157 nm Lithography Moves Forward

A year ago, the prospect of developing a thin, single-layer resist with adequate etch resistance looked doubtful, requiring considerable invention, according to Harbison. However, a couple of suppliers presented positive data identifying a feasible platform, the fundamental chemistry for 157 nm lithography. At this point there have been no demonstrations, so resist continues to be red-flagged, a potential show-stopper. "There has been significant progress in the development of single-layer photoresist systems, and I anticipate that with the current progress, by November, the next checkpoint, resist, will go from red' to yellow,' " said Motorola's Will Conley.

Emissions at 157 nm are absorbed by water, oxygen and hydrocarbons, presenting a challenge for system designs. Unlike 248 nm and 193 nm systems, 157 nm lithography tools require an argon or nitrogen ambient with complex filtration to maintain beam intensity. Once the likelihood of nitrogen purging to O₂ levels <3 ppm was doubtful, and an 8-in. CaF₂ cube required for catadioptic systems seemed improbable. While there are indications all suppliers plan to use a catadioptic design, combining reflective and refractive optics, there are various approaches. The use of a CaF₂ beamsplitter to effectively reduce beam path and system footprint is just one approach. To this end, SVGL (Wilton, Conn.) has shown CaF₂ ingots in the size range of 12 to 15 in. that demonstrated large single-crystal feasibility for their beamsplitter catadioptic design. Post-processing to refine the optical properties is underway. Though final material quality is still being finessed, the hard part has been demonstrated, according to Jim McClay, head of the 157 nm program at SVGL. The tool manufacturer also has shown that <1 ppm nitrogen purging is achievable and that optical trains can be fabricated without hydrocarbon degradation. Lambda Physik (Fort Lauderdale, Fla.) and Cymer (San Diego) presented data showing beam bandwidth 0.6 to 0.7 pm. Typically thought to be 1 pm FWHM, the narrower bandwidth will improve overall performance in catadioptic systems but is not narrow enough for all-refractive designs, according to McClay.