|
SPIE Litho nuggets: Why doesn't the industry support nanoimprint? |
|
|
|
Feb 28, 2007 at 10:08 AM |
My coverage of the second day of the SPIE Advanced Lithography show began and ended with nanoimprint lithography, or NIL, a shorthand sometimes pronounced like the British sports expression for "zero," as in "Leeds defeated Chelsea two nil in one of the most astonishing upsets in FA Cup history." Other times, NIL is spelled out loud like a proper initialism.
I polished off my last cup of coffee listening to Wei Wu of Hewlett-Packard Labs discuss his team's groundbreaking work in optical negative index materials.
That's right, negative index materials, something not found in nature but deemed theoretically possible since the '60s. Wu and Co. have used NIL techniques in their research, specifically a UV-curable imprint with double-layer spin-on resist process. He believes NIL is a promising candidate for fabricating these unnaturally occuring materials as well as for other optical components.
That evening, I sat in on the NIST-sponsored panel on the needs and prospects for NIL metrology and materials. Near the end of the almost two-hour discussion, Grant Wilson, professor at the University of Texas and an eminence grise of the NIL community, mildly ranted about the lack of investment in and support of nanoimprint by consortia.
Wilson cited the prodigious amount of money being pumped into all manner of EUV litho research, by the likes of Sematech, IMEC, Selete, as well as their industrial partners. "Give me one person out of Albany Nanotech's EUV mask team to work on NIL templates," he pleaded. He said he doesn't understand why it's not happening, and why there is so much resistance to NIL investment, given that nanoimprint litho has been added to the International Technology Roadmap for Semiconductors.
The good professor believes there must be consortia-scale activity for CMOS-type NIL to succeed. People are looking for extensive demos in areas such as defectivity that the current array of research labs, small equipment and materials companies, and modest alliances finds difficult to provide. Skeptics want the data but are not willing to fund the kind of work that is needed to provide a demo replete with said data.
Wilson's comments resonate, and do raise questions about the wisdom of those chipmaking cognoscenti who have hitched their litho wagon to pushing immersion as far as possible (good idea) and beating the EUV horse to near-death (not-so-good idea). The progress shown in nanoimprint technology has been impressive the past few years. By many accounts, the challenges that must be addressed for NIL to have a shot at CMOS fab applications, though daunting, are of an engineering variety, not technical showstoppers.
In other words, if there was at least a modest fraction of the focus and investment that the industry has devoted to immersion and EUV, NIL could move from its long-shot or niche position to legitimate production contender. Given the eyeball-bulging figures associated with production-level EUVL systems--"hey, $70 mill will get ya an extreme ultraviolet stepper, buddy"---you would expect a clamor for the kind of cost-effective, sub-10-nm approach that NIL potentially offers.
There a few supporters---HP, Fujitsu, some heretics at AMD Dresden, and others---but no major company taking on the role of champion of NIL, similar to that played by IBM for the implementation of chemical mechanical polishing in the 1980s. It's time for someone to step forward, take the risk, and fight the optical/photons-forever powers that be. Shoudn't nanoimprint litho get the resources it needs to prove (or disprove) its volume-manufacturing capabilities before it's too late?
|
Blogs 
