling in 1998. One industry observer even expressed fears that the stampede toward smaller CDs would lead to a "technology glut" in which there would soon be too much capacity at the leading-edge, which might cause another downturn in equipment sales.
Be that as it may, most exposure tool companies were up-beat, relishing the challenge of producing more DUV systems than predicted. John C. Wiesner, senior VP of Nikon Precision Inc., described the second-generation Nikon NSR-S202A KrF laser scanner, capable of 85 wph at 0.25 micron and below. Optically identical to the earlier NSR-S201, the S202A achieves higher throughput, lower cost, and better maintainability with an entirely new "noncontact" stage in which the wafer chuck is supported on air-bearing pucks directly referenced to a very flat granite plate. The wafer chuck assembly is then moved in x and y by an actuator arm. The new design reduces focus shifts, tilt, and yaw and is capable of scanning at up to 140 mm/sec. Wiesner reported that Nikon would produce 200 DUV tools in 1997, with one-third of the 1998 production being step-and-scan systems. He also observed that "CMP has silenced concern about depth of focus in the 0.25-micron era. The industry now seems to feel that 0.7 micron is adequate, and thus chip development is accelerating. Equipment development, however, still takes at least six months."
Nikon Precision also introduced "Nikonnest," a computer retro-fit that will extend the lifetimes of Nikon`s 6-10 model steppers. By replacing the obsolete PDP-8 computers with 200 MHz Pentium Pro-based motherboards equipped with a 36 MHz co-processors, Nikonnest speeds operation by 200 percent. A Windows-NT operating system will ship later this year.
In other lithography-related news:
Edward A. Dohring, president of SVG Lithography Systems Inc., reported plans to deliver more than 70 Micrascan III step-and-scan systems in 1997, ramping to 200 or more by the 4Q98. The new 0.6 NA catadioptric scanner will sell typically for $7.2 million, depending on options. Prototype ArF versions are also under development, with initial packages offered for $10.5 million.
Integrated Solutions Inc., developer of the small field ArF stepper widely used in process development, showed their mag-lev stage, which promises improved performance in conventional stepper applications. The wafer carrier is supported and fine-aligned by a magnetic suspension supported by a conventional x-y stage stack.
Jim Greeneich of ASM Lithography observed that 300-mm wafers do not automatically increase the productivity of lithography systems, unlike other wafer processing activities. Because each field of a wafer is exposed separately, decreasing cost of ownership requires improvements in stage acceleration and scan-speed, laser power, and overhead time. Achieving 100-wph throughput for 300-mm wafers is less difficult if the number of exposure fields can be reduced by exploiting the larger maximum field size of a scanning system. However, doing so requires 9-in. reticles and no manufacturing infrastructure currently exists for such large high-resolution masks. Track manufacturers will also have to improve the throughput of their 300-mm systems.
In light of the difficulties in achieving the theoretical advantages of step-and-scan exposure in the near-term, Phil Ware of Canon advocated continued use of DUV steppers. Canon is delivering four 300-mm DUV steppers in 1997 for process development, but recognizes the productivity advantages of field sizes larger than 22 ? 22 mm. According to C. Douglas Marsh, president of ASML`s US Operations, the 300 series is the last ASML stepper, and a suite of 300-mm ASML step-and-scan tools will be available by 1999, including a cost-effective i-line scanner.
A leader in low-cost lithography for mix-and-match, Ultratech Corp., announced two new subsidiaries to commercialize noncore businesses. UltraBeam Lithography Inc. will promote a new vector-scanned electron beam lithography system for mask-making, based on technology obtained in the purchase of Lepton, Inc. The new system is expected to be more compatible with the rigors of 0.25-micron mask production than the R&D-based Lepton devices. Verdant Technologies will endeavor to produce production-worthy manufacturing tools based on Ultratech`s P-GILD doping process, developed in collaboration with DARPA and SEMATECH. By melting and recrystallizing a well-defined region of a wafer using a projected pattern of laser light in the presence of a doping gas, P-GILD produces shallow doped regions without the need for further annealing. A related process, termed gas immersion laser annealing (GILA) potentially replaces conventional RTP in forming low resistance silicides for 0.13-micron geometries. Both processes eliminate conventional processing steps by exploiting the resolution and power of images projected by excimer steppers.
Olin Microelectronic Materials announced a collaboration with Lucent Technologies to commercialize a single-layer, 193-nm resist based on the work of Elsa Reichmanis at Bell Labs. The new resist design complements the bilayer 193-nm resists developed by Olin internally and extends Olin`s strategy to lead in the commercialization of advanced photoresists. Murrae J. Bowden, the recently-appointed director of R&D of Olin`s Photopolymers Group, commented that Olin is supporting an aggressive R&D timetable to serve rapidly-evolving process demands in the semiconductor industry. Olin also introduced two new "production-capable" DUV resists - ARCH 5200 for 0.30-micron contact holes and ARCH 5600 for 0.25- and 0.20-micron gate structures.
Track maker FSI Inc. showed its new model 2200 cluster with two robots, capable of processing 100, 200-mm wafers/hour in an 80 ft2 area. This DUV-capable system is enclosed in its own Class-1 minienvironment, and is protected from amine contamination by Donaldson charcoal filters. It is also designed to minimize downtime during recipe changes. Pete Steege of the FSI marketing department hinted at an even more compact track system designed for either 200- or 300-mm wafers in the same footprint.
KLA-Tencor unveiled the KLA-2138, a new die-to-die inspection tool for CMP wafers. The ultra-broadband illumination and recipe-defined segmented-auto-thresholding defect criterion logic is said to improve defect capture rates by 30-4000 percent while reducing false defects from film thickness and texture variations.
ETEC described the ALTA-3500 scanned laser mask pattern generation tool. A new 0.8 NA imaging lens and a four-offset multipass raster exposure strategy (and other improvements) are predicted to facilitate production of 0.25-micron-generation masks and multilevel phase-shift masks by laser-writing. Micronic Laser Systems AB, ETEC`s Swedish competitor, described a large-area photomask technology suitable for flat panel displays and noncritical semiconductor layers in the 9-in. reticle era.
Production of DUV optics may be aided by a new dual-wavelength UV interferometer system developed by WYKO. By measuring the transmitted wavefront of a component near the operating wavelength, the WYKO DUV Optical Component Inspection System should facilitate production of undistorting lenses.
As SEMATECH`s Brown noted, the transition to DUV is now. Equipment companies are competing to be first to the market with production-capable DUV systems and semiconductor producers are competing to be first to market with higher value chips produced by these systems. This transition is fueling equipment sales now; a slow-down may come as customers work down the learning curve of their new processes. But it may not! Moore`s law may really have accelerated, with the demand for even more advanced technologies stressing the R&D engineers just as the production people begin to learn how to deliver the last innovation in volume. Second and third generation DUV technology is becoming available, with the expected real-world improvements. If all segments of the industry do what is required to realize predicted productivity gains, the next boom could really have begun. M.D.L.