Tulvio,
1. About the throughput: in your example, it will still be 80 wafers per hour, because the first machine does not stay idle while the wafers are processed by the second machine. Trust me on this one :)
2. About mix-and match yield per wafer: this is more difficult to answer. Clearly, mix and match can't produce smaller chip than .25nm alone, but depending on particular IC, it may still produce the same size chip (non-critical layers done with .35nm).
As I'd stated in my previous note, my gut feeling is that mix and match is not optimal for the relatively simple chips, like DRAMs - it will be inferior to one of two "pure" alternatives (.35nm i-line or on .25nm laser). Mix and match is aimed at more complex chips (e.g., processor+memory on a chip).
As for the choice between Nikon i-line and laser products by DRAM producers (and that's the market Nikon is targeting), I can only list some factors that could influence the decision:
1)Difference in price (should be low enough - laser costs less than 10% of the stepper cost)
2) Difference in service/parts cost (unfortunately, for now may be considerable, the only article I saw said 25ok vs.25k)
3) Immediate availability and time until delivery - BIG question
4) strategic considerations (ability to go to lower line widths) - these should favor laser
5)- and this is what is killing the whole sector now- ability to get financing
Regards,
Y. |
