Developers Call for Backing of High-Index Immersion Lithography

Aaron Hand, Executive Editor, Electronic Media -- Semiconductor International, 10/9/2007 11:36:00 PM

While past years have focused more on solving the challenges of water-based immersion lithography, this year’s International Symposium on Immersion Lithography, being held this week in Keystone, Colo., has turned much of its attention to further extensions of optical lithography, including high-index immersion lithography and double patterning.

After a full afternoon of presentations on the latest progress in high-index immersion — including the readiness of high-index lens elements, fluids and tools — a group of panelists representing a cross-section of the industry gathered this evening to discuss the pivotal questions for the technology: What is the biggest roadblock to successful high-index immersion lithography? What needs to be done to accelerate the technology’s research and implementation? Will it ever be used by the industry?

Moderated by Bryan Rice, an Intel assignee to Sematech and chair of this year’s symposium, the panel also included L.J. Chen of Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC), Harry Sewell of ASML, Roger French of DuPont, Will Conley of Freescale Semiconductor, Soichi Owa of Nikon, and John Hoffnagle of IBM’s Almaden Research Center.

TSMC’s Chen, who gave the first presentation, said that he considers high-index immersion lithography to be promising in terms of cost-effectiveness, but timing is an issue. He is pushing for high-index scanners to be available before the first quarter of 2009 with a numerical aperture (NA) of 1.85. Although others scoff at the feasibility of a 1.85 NA system (most consider 1.7 as the goal), Chen insists that 1.85 is needed to keep k₁ above 0.3, thereby maintaining a workable process window.

But there is little chance of high-index immersion staying within such a demanding time frame at the rate it’s going, especially to get past an NA of 1.55. Currently, water immersion has reached what is considered to be its maximum attainable NA of 1.35. To get to 1.55 (probably more like 1.45-1.50 of actual available NA) requires an immersion fluid with a refractive index of 1.65 — second-generation fluids that have been demonstrated already by companies such as DuPont and JSR Micro. To reach the next goal of 1.7 NA, a third-generation fluid must be developed with a refractive index of 1.8, and a high-index lens material is also needed. Lutetium aluminum garnet (LuAG) is considered the leading candidate.

However, quite a lot of progress still needs to be made to hit on a feasible Gen3 fluid, and a presentation earlier in the day from leading LuAG developer Schott indicated that they have not yet gotten the results they had hoped for, with progress running about half a year behind schedule.

For many, the availability of the high-index lens material is the biggest roadblock hindering high-index immersion acceptance. It’s the key question holding up development in other areas. ASML, for example, can’t make a move toward developing a high-index immersion scanner until Carl Zeiss, its optics partner, gets hold of the necessary lens material, according to Sewell. Nikon’s Owa echoed his concern about the lens material, adding that development acceleration needs to happen to address the second roadblock, which is timing. “If the end user rejects a 2011 timing, then there’s no market for high index,” he said. Therefore, the technology needs to be developed faster, he added, noting that he is not optimistic.

But DuPont’s French voiced his frustrations with these types of assumptions, concerned about the bind that high-index immersion development has gotten into by waiting for LuAG development. Instead, he urged, the industry needs to reconsider the prior assumptions that were entered into two years ago. Particularly considering the bifurcation of the roadmap into various device needs, he doesn’t understand why the timing has to come down to a binary yes or no.

French argued that the toolmakers need not wait for the perfect last lens element to develop a high-index immersion exposure tool. Instead, much as they have done with water-based immersion tools, they could introduce incremental improvements, developing first an exposure tool that uses Gen2 immersion fluids with existing 1.2 or 1.35 NA lenses.

Although ASML and Nikon have both taken similar approaches with their water-based immersion tool introductions, ASML’s Sewell noted that the situation with high-index immersion was different. Considering the cost of developing a tool with a relatively minimal gain in NA, he said, they needed more confidence that a feasible LuAG material would indeed be coming. “It makes a big difference in the attractiveness of an interim solution,” he said. “What underpins it is this question of will there be a high-index lens material to follow? Nobody’s going to spend the money to go from 1.35 to 1.45.”

Actually, according to Freescale’s Conley, the solution is easy: Spend more on high-index immersion development and less on extreme ultraviolet (EUV) lithography. He insists that high-index immersion will ultimately be used by chipmakers because there will be nothing else for them to use (i.e., EUV will never come to fruition). “If you think EUV is going to happen, then you’re going down the right path,” he said. “But if you think it’s not going to happen, then you have to do something else.”

Compared with the billions of dollars that have been spent on EUV development, high-index immersion would require a relatively small amount of money to accelerate development to the levels necessary to meet 2009 goals. Peter Krüll of Schott estimated that an investment of ~$5M could accelerate learning on LuAG by a factor of 2.

Rice pointed out that materials readiness and the issue of timing has been a recurring theme for high-index immersion. End users want a solution by a given time, but that solution requires a significant investment in capital. In that case, the industry needs to choose to support the technology, he said. He urged symposium attendees and presenters to go back to their respective management teams and do what they could to get their companies to invest in making high-index immersion happen faster.

Although companies continue to pour money into EUV development, much of the focus has turned lately to double patterning and double exposure techniques for increasing pattern density. While many of the Tuesday presentations at the Immersion Symposium focused on high-index immersion status, Thursday’s program includes a session on double exposure/double patterning, with another panel discussion to round out the agenda. In fact, the audience may again hear similar woes about the readiness of double patterning, and the investment that’s needed to speed development.