Robert - Wow! Pretty thorough. But some comments:
1) Do you know what the duty cycle is for Cymer equipment? If Cymer equipment is, as I would expect, on 24hrs per day then DUV248's post about the 1 quarter lifetime of the bearings is mostly accurate since 141 days is about 4-1/2 months - only a little more than 1 quarter.
2) The introductory paragraph in the article about the Cymer 6000 is written in a way that I find somewhat worrisome. They are subtly denigrating Cost Of Operation. I don't mean to imply that they are ignoring it, but they are implying that it is mundane. It isn't! ("While the conventional wisdom towards reducing the laser's contribution to the cost of the lithographic process has been and will continue to be reducing the laser cost of operation, laser-based enhancements in system throughput offer another powerful advantage to semiconductor manufacturers.") This, BTW, is one of the biggest dangers when you have tech company full of scientists. They tend to focus on the 'cool' problems and ignore more important ones - I myself have managed in multiple situations like this and it is often very hard to get the techies to focus on the right problems (I am ultra sensitive to this sort of loss of focus precisely because in my own personal experience it is *the* most common place for a group of engineers and scientists to founder.)
3) In searching for a Cymer 5400, I found that Berkeley's training lab is using other companies' lasers in addition to Cymer's:
www-microlab.eecs.berkeley.edu:8080/text/labmanual.html
4) Although I dug up some of DUV's previous posts, I did not notice DUV's continual reference to 5400. Since I do not work in the industry I assumed that the 5400 was either a line of which I was unaware, or was a vendor relabelling (the latter of which often happens in other industries as a method of differentiation. The manufacturer uses one numbering system and the OEM another.)
5) Note that in DUV's posts from last year he accurately predicted that there would not be a lot of orders for DUV systems in the next year because of the general slow down and difficulty in getting customers used to DUV type problems. He made this prediction when most people thought DUV would be immune to the slow-down. He also specifically stated that Lambda et al were not serious competition at the time. This also turned out to be true.
I would love to be able to rule out DUV's posts, but they have more truth in them than I would like. I'll try to get more info about the 'Cymer 5400' and some general info on Cymer reliability. Thanks again for the large amount of research you obviously put into your post.
Clark
PS Note that even if you take everything DUV says in good faith, it doesn't necessarily mean that it is true. I would be surprised if Lambda, Komatsu et al weren't claiming better COO and MTBF, and I would also be surprised if the chip companies weren't giving them more credence than they deserve. People hate having a monopolistic vendor and in their passion tend to take the competition's claims too seriously. (Look at a lot of people's attitude toward Microsoft. They would rather use any system other than microsoft even if it costs them money.) This is a worry, but nowhere near as big a one as it would be if the Komatsu or Lambda system really had a better system. |
