Debra Vogler, senior technical editor
March 3, 2011 — At the SPIE Advanced Lithography symposium this week in San Jose, Kurt Ronse, director of lithography at imec, discussed the research center’s new ASML pre-production EUV scanner, EUV readiness with source power (still a concern) and resists (practically there). He also discusses the perhaps overlooked topic of pattern collapse.
imec is currently installing ASML’s preproduction EUV scanner, the NXT3100, over the next few weeks. Kurt Ronse, director of lithography at imec, Ronse says it should be completed by late May to be up and running in June. (Here’s a Youtube video showing some of the tool’s move-in and installation.)
Regarding concerns about EUV readiness (described by, among others, TSMC’s Morris Chang), "source power is still a major concern," Ronse admits, but points to what he says was a major achievement just a couple of weeks ago: a ~20x; increase in power level at ASML, measured for several hours with "reasonable reliability." Another ~10x; factor increase will be needed, though, for high-volume manufacturing (HVM), he adds.
Resists, meanwhile, are "almost where they have to be," he said (~10-15mJ/cm2). The major issue really is getting line-edge roughness into spec, but there’s "steady progress" in this area, and also a lot of work on resist smoothing, including LER after etch. And a less-talked-about but still important consideration: pattern collapse, which becomes problematic below 20nm and will need to be solved to scale up EUV, he says.
Regarding comparisons to multibeam lithography, Ronse notes that EUV has "gained quite a lot of momentum" in the past 1-2 years and is now seen as the mainstream choice for HVM. But, e-beam does can play a major role in niche markets such as mask writing, he says.