Lithography's old-home week

When SPIE Microlithography convenes each year, it's old-home week for the stepper, track, mask, metrology, resist, and chemicals crowd. More than 3500 conference attendees and exhibitors showed up for this year's event, which took place during the final days of February in Santa Clara. As usual, the level of technical interaction was at a fever pitch, both in the sessions and in the hallways. Although the symposium had its usual share of high-end presentations on next-generation lithography, bleeding-edge manufacturing practices, and emerging nanotechnologies, the practical matters of dollars and cents carried the day in most quarters. Squeezing as much as possible out of existing processes and products and lowering the bar for return on capital investment weighed heavy on many people's minds.

Executives from number-three stepper/scanner supplier Canon laid out a plan to pass archrivals Nikon and ASML within five years. The cornerstone of their strategy is what they call return on lithography investment, or RoLI (although they're not sure whether to pronounce the acronym "row-lee" or "rah-lee"). Ray Morgan, the company's U.S. strategic marketing manager, described a gameplan of deeper technological development and cooperative partnerships, new product R&D in the EUV and maskless areas, high-productivity tools, improved customer support across the entire litho cell, and automated process and equipment control. The launch of the new high-resolution, high-throughput FPA-6000 scanner platform, which Canon sees as extendible to the 45-nm node, is one of the linchpins of its plan for market dominance.

An overriding theme cited by many people I spoke with was the need to extend 193-nm lithography as far as possible. Photronics' Steve Carlson told me that recent announcements such as Intel's pushout of the EUV time line "affect everything.... That puts more focus on pushing 193 nm to the limit, including using mask-enhancement techniques to move down the image-fidelity and feature-size path.... How do we get more capability without a new tool or a new technology?" Carlson says the "inevitable conclusion is, the design infrastructure and reticle infrastructure have to collide, have to be more integrated....This will enhance our ability to continue innovation within the same wavelength."

In order to perform this kind of innovation within the 193-nm realm, metrology's role becomes more important than ever. KLA-Tencor's Ingrid Peterson described how several new applications have come to the fore in the past year or so. "Wafer defect inspection has become an essential part of the qualification of new resist processes. We have been asked by the resist companies and IC manufacturers to provide a sensitive method to qualify resists." She noted one case where a DRAM fab was beginning 193-nm process development and was plagued by "microbridging defects" that an earlier-generation inspection tool could not detect. The use of a newer inspection tool helped bring the defect excursion under control. Another case showed wild swings in wafer-to-wafer CD variation during the qualification of a 193-nm process because of what turned out to be certain characteristics of a new resist chemistry. "At 193 nm," Peterson concludes, "it is essential to deal with defect density issues for multiresist applications."

The combination of souped-up yet sensitive 193-nm technologies and the rollout of 300 mm increases the blood pressure of a lot of fab managers and operations VPs. One essential part of keeping ramp-ups on track is the implementation of advanced process control (APC), an area getting increased traction at both the microlithography conference and its sister symposium on advanced microelectronic manufacturing. "At 300 mm, APC will become a must-have capability analogous to what contamination-free manufacturing (CFM) was for 200 mm," said AMD's Thomas Sonderman. His presentation on his company's APC roadmap underscored some of the excellent results AMD has experienced through the use of fabwide control systems, which has led to what he called the enhancement of the three Ps: productivity, predictability, and profitability. Sonderman envisions an advanced system of run-to-run control for one or more process and metrology tools; tool and process fault detection; and fabwide supervisory-level control of electrical parameters and yield, sampling, and dispatching. "The goal is to automate the control of the entire fabrication process."