June 21, 2007 – A new study by Wright Williams & Kelly Inc. and Strategic Marketing Associates sheds light on what people in the industry think will be most likely to hit production process lines in the next 2-5 years, and which ones may take longer than expected to be adopted into manufacturing lines, if ever.
Manufacturing technologies included in the survey are most lithography and gate technologies currently in development — new methods of cleaning, deposition, and test/metrology; and several cost reduction or productivity enhancing manufacturing technologies, including 450mm, energy saving methods, remote diagnostics, and manufacturing simulation, according to Daren Dance, WWK’s VP of technology, who shared some insight into the survey results in an e-mail Q&A with WaferNEWS.
Getting the nod by ≥50% of respondents to be seen in production manufacturing by 2008-2010 are high-k dielectrics and metal gates; adaptive test; and equipment with energy saving “sleep” states, according to the survey. Meanwhile, ALD is viewed as ready now for production by 50% of respondents, and half of respondents think both FUSI and integrated in-situ metrology will be ready within the next year or so.
Another ≥50% of respondents think a group of technologies, including damascene gate formation, 193nm high-index immersion lithography, and wafer-level reliability testing, should be ready for production sometime between 2010-2012.
Dance pointed out that it’s no surprise that respondents are cautious about lithography forecasts, since “new litho technologies have a way of falling by the wayside, leaving optical as the industry workhorse.” He added that although similar caution exists regarding gate technologies, multiple advanced gate approaches are likely to be used, as each has different benefits for certain architectures and applications.
One of the more interesting results from the survey was respondents’ opinions on a change to 450mm wafers, one of the more polarizing debates currently happening in our industry. Right now, sentiments seem to be of three types:
– “Yes it’ll happen” — pushed heavily by the biggest chipmakers like Intel, or TSMC seeking economies-of-scale (and with them probably heavily funding much of the development costs);
– “Yes, but way later than the ITRS‘ 2010 deadline” — noting how long it took 300mm to gain traction, and how long it’ll likely take everyone involved to come to a consensus the next time;
– “No, it’ll probably never happen” — citing tool suppliers’ aversion to bear any brunt of the massive development costs foreseen in a new wafer-size change implementation.
In this WWK survey, respondents’ sentiments fall squarely in the latter camps. The most frequent response to 450mm wafers was that they won’t see the light of day in production manufacturing until 2013 or later — but 39% of respondents think 450mm won’t ever happen at all. Respondents also doubt the viability of EUV lithography and imprint lithography as finding inroads into production until 2013 or beyond.
Dance noted that the survey tried to avoid bias by asking for a “date when expected,” with the option to choose “Never” added at the last minute. That option was ultimately checked at least once by ~40% of the questions, though he pointed out that some of them referred to technologies that are actually seen in current production. J.M.