Immersion Lithography Looks Like Reality

There is still time to jump off the fast train to immersion lithography, but most of the industry seems pretty content to stay on board. It has been less than two years that the industry has been seriously considering immersion techniques for inclusion on the roadmap, but significant changes to the roadmap are already underway.

ASML (Veldhoven, Netherlands) and TSMC (Hsinchu, Taiwan) made the latest bold move. ASML introduced early last month the industry's first immersion lithography system, also announcing that TSMC would be its first customer. The foundry powerhouse, which has been the key driver for immersion lithography, is scheduled to take delivery of ASML's Twinscan XT:1250i in the third quarter of this year.

The pre-production lithography scanner is based on ASML's Twinscan platform, and follows just behind the company's dry XT:1250 ArF tool. Like the dry version, it is a 0.85 NA, 193 nm system, with wafer measurement done on one stage while the imaging process is carried out on the other. In the wet version, the wafer measurement stage remains the same — only the imaging half of the machine converting over to an immersion technique.

Immersion lithography replaces the air between the final imaging lens and the wafer with water or another fluid. Because water has a higher refractive index than does air, this effectively increases the NA of the imaging system and improves depth of focus, thereby taking the k₁ factor back up to more workable levels. This is comparable to lowering the light source wavelength from 193 to 134 nm.

However, ASML's first immersion tool does not take full advantage of underwater capabilities. Both the 1250 and 1250i operate at the 65 nm node, with half-pitch resolution at 70 nm. At this point in the game, the 1250i is geared toward letting customers test and qualify immersion processes.

"What the initial immersion tools will give, is that they'll improve the depth of focus," said Boudewijn Sluijk, director of product marketing for ASML. "We're not gaining immediate resolution, but we are gaining depth of focus." This will allow users to get a higher yield, or to push their designs a little further, he said, adding, "With larger NAs, immersion will deliver on resolution and also on depth of focus."

With the immersion buzz surrounding SEMICON West last year, the three major tool suppliers promised to announce their decisions (yea or nay) on the feasibility of immersion lithography by early this year. A couple days before ASML's announcement of its prototype tool, Nikon announced that it would indeed plunge ahead into immersion tool development. The announcement detailed a schedule that shows completion of an engineering evaluation model (with a 0.85 NA) by the second half of this year, a pre-production version (0.92 NA) some time next year, and production tool (>1.0 NA) in 2006. As of late December, Canon had not made its decision official. In November, however, Canon USA announced that immersion work was looking good, and it is indeed a potential successor to dry 193 nm technology.

If the industry decides to push ahead with 193 nm immersion lithography — which there is little question of — it could possibly obliterate 157 nm lithography from the roadmap. But nobody seems willing to set all that work aside just yet. When TSMC takes delivery of the XT:1250i, ASML will have an installed base of both immersion systems and 157 nm dry systems, Sluijk noted.

"Both tools still need to be proven in volume production," Sluijk said. "We think immersion will be easier than 157, but it's still too early to know for sure." Although ASML expects to release its last dry 193 nm system late in 2004, it has a dry 157 nm system slated for delivery in the fourth quarter (Figure ). "By mid-2004, we'll decide whether immersion lithography will deliver on its promises or not," Sluijk said. "We'll decide whether to continue with immersion at an NA of >1, or pursue 157 dry instead." Assuming 193 immersion delivers, the industry will ultimately decide whether to bring 157 immersion onto the roadmap as well.

ASML expects to make a decision by mid-2004 as to whether it will continue to pursue immersion lithography at NA>1, or 157 nm lithography instead. (Source: ASML)

Rumors have also been swirling about the demise of EUV lithography, or at least its delay into the next decade. But again, it's too early to know. EUV lithography is now slated for the 32 nm node, and it's possible that immersion could extend to that point or beyond, especially if it is applied to 157 nm lithography. With immersion lithography unproven in production, however, these questions remain unanswered.

International SEMATECH (Austin, Texas) will sponsor its third Immersion Workshop Jan. 26-27 in Los Angeles to review the progress that's been made thus far, and to further evaluate immersion's feasibility.

For additional information on lithography, go to www.semiconductor.net/lithography