Jay,
There are many different combinations of technology to achieve 0.18, 0.15, and 0.13 resolutions using a 248nm light source. A 0.7 NA scanner like ASML's latest model can resolve 0.15 features using binary masks and OPC. High-volume DRAM companies will use PSMs when it is cost effective, like now. When a shortage for DRAM becomes apparent, those same companies will purchase new tools to both expand capacity and acquire tools that are capable of finer resolutions obviating the need for very expensive PSMs. I have no idea as to how this will play itself out, but sooner or later DRAM companies will purchase new high-end tools. It is my understanding that Micron already has tools in place at their two Boise fabs for 0.18 production using PSMs, but will need to upgrade the acquired TI facilities with new DUV tools. According to articles, Samsung has 25% of its capacity at 0.25 and under. Samsung is considered a technology leader, so others are most likely lacking compared to them, excluding Micron. The Hyundai/LG combo is a big question mark. LG wants over $4B for their semiconductor operations from Hyundai, but the D/E ratio for the combined company will still be in the 700% range. If a DRAM shortage materializes, they will most likely find the financing, if they haven't already.
I agree with the comments from this industry expert.
Feature size shrinks, on the other hand, are proceeding rapidly.
Quarter-micron and smaller features will account for 90% of equipment
investments by 2003, Rinnen said. By the same year, 76% of installed
steppers will use 248 nm exposure.
The time from introduction of an exposure technology to 10% market
penetration is also getting longer. Projection aligners reached the 10% mark
only two years after production systems were introduced, versus 10 years for
248 nm systems. In light of this trend, Rinnen suggested that forecasts calling
for 193 nm lithography to become mainstream by 2002, only 4 years after the
first production-ready system, may be overly aggressive.
While all links in the lithography food chain are experiencing above-market
growth rates, mask growth is much stronger than stepper growth. In many
cases, advanced masks allow lithographers to extend the lives of their existing
steppers. Rinnen expects shrinking feature sizes to drive expanded use of
optical proximity correction (OPC) and phase shift masks (PSM), with
resulting strong growth in mask revenue. Most predictions place 193 nm
lithography's insertion point at the 0.13 µm generation or later, well below the
exposure wavelength, so even early 193 nm masks will require OPC.
Because of their high cost, PSM volume is likely to be counter-cyclical. When
markets are healthy, companies will likely buy new equipment instead of using
PSMs. During downturns like the current one, high volume chip companies
will use PSMs to delay capital spending.
news.semiconductoronline.com
Bob |
