High-Index Lens Lags Roadmap by Several Months
Aaron Hand, Executive Editor, Electronic Media -- Semiconductor International, 10/10/2007 5:21:00 PM
Although double patterning/double exposure has been capturing the limelight lately for 32 nm half-pitch lithography, another competing technology with a shot at keeping optical in the fab is high-index immersion lithography. Using single exposures with a numerical aperture (NA) of >1.5, high index will be realized through an immersion fluid with a high refractive index (n>1.6) and an optical material with n>1.8. Combining a high-index lens element with a third-generation immersion fluid (n>1.8) could extend 193 nm optical lithography toward the 22 nm half-pitch by enabling an NA of ~1.7.
But the technology was delivered a bit of a blow this week as Schott Lithotec, the key developer of the leading high-index lens candidate lutetium aluminum garnet (LuAG), announced that they were behind schedule on some key material attributes. Schott’s roadmap aims to provide a full-size LuAG lens blank to toolmakers in 2009. Although LuAG’s refractive index of 2.1 is well within the goal, there have been challenges with regard to its transmission properties and the purity of the raw material.
Speaking at the International Symposium on Immersion Lithography yesterday in Keystone, Colo., Schott’s Lutz Parthier presented several positive results with regard to the company’s progress in crystal quality and material purity, but he conceded that improvements in transmission were behind schedule. Absorption of the 193 nm wavelength by the LuAG material has been reduced from 0.20 to 0.11 cm-1, he said, but that still fell short of the current target of 0.05 cm-1. The ultimate transmission target, aiming for the first quarter of 2008, is 0.01 cm-1. But that will be a very difficult target to reach given that Parthier estimates they are about half a year behind schedule.
Meanwhile, Schott has made significant progress in other areas, Parthier detailed. Both stress birefringence and homogeneity improved much faster than expected, and are close to their final targets. Homogeneity was reported at the Advanced Lithography conference in February as 0.12 ppm RMS at an 80 mm diameter. As of September, Schott had reached 0.029 ppm RMS at a 45 mm diameter. “It is one of the highlights of our development activities of last year,” Parthier said.
Schott has been able to create a crucible quality that meets requirements and has no further considerable impact on the crystal quality, according to Parthier, who noted that the achieved purity of the crucible reduced the predominance of the crucible impurities on the crystal absorption. Schott has even been able to qualify a new supplier in order to provide the industry with a second source, Parthier said, and the other supplier’s purity level is also in the range of Schott’s target.
Significant progress has been achieved on the purity of both raw materials — Al2O3 and Lu2O3 — and could be reduced close to the actual specified target (<6N). However, especially for Al2O3, researchers have measured a strong fluctuation of impurity concentrations — from lot to lot as well as within a single lot. The purity level of the Al2O3 raw material seems to dominate LuAG’s absorption factor, according to Parthier. And although Al2O3 purity has improved, it is not stable and it is insufficient. Further improvements are required and are being addressed at the supplier, and the introduction of additional Al2O3 purification processes at Schott will be necessary to push progress further.
The researchers have begun a root cause analysis, Parthier said, and there are several areas of potential improvement. But they need more time. “We expect that we can significantly reduce the impurity level,” he said. “We need three or four months to realize this at the supplier level.”
But time is something that high-index immersion lithography is running short on. There were already concerns from the industry about the availability of high-index materials in time for a 32 nm node insertion, and the latest announcement from Schott was cause for further concern. Through presentations throughout the day on Tuesday, as well as a High-Index Materials panel discussion Tuesday evening, chipmakers and toolmakers voiced their concerns and demands for faster progress if high-index technology is to be ready in time.
