High-Index Immersion Not a Likely Winner

With the latest tool announcements from ASML and Nikon, numerical apertures (NAs) have essentially reached their limit for water-based immersion lithography. The announcements centered around SEMICON West, when Nikon introduced its NSR-S610C, with an NA of 1.3, and ASML introduced the Twinscan XT:1900i, with an NA of 1.35.

The fundamental limit for any NA is the lowest refractive index in a lithography system. In current systems, that lowest number is the 1.44 index of water. About 90% of that index is what is considered the practical NA limit, so therefore ~1.3 for water-based immersion. With this in mind, the latest introductions — the fifth-generation immersion tool for ASML and the fourth-generation for Nikon — will be the final full tool introductions for immersion lithography from these two suppliers unless the industry moves on to higher-index materials.

The question on many people's minds these days is what is likely to follow water-based immersion. In a presentation at SEMICON West, Gene Fuller, principal engineer at Nikon Precision Inc., listed three main choices under serious consideration: 1) Continue to increase the NA and improve resolution through higher-index fluids, resists and lens materials; 2) retain the same water immersion scheme, but adjust the process with something like double exposure or double patterning; and 3) extreme ultraviolet (EUV) lithography. He also mentioned the possibility of "something completely new." Whether nanoimprint lithography qualifies as something new he didn't say, but certainly some in the industry are putting their bets on that 1× technology.

Fuller detailed the pursuit of higher-index materials to extend immersion lithography's capabilities. With the latest advances in immersion fluids with a refractive index of 1.65, researchers are pursuing higher-index resists and lens elements as well, and toolmakers are employing catadioptric designs to better take advantage of NA capabilities. Although considerable progress has been made in achieving high-index fluids, the refractive index is not the only barrier. The fluid must also have high transparency, low variability in relation to temperature, good viscosity characteristics for high-speed scanning, and it must be recyclable.

The high-index fluids are a step in the right direction, but that change alone enables only an incremental increase in NA. Even with higher-index lens materials thrown into the mix, a high-index resist would be needed to take NA higher than 1.5 (Table).

Based on a practical lens diameter, and assuming high-index resists and a 6 in. mask size, high-index immersion (fluid=1.64, lens material=1.91) could take the NA to ~1.5, and ultrahigh-index immersion (fluid=1.82, lens material=2.10) could take the NA to ~1.7. (Source: Nikon Precision)

In general, the lithography tool suppliers have not made it a secret that they don't put a lot of stock in high-index lens materials or fluids being the most reasonable path to take to extend immersion lithography. "We are still pursuing this, and I'm sure other people in the world are pursuing this, but there are a lot of challenges," Fuller said.

Probably the biggest challenge is not technical, but economic. Although fluid suppliers have the opportunity to provide a relatively pricey consumable to the market, the lens suppliers would be presented with an extremely limited market. The industry will not move ahead with only an improvement in the fluid's refractive index. The moment that rises to 1.65, it's the lens that now becomes critical, limiting the system's capability for higher resolution with quartz glass material's index of 1.57.

Moving to higher-index materials offers a resolution gain of ~4%, noted Martin van den Brink, executive vice president of marketing and technology at ASML. Compare this with the move from i-line to DUV lithography, or the move from 248 to 193 nm lithography — in both cases, it was almost a factor of 3 improvement. "You go now from dry to wet, and it's not a factor of 3, it's 40%," he said. "But for that 40%, you stay in the same wavelength. Every time you change wavelengths, you have more material challenges."

Even with high-index lens materials of ~1.8, the resolution improvement would be only ~15%. As a reminder of how the industry feels about relatively low improvements in resolution, van den Brink trotted out lithography's most recent failure — 157 nm lithography, which offered barely a 20% resolution advantage over 193 nm lithography. "The 20% advantage was too low to motivate customers to go through all the barriers," he said.

High-index lens materials for immersion 193 nm are expected to be used only in the last lens, which creates a market of only a few hundred kilograms worldwide, van den Brink noted. "What is the business incentive for the material suppliers to do this?"

Despite all the arguments against it, however, nobody is willing to say that high-index immersion is not going to happen. "I cannot say it will not happen, because the alternative for this is double patterning, which is expensive," van den Brink said. "The other alternative for this is EUV, which has numerous other challenges."