Yo ex-, sorry for the delay responding, I was reading up on stuff related to ion implant and oxidation/diffusion (classes to teach this week).
Intel was mostly responsible for the 125 mm to 150 mm step, in the mid '80s, I believe. The jump to 200 mm was carried mostly by IBM, although Sematech had a supporting role. Still, IBM once published that they'd spent over $8 billion of their own $ to get the industry over that hurdle, and they thought they'd let some other sucker, er, industry partner suffer through the next wafer size expansion. Which is slated to cost $18 billion to work the kinks out of, according to some industry analysts. That's more than a small pile of change!
I don't know of any full-scale, 300 mm fabs actually running. Siemens/Motorola in Dresden (Siemec, is that the name?) was slated for that, but I believe they populated with 200 mm tools when all the equipment makers pushed their 300 mm release dates out. Intel has a 300 mm pilot line in Oregon, and I understand it's making Pentium III chips; enough they call that fab "production" but it's really still a beta-test facility. <grin>
There was one other small facility up, but I misremember where. Best information on this is available from SEMI, by simply purchasing their "Fabs on Disk" which comes out regularly (quarterly, I believe). All you need is about $800 or so in loose coins from the sofa. (My wife took all my loose sofa change and installed some tile flooring, hence I don't have a recent copy of Fabs on Disk. <sigh> )
Matching the top with "when the next generation wafer sizes start kicking in" is much more problematical; I tried, really I did. Here's the rub: I can't find solid data on MSI (millions of square inch) Si devoted to each size wafer over time, preferably reported monthly, since about 1985 or so. That would make it possible to "correlate" growth in MSI devoted to each tool size over time, and where that falls on the rise/fall cycle chart for our industry.
Anybody got raw data of that type? MSI Si consumed each month, by wafer diameter, over the last 15 years or so? Then I might could give a guess about whether or not it correlates to the financial cycle.
At a guess, the correlation is probably not very good, as the economic cycle may not match the implementation cycle. We've had more tops (and bottoms) in the financial end (5? 6? since 1985) than we have wafer size changeovers (not yet 3 complete), and that's a Real Problem to sort out....
Mitch |
