The Fine Print

For the SEMICON West Daily News, I reported on the Tuesday afternoon Device Scaling TechXPOT, which featured presentations and panel discussion on lithography options at 22 nm. Although I wanted to include comments from moderator Lars Liebmann, I had to cut back to make the story fit on one page (all the news that fits, right?). Granted, lithographers and their suppliers have a mighty difficult task ahead of them, but Liebmann had a particular knack for making each technology candidate look bad — and for upsetting several participants and attendees along the way.

This TechXPOT session offered up six speakers to present on six potential lithography routes: immersion lithography (including high-index and double patterning), EUV lithography, imprint lithography, increased layout regularity, and evolving design flows. IBM’s Liebmann offered them up — then tore them down, one by one.

When the presentations were over, before moving onto the panel discussion, Liebmann decided to poll audience members to get their take on where the industry stood with regard to 22 nm lithography readiness. Very few people raised their hands when asked if they thought the 22 nm node would continue on a two-year cycle. Turning to the panelists, Liebmann said, “We were not able to convince the audience that we have a technology that will keep us on a two-year roadmap.” The panelists snickered softly, but as I sat in the audience, I had to wonder if some of them might be feeling a little animosity toward Liebmann at that point because of his string of deriding presentation follow-ups that may or may not have influenced the audience’s final viewpoint.

A few examples:

• Before introducing AMD’s Bruno LaFontaine to speak about EUV lithography, Liebmann reminded the audience that the X-ray lithography program was far more mature than EUV is today before the industry finally called it quits. Also a reminder that EUV readiness is slipping.
• After La Fontaine’s presentation, Liebmann again reminded the audience that EUV is X-ray, then said, “It’s looking a little grim here.”
• Larry Pileggi, a professor at Carnegie Mellon University, advocated a move away from arbitrary and toward more regular design patterns, noting that you can get better gate density by putting down very regular poly gratings, then cutting away the line ends with a separate step. To cap this presentation, Liebmann commented, “We’ve gone through this exercise before with alternating phase shift. A simple two-step operation turned into a nasty mess. I’m glad we decided to abandon it before we had to actually get into manufacturing.”
• After the presentation from Cadence’s Rick Brashears about design flows and the need for them to evolve, Liebmann joked about the need to make more computers basically just to run the designs to make the chips to go in those computers.

AMD’s Bruno La Fontaine argued why EUV is on the road to high-volume manufacturing, despite derogatory comments from moderator Lars Liebmann before and after his presentation.

Liebmann probably thought he was being funny, and I admit that I let a chuckle slip more than once. But after a while it got old, and some folks understandably took it more personally than I did. Stefan Wurm, for example, who was sitting in the row in front of me, got up and left shortly after the EUV presentation, muttering something to Canon’s Phil Ware, who was sitting next to me, before exiting. I got to talking with Wurm, program manager, EUV strategy at Sematech, as well as Molecular Imprints’ Ben Eynon (one of the session’s presenters), during Sematech’s reception later in the week. Wurm was pretty annoyed, as it turns out, about all the disparaging remarks from Liebmann with regard to EUV, and wasn’t going to hang around to hear more. Eynon agreed that it was all a bit much, and I talked to others during the week who also felt that Liebmann’s jibes were over the top.

As many of you have undoubtedly figured out by now, I have my own doubts about EUV and its viability, although I really am trying to keep an open mind. But I do think that if somebody is asked to moderate an event like this, they might do the audience a better service by holding back on the insults of each presenter’s platform. Liebmann seemed to set out to make many of them look bad, then ask the audience to confirm that they looked bad. It was certainly a foregone conclusion.