This year, in the two-year chip innovation cycle, the ITRS is only an update of last year’s edition, to be edited and co-developed by about 1000 industry volunteers and to be finished soon now, before it becomes the global planning tool for the high-tech industry. Yet, in the regular flow of things, the 2004 ITRS update has created an unusual ruckus. Why? It spells out, in the words of its chairman Paolo Gargini (also director of technology strategy at Intel), a path to and the timetable for nanotechnology in chip manufacturing. And, even more disturbing to equipment makers, it initiates the migration to a wafer diameter of 450mm, to be production-ready in 2012.
In other words: the headache of migrating to 300mm will return, but much bigger. In the “proposed 450mm-wafer transition timeline” (Gargini in his keynote speech at Semicon West 2004), next year will see the first wafer standards evolving as well as the determination of carrier and lot-sizes. In 2006, just when the next chip recession is probably going to roll around, the industry will have to deal with new and more expensive factory control systems, transport and equipment standards – all to continue through 2010. Then it’ll time for the first 450mm alpha tools to arrive and begin a three-year cycle of interoperability testing and reliability verification. It took seven years of definitions and pilot production for 300mm to almost arrive in mainstream semiconductor production, Gargini says. “For a 2012 intercept that means that the analysis should begin in 2005”. And that’s not all of it, folks. As one of Gargini’s statements revealed, the next monster wafer size – only a joke? – is waiting around the corner: 675mm-diameter wafers are supposed to arrive in 2019.
The other key element of the update is the timeline to nano devices, tools and materials. Until 2010, Gargini contends, “evolutionary CMOS” will continue: with downscaling as usual, but with new ingredients such as strained Si, high-k gate dielectrics, metal-gate electrodes, advanced MOSFETs (FDSOI) and enhanced v-sat devices with quasi-ballistic charge transport mechanisms. Then, around 2012, “revolutionary” CMOS will step up on the stage: with nanotubes and nanowires, optical interconnect and molecular self-assembly replacing mechanical pick-and-place. To be followed by “exotic” devices. Which, that much is known today, will be “really different.” Stay tuned for single-transistor switches and spintronics. Just what kind of exotic, really different manufacturing tools will all this require? In time, we will know a bit more on this predicted technology future.
Werner Schulz
EPP EUROPE 404
