by James Montgomery, News Editor
In what should be the last roundup of water-based 193nm immersion lithography systems, two of the big lithography tool vendors are coming out with their latest tools targeting 45nm (logic) and 32nm (memory) semiconductor manufacturing. Nikon Corp. says it will begin shipments of its new NSR-S610C ArF immersion scanner, with a 1.30NA projection lens, by the end of this year. ASML has indicated its latest immersion tool, the Twinscan XT:1900i with 1.35NA optics, is being prepped for deliveries sometime in mid-2007. Both systems are expected to clock in with throughput of ~130 wafers/hour, although preliminary data from each company suggests ASML's tool achieves exposure rates of 125 exposure/wafer, vs. 76 exposure fields with Nikon's tool. Each tool has an exposure area of 26x33, and both companies claim to have eliminated immersion specific defects, enabling total wafer defects in the single digits, equivalent to today's leading-edge dry scanners.
Nikon's system utilizes a multi-axial catadioptric lens design, and the "tandem-stage" platform used with the company's previous-generation immersion system (1.07NA), which combines a moving wafer-exposure stage with a dedicated calibration stage (see WaferNews, V12n10, March 7, 2005). The system uses a proprietary nozzle design, and an airless fluid handling process (300m/min) for water containment, instead of an air curtain which Nikon claims contributes to evaporation and deforming of the wafer holder, thus impacting the overlay. The tool incorporates Nikon's Polano fourth-generation polarization technology. Alignment accuracy has been reduced to ?6.5nm.
Meanwhile, ASML is readying what it calls the "last-generation" water-based immersion lithography system, preceding a new crop of immersion systems utilizing high-index fluids besides water, and an influx of new EUV-based systems. The Twinscan XT:1900i offers optics of NA=1.35 and throughput of 131 wafers/hr (125 exposures/wafer), targeting 45nm chipmaking processes and 40nm shrink. (Nikon, in response, claims that "a 1.35NA lens provides negligible resolution improvement and smaller depth-of-focus, with far more engineering risk and elevated flare levels.")
In a pre-briefing, Peter Jenkins, VP of marketing at ASML, noted that the company will continue to make improvements to its final water-based immersion tool, addressing productivity and overlay, and the inline catadioptric lens design will be scalable to new fluids. But he also pointed out that the short window between development of high-index fluids and the need for EUV lithography suggests there's little benefit there for either ASML or its customers.
"We have not been very aggressive in our belief that high-index NA is a logical step for us and the for industry to make," he said in the prebriefing, explaining that there would be limited shrink improvement, plus additional costs of developing new processes required by these new fluids, and time required for the optical materials to mature. Also, customers want to amortize tools for two or three generations of production, but high-index fluids are likely to deliver only a half-node shrink, whether for single-pass or double-patterning. "There's only going to be a limited step forward in terms of resolution," Jenkins said.
Instead of focusing on high-index fluids, Jenkins said ASML will focus on parallel efforts, namely double-patterning (including improvements to the 1900i to enable two-pass processing, and halve the shrink capability of the high-NA), and a transition to next-generation EUV lithography that can offers the prospect of multiple generations of production use (ASML is prepping two EUV tools to ship to IMEC and Albany Nanotech, targeting the 32nm node). -- J.M.