To: SemiBull who wrote (912 ) 2/13/2004 9:40:18 PM From: Proud_Infidel Respond to of 43402 EUV Litho interview: with Noreen Harned of ASML Silicon Strategies 02/13/2004, 10:50 AM ET Noreen Harned is vice president of marketing, technology and business development at lithography equipment supplier ASML. Silicon Strategies asked Harned for an update on the progress in extreme ultra violet lithography, a research topic which seemed to have been eclipsed by progress with the immersion extension to optical lithography Silicon Strategies: How does ASML divide up the work on EUVL; such as masks, optical systems, illumination sources? Harned: We have a development program that is expected to yield a full field scanner. We do divide it up like and there's also vacuum technology. That's very important because we want to take what we already have and, if possible, adapt it for operation under vacuum for EUVL. So that's things like stages and transport mechanisms. In the mask area the big change is from tranmission to reflection. It's still a four-times reduction mask, just like today's masks. Indeed the EUVL masks would look like a 193-nm mask but it's a multi-layer reflective mask. Work is carrying on but ASML has been very active in promoting mask standardization: in terms of blank dimensions, coatings and protection. Masks still have some things to be resolved but Schott in Europe and Hoya of Japan. ASML has been instrumental in developing EUVL mask protection. EUV does not transmit through just about anything so you cannot have a pellicle. We've developed a special frame that protects the mask simply by defining a zone around the mask. Silicon Strategies: And what about the optics? Harned: Well we started this work six years ago and everyone is trying to minimize the number of reflective surfaces. Nikon, Canon and ASML have come up with their own designs. In our case I should say ASML Optics and Zeiss. But the challenge is to make the mirrors to the required smoothness and cover the entire spatial and frequency domain. But all of us have come up with more or less the same result, six mirrors plus the reticle, which is a mirror, for the illuminating optics. Silicon Strategies: The more surfaces, the more wasted illumination?> Harned: The source is the key. But ASML has been spending to enable this infrastructure. We don't want to be a builder of sources although we have put development money into this. We collaborate with Cymer, with Lambda Physik, with Philips, Extreme, Powerlase. Everyone wants to get very high throughput by which we mean 80 wafers per hour, even 100 wafers per hour and that requires a ten-fold increase in power from where we are now. Silicon Strategies: And how much power is that - the goal? Harned: We need 115 watts at the intermediate focus. Silicon Strategies: So how are you going to get from 10 or 11 watts to 115 watts? Harned: There's two main technologies: discharge lamps but the electrodes tend to breakdown; and laser-produced plasma which should scale up but can it be done cost effectively? But we can certainly get to a process development tool at, say, 15 watts. Silicon Strategies: What are your major collaborations in EUVL, who with and what timetables? Harned: There are many. We are in MedeaPlus programs, European Union programs, Sematech, when EUVL LLC was going we were involved, we're working in Japan. In MedeaPlus we are in the lead in a set of programs around an EUV scanner. There are MedeaPlus programs on source, optics and mask standardization work. In International Sematech there is mask work. Silicon Strategies: And in Japan? Who are you working with? Harned: That's more informal. But we are sitting down with Canon, Nikon and Selete. Silicon Strategies: When do you expect to have an alpha tool?, beta tool and a production too for EUVL? Harned: For the alpha tool we are on-track for the latter part of 2005, the fourth quarter of 2005. That will be a process development tool to allow engineers to develop manufacturing processes and other parts of the infrastructure. The latest estimates are that high volume EUV production won't start until 2009. We are very much market driven so we'll launch a production machine some time between those two dates, starting with some pilot tools. Silicon Strategies: 2006 seems unlikely after launching a process development tool at the end of 2005. Harned: 2007 would be a possible start for those early tools.