by M. David Levenson, Senior Technical Editor
The 380 attendees at the Third International Symposium on Immersion Lithography earlier this month, found a field moving out of R&D and into a competitive commercial phase. Key technical presentations addressed hyper-NA, resists, exposure tools, optical materials, process, photomasks, immersion defects, and alternative immersion fluids. More than 20% of submitted papers dealt with immersion defects, revealing the industry’s momentum in preparing 193i for volume manufacturing at sub-65nm half-pitch.
Two vendors — Nikon and ASML — claimed to have demonstrated water immersion exposure tools with performance consistent with volume production. Both announced 7nm single machine overlay and 11nm overlay for mix-and-match applications with a dry lithography tool. Focus and leveling accuracy of both tools was stated as 26nm, in spite of the great difference in the technologies that the two vendors used for focussing. Defectivity was claimed to have been reduced to about 0.04/cm2, partly through the use of new on-tool cleaning systems that suppressed particles, even those carried onto the tool by the wafer.
Where the offerings of the two vendors differ is in the numerical aperture and hence resolution. ASML’s 1700i with an NA of 1.2 required polarized dipole illumination to produce 45nm half-pitch lines and spaces (in one orientation only). Nikon’s NSR-S309B, with maximum NA of 1.07, could not do that, but was targeted for 55nm resolution in Manhattan geometry with azimuthally polarized on-axis quad-pole illumination. The Nikon team demonstrated full-wafer CD uniformity of 3.8nm for V and H oriented lines at that half-pitch.
Both vendors are building more advanced tools that approach the NA limit of water immersion fluid at ~1.3, and thus should give 2D imaging at 45nm half-pitch under ideal conditions. Nikon predicted its first NSR-S310C tool would ship before the end of 2006. Canon, which did not present data on their advanced tool, nevertheless predicted first shipment in 1Q07.
The key trick to making real circuit patterns is the proper combination of resolution enhancement technologies (RETs) — such as azimuthally polarized quad illumination and OPC — on a highly capable exposure tool. Right now, no full-field tools are available with the capability to print chip-like patterns at 45nm hp. However, engineers at International SEMATECH-North reported using an Exitech 193nm microstepper with NA=1.3 and azimuthally polarized on-axis quad illumination to print 84nm full pitch patterns (36nm line CD) simultaneously in H and V orientations in a single field (see figure, above). Still, 1:1 duty factor gratings, corners, Tee’s and other circuit patterns remained dicey. Suitable OPC methods are being sought and tested.
Defectivity remains a controversial topic. While control of bubble and water-mark defects seems to have become adequate, defect types due to particles mobilized by the immersion fluid remain above dry lithography levels, though perhaps not by much. Nikon claimed to have suppressed all immersion-related defect types, while Bob Streefkerk of ASML reported three per wafer (out of a total of 23) at fast (550mm/sec) scan speeds. There seems to be some evidence that scan speed matters for defectivity, which may limit throughput in production. Details of topcoat materials also matter for the resist profile, process window, and defect level.
Beyond water immersion, the situation remains murky. Interest in extending the technology below 40 nm half-pitch was demonstrated by a number of submissions on high-index fluids comparable to those on water. Second-generation (higher index) fluids are being characterized, but the resolution improvement may not be sufficient to motivate insertion. Many participants appeared convinced, however, that 193i technology could reach and possibly exceed 1.55 NA.
The third-generation fluids required for sub-40nm patterning will require new optical materials, and development of high-index lens materials was the consensus top priority mentioned in Andrew Grenville’s closing presentation. Second priority, according to the SEMATECH Steering Committee, are the higher-index immersion fluids, followed by resists — which need to leach less into the immersion fluid and also have higher refractive index. Defectivity was fifth priority, after double patterning technology — a testament to the industry’s success and optimism for overcoming current challenges. In one year, attendees at the next Immersion Symposium will learn whether it has been justified. — M.D.L.