SEMATECH Litho Forum: Steady progress on all fronts

07/01/2008

With almost all major R&D lithography efforts reporting at SEMATECH’s Litho Forum (May 12-14, 2008, Bolton Landing, Lake George, NY), the common theme resonating throughout a debrief with forum co-chair Mike Lercel (and SEMATECH’s lithography director) and program chair Bernie Roman is one of great progress made on many fronts with a growing interest in maskless/e-beam lithography.

The event gauges the level of interest in the various lithography options by conducting both pre- and post-event surveys of end users, as well as suppliers. According to Roman, the results from both surveys among end users were fairly close in their assessment of technology directions. Unsurprisingly, end users gravitated to double-patterning for use by 2010, and most also selected double-patterning for 32nm (half-pitch) by 2013, although Roman noted that there was increased interest in using EUV by 2013 as well. By 2016, EUV was the “only game in town” for the most part.

There was, however, a noticeably increased interest on the part of end-users in maskless lithography as an option or substitute by 2016, which “was one of the surprises that came out of the survey,” Roman told SST.

Three companies involved in e-beam maskless litho, all with tools in either the design or prototype stages–Mapper Lithography, Multibeam Systems, and IMS Nanofabrication–presented at the forum. “IMS Nanofabrication showed imaging results from its “many beams” proof-of-principle test bench tool (PML2),” said Roman, adding that this technology has seen increased activity in the past two years.

Papers on imprint lithography were also featured at the Litho Forum, and Lercel pointed to DNP’s improvement in template quality as particularly impressive. “They showed template resolution down to 15nm hp,” Lercel told SST, “and large (close to full-field) templates at 32nm hp were done on variable-shaped beam tools.” The significance of this work is that it shows that one could “almost” do a 32nm hp template with a standard mask process–something Lercel characterized as a fairly dramatic statement. He added, however, that defects were still a problem due to a lack of inspection capability. Besides the difficulty of working with 1?? templates (versus 4??), the charging of glass templates that is encountered during the inspection process is problematic.

Lercel also noted that there was tremendous progress shown in high-index immersion fluids. SEMATECH and supplier research showed that Gen 2 fluids do exist that can enable an NA of ~1.55 to be achieved in scanners–and he hinted that SEMATECH’s work with metal nanoparticles may make it possible to go beyond this number. “Metal nano-particles have a very high index of refraction, and when dispersed in a fluid, they dramatically increase the index of refraction of that fluid,” he explained, though this is a promising approach that still needs a lot of work.

Also noteworthy was the work on the high-index immersion lens material LuAG (lutetium aluminum garnet) by Schott Lithotec, which reported “absorbance very close to the target they need,” as well as “significant progress on purifying the material as well as growing crystals up to 80mm in diameter,” explained Lercel. “They also showed that residual absorption is purity-based, so with further refinements, they should be able to make the LuAG material successful.”

Despite the impressive work on high-index fluids and materials, the proceedings took a different turn when Nikon announced at the forum that it had canceled its high-index program. And in general, Lercel noted that all three litho toolmakers (ASML, Canon, and Nikon) were concerned with issues relative to fluid containment and fluid lifetime. “There has to be some re-design of the fluid containment to get high scan speeds,” he said, pointing out that fluid lifetime was related to defectivity. –D.V.