By demonstrating that less expensive lasers using long, rather than the usual ultra-short pulses, the researchers hope to pave the way to EUV lithography beyond the 32-nanometer node.

"We have found that instead of 20-nanosecond long pulses, that we get better results with pulses over 100 nanoseconds," said Mark Tillack, a research scientist at UCSD. "We applied for a patent, because these longer pulse are cheaper and easier to make than short ones."

Some researchers are turning to spark-like ignition sources to produce EUV flashes, but Tillack and Cymer believe that their new approach will enable relatively inexpensive carbon dioxide (CO₂) lasers to provide the energy that produces EUV light.

At 32-nm, many leading-edge chip makers will insert 193-nm immersion tools and deploy double-patterning techniques.

At 22-nm and beyond, there are several lithography candidates vying for dominance in future production fabs. EUV, double-patterning, nano-imprint, multi-beam are among the choices. But for the most part, it appears that leading-edge chip makers are banking on EUV.

Recently, however, Intel Corp. reportedly said EUV would not ready for the 22-nm node. The problem: The sources, resists and photomask are not ready. Cost is an issue: Some say an EUV could run $50 million or more.

For example, an EUV tool will require a 100 watt power source for mass IC production. At present, EUV power sources are only in the "single-digit range."

ASML Holding NV and Nikon Corp. are separately supposed ship ''pre-production'' EUV tools by the end of 2009. The early tools will use Cymer's laser produced plasma (LPP) technology as the source.

Cymer isn't the only source vendor in the market. Japan's GigaPhoton also has an active LPP development program, according to sources.