LONDON – Eric Meurice, CEO of lithography equipment supplier ASML Holding NV (Veldhoven, The Netherlands), has provided background on the roll out of its next generation extreme ultra violet (EUV) lithography machines and the throughputs customers can expect.
EUV lithography, considered by most of the industry as necessary to allow continued miniaturization of circuits, has been in development for nearly a decade and is still plagued by low power light sources that prevent machines reaching benchmark wafer throughputs in excess of 100 wafers per hour.
Meurice, speaking on a conference call with financial analysts held to discuss the company's somewhat tepid second quarter financial results, provided more EUV information than previously as the company gets close to shipping its NXE:3300 commercial EUV machine. The roll out of the first 11 "process development" machines, capable of handling 300-mm diameter wafers is now expected to be a 2013 event.
While in the short-term ASML is not promising much by way of throughput, possibly as little as 30 wafers per hour by the end of this year Meurice said there is now a roadmap to a throughput of 70 wafers per hour in 2014 and 125 wafers per hour in 2016. As and when customers call for 450-mm capable machines that should provide ASML with about a 10 percent adder on the sales price, Meurice said.
ASML has 11 of the NXE:3300 machines on order but these are now likely to be delivered to customers in 2013, Meurice said. The final integration of the first NXE:3300 at ASML is expected in October or November he added. The NXE:3300 systems shipping in 2013 will only be used for developing manufacturing process technologies, Meurice said but will still result in 800 million euro (about $980 million) of revenue for ASML in 2013.
Yes EUV is difficult - no question. But we've been at this several years now. You would think that at some point the estimates of future progress would become more realistic. Maybe you would even err on the side of caution so you could actually meet a milestone, or heaven forbid even beat one. I like ASML, and I think they make some great products, but ASML has been much more successful at marketing EUV than actually delivering on the technology.
@any1- I hear you. At a certain point it gets to be like the little boy who cried wolf. But in defense of ASML and everyone involved, what they are attempting to do is really difficult.
Have the pushouts resulted in some tarnished credibility? Probably a little bit. But I don't believe ASML would continue to spend millions of dollars on development and offer new targets for production if they didn't believe they would get there eventually. And I don't think Intel would have opened up its wallet if it wasn't presented with convincing evidence.
On the 70 wafers per hour being good enough, you may be right about that. But from what I heard last year at BACUS, I think that at this point chip makers would be more than happy with that as a starting point. Note that Meurice is now saying 125 wafers per hour by 2016, and that will I am sure be workable.
All this said, I do agree that the jury is still out on EUV.
I'm amazed that ASML can continue to miss scheduled milestones in EUV litho and have any credibility left. I understand that they are not directly responsible for the source problems, the resist issues, the whole EUV mask infrastructure, etc. But forgive me if I don't have much faith in this four year throughput projection. I know ASML believes that 70 wafers per hour is good enough for adoption for high volume manufacturing, but I question that assertion as well.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.