This is a key year for Nikon. The venerable Japanese company is entering the immersion lithography market in earnest with the first commercial shipments of its S610C immersion scanner, with a 1.3 NA lens capable of 43-nm lines and spaces.
“This is the real ramping-up year for water-based immersion,” said Toshikazu Umatate, general manager for development at Nikon’s Tokyo headquarters.
Nikon has to move fast to reverse the ascendancy of ASML, now twice as large as Nikon in terms of lithography revenues. The Dutch company has shipped 31 immersion scanners thus far, and controls more than three-fourths of the total 193-nm scanner market.
However, Nikon claims that it has a superior approach to immersion which does not use air flow to contain the water. Suichi Owa, a Nikon fellow, said the no-airflow approach eliminates the problem of watermarks from water evaporation under the air flow. A no-airflow approach also reduces cooling of the wafer and the resultant wafer shrinkage, which impacts critical dimension uniformity (CDU), Owa said.
To roll out the S610C, Nikon is leveraging its long-held relationships with Toshiba – which used Nikon’s first immersion scanner, the S609B, for 56-nm flash – as well as with track vendor Tokyo Electron Ltd.
TEL manager of immersion process development Jun Kitano said TEL and Nikon combined the S610C and TEL’s Lithius i+ track to study the defect rates of immersion clusters with topcoated resists and topcoat-less resists. Now that immersion systems are well-accepted, the next push is to eliminate the topcoats, which increase defect rates and complexity, Kitano said.
Defects already are a problem for immersion systems compared with dry scanners, despite protestations from IBM and others that dry and immersion defect rates are essentially equal. The chip vendors which can move first to a topcoat-less resist with low defects will gain yield and cost advantages at the 45-nm ITRS node.
Nikon got out-maneuvered in the early days of the immersion race, making early customers come to the Kumagaya development center to study immersion instead of sending out systems to research consortia and beta customers.
ASML managers also debated whether to ship kludgy, retrofitted dry scanners to customers two years ago for development purposes. They went ahead, taking dry systems, adding jerry-rigged immersion modules, and sending them out to IBM’s Albany research center, to IMEC, and elsewhere. Nikon, by contrast, had early customers come to Kumagaya.
Can Nikon turn things around? With double-patterning providing the next big challenge, Nikon has an opportunity to pull off a technologically superior DP system with reduced overlay. To succeed, Nikon needs to learn from its early immersion shyness and get its double patterning ideas out to the larger device community quickly. At this stage, it is all about communication.
However, ASML’s relationship with IMEC again may give the Dutch company an advantage. At the SPIE Advanced Lithography conference, IMEC research manager Maaike Op de Beek presented detailed double patterning studies, based on – you guessed it – ASML tools.