Economics 101: Sticking With 193

Every year around this time, we fill the pages of Semiconductor International with information about the latest developments on the lithography front. There's never a shortage of material to write about, to say the least, and this year is no exception.

But the playing field has changed a bit this year. For example, 157 nm lithography, which was still somewhat up in the air last year following Intel's announcement that it would drop it, is pretty well off the table now for most commercial tool developers, as well as those within the supporting infrastructure. Part of the reason for this is the significantly increasing confidence in 193 nm immersion lithography — a colossal achievement in itself, considering the short time it's been on the docket. But it's also because of the countless other things that have gone into stretching the limits of the current optical wavelength, such as improved resists, optics, illumination, and increasingly complex resolution enhancement techniques.

Even 157 nm immersion is a long shot, sources say. It would require developing all the same infrastructure that was already behind the 8 ball in 157 dry development, and — at the end of the day — it's just easier to keep stretching 193, coming up with high-index fluids to put into the immersion systems, or novel high-index materials, for example.

What it really all comes down to is economics. We've heard an awful lot of talk through the last horrendous downturn and the subsequent all-too-short upturn about semiconductors becoming a commodity item. Whether you believe all the commodity talk or not, it's clear that this industry is becoming ruled even more than ever by the economic considerations that drive commodity markets. And that translates into taking what you've got, and keeping it working.

In this issue's cover story, Senior Editor Laura Peters details the work that has been done to make 193 nm photoresists production-ready (see "Photoresists Meet the 193 nm Milestone "). This year will see several chipmakers ramp 90 nm technology into full-scale production, using 193 nm lithography for the critical layers, with it really falling into its stride at the 65 nm node. The ArF resists have faced considerable optimization efforts in several areas, including line edge roughness (LER), defectivity and mask error enhancement factor (MEEF).

According to a report from Senior Editor Alexander Braun, however, LER continues to be a worrisome hurdle (see "Line Edge Roughness Is Here to Stay "). As the 65 nm node approaches, LER features that previously could be ignored now increasingly challenge 193 nm lithography, as well as etch. As one source put it, LER has evolved from a nuisance to something requiring attention. Immersion lithography provides some relief from LER because 157 nm resists faced even more troubles in this area. But the fact that immersion will likely push optical lithography out a few more nodes will cause LER to stick around for the foreseeable future.

There's no doubt at this point that immersion lithography will make it to full-scale production, and likely quite soon. As noted in my article this month, both IBM and TSMC have already demonstrated immersion lithography's capabilities on critical layers of commercial chips (see "Chipmakers Immerse Themselves in 193 Wet "). They might even share some of their results later this month at the SPIE Microlithography conference in San Jose.

No matter what it takes, this industry will make 193 nm lithography last as long as humanly possible (and then some) because it makes economic sense to do so. The likely ArF development costs, whether wet or dry, will never reach the relatively insurmountable costs of extreme ultraviolet (EUV) lithography. So engineers and researchers will just keep coming up with more neat tricks, which — I can only imagine — have a lot more to do with hard work than with magic.