|
SPIE: The trouble with immersion? |
|
|
|
Feb 26, 2007 at 09:56 AM |
The Advanced Lithography event of the year may yet have to officially start here in rainy San Jose, but it seems clear from attending Nikon's Sunday symposium that decisions over what litho techniques will be used at the 32nm node are far from certain.
The LithoVision's event was once again a good event to have been privileged enough to have been invited to, and it didn't disappoint!
Nikon's Geoff Wild stated that the company expects to ship 30 immersion tools in the current financial year, which is significantly above previous projections and, perhaps more importantly, well above last quarter's figures that barely indicated that the average quarterly shipments would be no more than five units.
Nikon expressed confidence in one particular customer starting volume production at the 45nm node in 2007, while inviting Toshiba to present on immersion developments that will see the NAND Flash memory manufacturer migrate this year to its NSR-610C, 1.30NA immersion tool, for sub-60nm volume production requirements.
Nikon also revealed that high-index fluid and high-index lens materials had not become bin-technology, though much further work was required.
Viscosity issues that had been highlighted before at SPIE were brought up again and have proven to be a sticky point (pun intended), but have not stopped development work in this area.
Ralph Dammel of AZ Electronic Materials presented data on how poor the viscosity levels were for high-index fluids, though the fact that such materials could give the required higher NA (up to 1.67) was encouraging. The downside was that scanner speeds would need to be reduced to 100mm per second, a far cry from the push to go to 500mm per second speeds in current immersion (H20) production requirements.
Nikon also had some good news regarding EUV Lithography with regard to their first tool shipping to an unidentified customer in the first half of this year. From the pictures shown, the tool looks more compact than the two Alpha tools already shipped by ASML, and Nikon was very confident in the Xe DPP light source performance of 10w IF.
The aim is to have a tool (EUVL2) capable of 50w IF and 50wph when required for volume production at the 22nm node.
A side note on this was that it would seem the lithography tool suppliers are now suggesting that wafer throughput is proportional to the level of power achievable. That level seems far less than what was claimed to have been needed by the likes of Intel!
|
Blogs 
