By Hank Hogan
AUSTIN, Texas—Following International Sematech's recent endorsement, the semiconductor industry may embark on a wet—but insiders hope not wild—ride down the road toward immersion lithography.
In immersion lithography, the space between the bottom of the lens of the tool and the wafer is filled with a liquid, creating another optical element and contamination source, says Walt Trybula, strategic planning manager for lithography at International Sematech.
The industry, he says, is aware of this and the most popular immersion lithography approach, dubbed “splash and flash,” takes this issue into account.
“Some people have described it as pushing a bubble of water all over the wafer,” Trybula says. “It's a different bubble. It's constantly being replenished because you worry about contamination.”
Various leading lithographic toolmakers, Canon, Nikon and ASML, have said they will make a decision on immersion lithography products early in 2004. If the decision is a go, extensive testing and possible production will follow.
The reason for this interest is that staying dry may be too risky. Currently, the most advanced manufacturing is based on 193-nanometer (nm) lithography. An industry rule of thumb is that features approximately one-third the wavelength can be comfortably imaged. So, 193-nm lithography yields minimum features of about 65 nm.
The next step involves 157-nm lithography, which means new tools, masks and photoresists. But this infrastructure may not be ready in time, a point driven home by Intel Corp. (Santa Clara, Calif.). “We made an announcement on our plans for the 45-nm node, which are to extend 193-nm,” says Janice Golda, Intel's assistant director of lithography capital equipment development.
Enter immersion lithography, the development of which is being followed by Intel and other Sematech-member companies. In theory, the index of refraction of the liquid effectively divides into the wavelength of the source to create a new imaging wavelength. The index of refraction of water is about 1.45, and that shrinks the minimum feature size for 193-nm tools to 45 nm or less.
Actual implementation is a bit more complicated than the theory. There are the edges of the wafer to consider and the optical properties of the water to contend with. Investigations have shown that doubly de-ionized, degassed water meets the tight specifications needed.
In the “splash and flash” approach, the tool steps over the wafer, injecting water into each field, exposing it to light, removing the water and then moving to the next field. Using this technique, Nikon has successfully imaged 45-nm features. There are, of course, problems still to be overcome. Contamination and material movement between the water resist and the wafer could be an issue.
Taku Hirayama, project leader of immersion lithography materials at resist maker Tokyo Ohka Kogyo (TOK), notes that for a resist company, the challenge is to develop materials “… which can minimize the dissolution of resist components into water and decrease the water uptake of the resist film surface.”
The solution may be a barrier layer, one of which TOK has already demonstrated. But there could be other approaches. To date, no major problems with immersion lithography have emerged, and deployment decision about the technology is expected early next year.