Greetings from our Nation’s Capital

I am writing late at night from Washington DC, where I am not following the advice of my lawyer to retire early. Tomorrow I am being deposed as an expert witness in a lawsuit – an experience, so I am told, unlike any other. But I’m not interested in talking about that, nor politics, nor the various memorials to famous dead people that I visited today. Instead, I want to talk about a bar.

I have just returned from the Brickskeller, a fairly famous local joint that is celebrating its 50th anniversary this year (it’s been a fixture of Washington life longer than most, but not all, of the politicians that tend to hang out it much more upscale bars than this one). What’s so special about this bar? It’s a cool little dive in an interesting part of town, not too far from the action but far enough to be a true neighborhood bar. The bartenders are friendly, carrying on conversations and knowing the city in Wisconsin where the beer you are drinking came from.

Oh, and they have a thousand beers to choose from.

Not on the order of a thousand, but actually one thousand beers. The menu goes on for pages. The number of beers on tap, of course, is much more limited, but I was still able to enjoy draft beers from Lyons (Colorado, not France), Wisconsin, Vermont, Oregon, Canada, the UK, and Russia.

In case you are wondering, the Brickskeller does not own the world’s record for beer selection at a bar. That honor goes to the Delirium Café in Belgium, which is reported to have 2,500 beers available (and so has earned the pink elephant that is their mascot). The Brickskeller, however, holds the record in the US, and the record in terms of bars I have been to. It almost makes me want to visit DC again soon.

Almost.

PMJ – Day 3

The final day at PMJ had the papers most interesting to me (that is, more related to the use of masks rather than their fabrication). Here’s a few things I learned at the conference (if you’re not a serious litho techno-nerd, skip this bit):

• Mask makers have started to ask for resists that are less sensitive (shot noise has become a problem), after three decades of complaining that resists are not sensitive enough.
• The non-ideal behavior of a scanner has a noticeable impact on proximity effects, and thus on the behavior of OPC tools.
• EUV is hard to do.
• No one really understands line edge roughness.
• Double patterning sounds good to mask makers – they get to sell twice as many masks, but mask errors probably won’t be the number one problem facing the lithographer.
• EUV is really hard to do.
• At the 45 nm node, lithography will be the #2 problem facing chip makers (strain variability will be #1) – how disappointing!

Overall, the average paper quality at PMJ is about the same as at other SPIE-sponsored lithography conferences, but with fewer really bad papers and fewer really good ones. There was one presenter who seemed to randomly change slides, both forward and backward, without any relationship to what he was saying. A couple of papers had none-too-subtle marketing messages sprinkled between graphs of data and self-serving conclusions. And of course, there was the guy whose company wouldn’t let him present any data, so he removed the numbers from the axes of every graph (why did he even bother to present?). But I’ve come to expect this from a conference that caters mostly to industry, and the number and egregiousness of the violations of paper propriety were relatively small. Most papers described solid but small incremental advances – the kind of thing that has been pushing Moore’s law forward for 40 years.

PMJ – Day 2

The second day of the conference has been eventful. One speaker went horizontal – fainted dead away – in the middle his talk. Apparently he is fine, but it was hard to tell as the paramedics took him away in a stretcher. It could have been nerves coupled with jetlag, or some bad tuna, or too much sake. I think, though, that he was done in by trying to make EUV masks work.

At lunch I was invited by some of the conference organizers to provide English lessons. As background, the conference, in Japan and run by Japanese, is carried on entirely in English. Now I am a typical American – unable to communicate in any language other than English, and just barely so in that language – so I have immense respect for anyone that is brave enough to commit to another language for an important endeavor. The people running the conference take their responsibilities very seriously, and asked for help in learning appropriate English phrases needed for conducting the business of conferencing. Is it better to say “In the interest of keeping on schedule, please see the author during the break if you have questions”, or “I’m tired of listening to this guy talk – let’s move on to the next paper”? I did my best to encourage authentic and appropriate phrases.

The day ended with a panel discussion from 6 – 8pm. The astute student of time might note the conflict between the timing of a such a technically fulfilling event and the emptiness of my stomach. Fortunately, the conference organizers anticipated this conflict and provided chips and beer. Less than fortunately, the beer was gone within nanoseconds of its availability – only the fast movers in this business reap the rewards of success. What did I learn from the panel? That panels always go better with beer.

PMJ Day 1

Photomask Japan (PMJ) is a relatively small conference – with about 500 attendees it is just under half the size of BACUS, the US conference on semiconductor mask making held each fall in Monterey. But small is much of what I like about PMJ. I can go to most of the papers, navigate the poster session, and attend the one panel discussion without feeling overwhelmed. The papers are generally good, with the normal mix of “quite good” with “quite abysmal”. Since the conference organizers take themselves very seriously (they are Japanese, after all), the marketing bottom dwellers of our industry rarely make appearances here to give sales talks thinly disguised as technical papers.

Day 1 was uneventful. I have to admit that the technology and science of making photomasks, the topic of the first day of talks, is not particularly interesting to me. By all rights it should be fascinating, since it is enormously challenging and immensely important. But for some reason, it does nothing for me. I’m glad there are people that revel in this topic, and I am glad that they are here this week. But for me, I am like the pre-med student sitting through calculus – I know it is important, but I cant wait till it is over. Now, the use of photomasks in a lithography process – that is what makes me wake up at 4am to read the book of abstracts. To each his own, I guess. That topic begins tomorrow – I can’t wait.

Tuesday night also included the conference banquet. This year’s banquet (standing room only – there were no chairs) was both novel and typical. It was typical in that the doors opened exactly on time, followed by a mad rush towards the buffet line. Don’t stand between a hungry lithographer and free food. And typical because it ended precisely 90 minutes later with all the food gone, the beer gone thirty minutes earlier, and no one lingering to chat. And it was novel because the entertainment was, well, entertaining.

Let me give you some background. BACUS (remember, the US version of this conference held each fall) is renowned for providing phenomenal entertainment programs, with maskmakers singing and dancing and putting on skits full of insider jokes and nerd humor (click here for pictures of past BACUS shows). But while the BACUS entertainment has imploded into misery over the last few years, the PMJ entertainment has gotten better and better. A rock band called “Progressive Masks”, made up completely of industry folks, did a great job. The song “Haze on the Photomask”, sung to the tune of “Smoke on the Water”, had us on the floor rolling with laughter. And the Hula Girls (and better yet, the Hula Guys) were great. Way to go PMJailians!

An American Lithographer in Japan

Its 4am, so I must be in Japan.

And it’s April, so this must be Yokohama, and the annual Photomask Japan conference. Photomask technology is an arcane but important subfield of optical lithography, itself an arcane and important subfield of semiconductor manufacturing, which is the slightly less arcane and possibly still important business that enables powerful chips to be made so cheaply that nearly all of my kitchen appliances have microprocessors running Linux. I can’t run Linux, but my toaster can. I’m sure my microwave regularly communicates with my dishwasher about the existential importance of semiconductors, and by extension, of photomask technology. This week, I’ll be trying to catch up with what my dishwasher probably already knows.

I am experiencing jetlag-induced early morning productivity – only because there is nothing else to do at 4am but be productive. I arrived in Tokyo late yesterday afternoon. The trip began routinely enough, with all the flights from Austin to Dallas canceled due to a ripple effect of bad weather in a part of the country that I never have been. Luckily, I convinced American Airlines to let me trade in my so-close-and-yet-so-far business class seat from Dallas to Tokyo for a middle coach seat from Houston to Tokyo on Continental. Lucky me. Eighteen hours later I was tired, stiff, smelly and at Narita Airport. After a relatively short 30 minute line through customs, a 90 minute train ride (I love taking the train in Japan – an exceedingly civilized way to travel), a 10 minute walk through a massively crowed train station, and 15 minutes in a cab, I was here.

I have to admit that Yokohama is my least favorite Japanese city. Not that there is anything wrong with Yokohama, it’s just that it isn’t very Japanese. Thanks mostly to it’s long history as an international port, Yokohama is the most westernized of Japan’s cities. Sure, you can find malls and Starbucks and McDonalds and Mr. Donuts everywhere in Japan, but in Yokohama (at least near the convention area) it’s hard to find anything that isn’t generically unappealing. I may be weird, but I enjoy trying to figure out, with my limited vocabulary, how to make sure I get a hot coffee instead of a cold Pocari Sweat (it’s a sports drink) or if I really heard right and my beer costs $12. But here in Yokohama, everyone seems to speak English – where’s the challenge in that?

But not to worry – this intrepid travel has been here enough to know what to do. A short 3 mile walk away is the entertainment district of Yokohama, where authentic yakitori bars and sushi bars and sake bars and bars I that definitely don’t want to go into (though I sometimes have) can be found in abundance. I know my way around pretty well, so I can find my way there and back even if I am drunk and it’s 4am, not that I’ve ever done that.

So if it’s 4am and you’re in Yokohama, come by and see me. If I’m not wondering the streets, I’ll be in the lobby of my hotel, searching for a wireless signal.

SPIE AL07 – One week later

It’s been over a week since the SPIE 2007 Advanced Lithography symposium has ended. Now that the dust has settled, what is my take on the conference and the state of lithography? I started out being pessimistic about 32nm half-pitch options, but now I am less so. Development of high-index materials for immersion lithography is going better than I expected (I’m always pessimistic about material development projects – they tend to be much harder than most people think), and double patterning is starting to show some interesting results. Coupled with significant improvements in scanner overlay performance, I think double patterning has a real chance of being a viable alternative at or below 35 nm half-pitch, maybe down to 22nm half-pitch. And my opinion as to the ultimate limiter of the capabilities of optical lithography remains the same – line edge roughness.

All in all, this year’s conference reinforced my earlier prediction that 2007 will be a very interesting year in the world of semiconductor lithography. Of course, interesting is not always good (like the Chinese curse), but I’ll take it over boring any day.

SPIE AL07 – Day 5 (Friday)

The symposium is finally at its end, with a half-day of “tool” papers on Friday. Unlike the rest of the week, the Friday morning papers are, for all practical purposes, reserved spots evenly divided among the major optical projection tool manufacturers, with a few subsystem providers thrown in for good measure. The quality of papers is usually mixed, and this year was no exception. Technical content tends to be low, and that is not necessarily bad, since my brain became full sometime on Thursday. Still, marketing content is too high, too many graphs are shown without labels on the axes, and the point of any given paper on Friday is not always clear. But it is useful to hear about tool roadmaps and get a sense for how hard the various suppliers are striving for technical leadership.

After a week of high energy, the 2007 SPIE Advanced Lithography Symposium didn’t come to a climatic finish, but instead faded away as the 4,000+ total attendees dwindled to a few hundred by the last talk. Granted, 200 people would be a great turnout for the vast majority of SPIE-sponsored meetings, so this shouldn’t seem like a complaint at all. But each year I walk away from the last paper with just a little sadness, knowing it will be another year before the excitement returns.

SPIE AL07 – Day 4 (Thursday)

It’s late on Thursday night, and I am exhausted. Only one half-day of conference to go. I gave my sole paper on Wednesday, and I am enjoying the idea of relaxing back at home rather than rushing to yet another paper.

But it was not always like this. 1993, in particular, was not like this.

I had the first paper slot on Friday morning of that year, for a paper entitled “Phase Contrast Lithography.” As usual, I was trying to finish it at the last minute, but things were just not coming together. My simulation results didn’t show many interesting effects, and while it is always worthwhile to give a paper that says “I looked at this idea, and it didn’t work out,” still, my intuition told me there should be something there. I just wasn’t finding it. Out of desperation my mind began to wander beyond the 90 degree phase-shifter/pupil filter idea that I was working on and I started scribbling down some equations. Within a few minutes I had derived an elegant result predicting a non-intuitive behavior of a 60 degree phase-shift mask and pupil filter combination.

I was ecstatic! I cranked up PROLITH to run some simulations and I began to write a whole new paper from scratch. It was midnight on Thursday night.

By 6 am I was running to Kinko’s with a floppy disk, desperately hoping it would take less than 2 hours for my powerpoint slides to print out as overhead transparencies. The acetate was still warm when I walked into the conference and started giving my talk. I’m still very proud of that paper.

That was 14 years ago, and such adrenaline-fueled excitement is now the pleasure of a younger generation of lithographers than myself. This year, I turned in my paper before I got on the plane for San Jose.

As a postscript, a few months after I gave that paper in 1993 I was reading through Born and Wolf’s classic textbook on optics when I ran across the very same equation I had derived with such late-night excitement. Ah well. There is very little new under the sun, and discovery is often actually rediscovery. But it’s still fun.

Spouses, Babies, and SPIE

I ran into one of the conference attendees at my hotel, pushing a stroller. Of course, I had to stop and admire his beautiful nine month old baby boy. My friend introduced me to his wife, explaining that I was the one he had told her about, that this time last year I had brought my new baby to the conference. She just glared at me, and immediately I understood the problem.

“You’re mad at your husband, aren’t you,” I said.
“Yes!”
“He’s been ignoring you all week, hasn’t he?”
“Yes!” Her look of disdain was directed at both her husband, sheepishly looking at his feet, and me.

I guess I was partly too blame for a week of unfulfilled expectations. My wife and daughter didn’t come to the conference this year. She told me she would never come with me again, since I completely ignored her all week and she was bored out of her mind.

I gave a sympathetic glance to my friend, knowing that there was nothing more I could do. This was a lesson most often learned the hard way.

SPIE AL07 – Day 3 (Wednesday)

So, what’s the minimum pitch found on a 45 nm half-pitch process?

No, it’s not a trick question. And yet, surprisingly, the answer is not necessarily obvious. I’ve talked to a few people and it seems that the phrase “XX nm half-pitch” is becoming a name, not a numerical description. We’ve known for years that node names, once themselves descriptive of the minimum half-pitch of the chip, are now the sole purview of marketing departments and can mean just about anything. That’s why people began to specify minimum half-pitch. “Here are the results from our 32 nm node process, which has a 45 nm half-pitch …” But if a company quotes the half-pitch to describe a process, what happens if that company decides to relax the minimum pitch due to lithography challenges? The easy solution is to relax the pitch without changing the name you’ve given to that process. We are starting to see “45 nm half-pitch” processes with a metal 1 (minimum) pitch of 100 or 105 nm. We in this business are nothing if not innovative.

This is a short blog today. Too many hospitality suites last night.

Musings of a Gentleman Scientist