<<Regarding Semitool, I don't know enough about either system to judge the relative merits, but you certainly can't ignore them. They certainly have substantial R&D experience, and may also have at least a beta production tool.>>
I think they have more than a beta tool to date, though I guess its still beta until true full blown production begins.
<<(It's hard to say,because the chipmakers are being very tight-lipped about what they're actually doing. Only IBM is claiming to be production-ready with copper, though. If there's only one customer, it's unlikely that both Novellus and Semitool are selling tools to it.)>>
I believe MOTO and AMD are as well. I believe MOTO's Hector Ruiz made that statement the week following IBM's announcement, and I've heard the same from AMD. I believe in both cases SMTL and NVLS are involved though its unclear how with the Damascus announcement. I beleive SMTL's ECD is deemed superior to NVLS (hence the buyout attempt last year) which may be why both are involved....or so I hope.
I guess the theory behind this, and please correct me if I'm wrong or this sounds illogical but as it was proposed to me is the that one supplier supplies the barrier layer tool, and the other supplies the seed layer tool.
<<Their big disadvantage vs. Novellus would be size and ability to provide the other parts of the equation.>>
Absolutely. Long term, if SMTL intends on making a business out of this they will need to team, partner or be bought out.
<<Novellus is also better positioned for alternatives to copper. For instance, a number of Japanese companies are doing serious R&D on aluminum damascene--the cost advantages of damascene, without the risky switch to copper, and DRAMs don't need copper performance yet
anyway. That's bad for Semitool, but still a potential market for Novellus.>>
Technical issues aside, I believe the domestic DRAM players (there are only two) have been looking at Cu. I'm not sure how that plays into what the Japanese are doing, but from a practical standpoint why spend all that research on developing an aluminum damascene process when most of the bugs have been worked out w/respect to Cu. Am I missing somthing here, or is this much like the Japanese insistance on employing SOG in place of CMP. The lowest cost mfg of DRAM in the world started using CMP way way back when few knew or could appreciate its advantages - lowering cost and improving yield. As Cu will improve yield b/c of aluminum's disadvantages, why buck the trend with respect to DRAM? Either way, though, I agree, SMTL will have a long term problem unless realize that they can't be a loner in this game. Your comments are sincerely appreciated.
Thanks in advance....SemiBull |
