Forum Suggests Lithography Driven to Drink
The water's fine. That consensus emerged from a meeting held in Los Angeles in January that was attended by more than 350 representatives from 120 companies in the semiconductor lithography industry. Under the direction of International Sematech of Austin, Texas, the participants handicapped the viability of various lithography technologies, including 193-nm immersion and 157- and 13.4-nm -- or extreme-ultraviolet -- processes. They selected 193-nm immersion as the most likely candidate for semiconductor manufacturing in 2007 and '09.
Today's state-of-the-art sources employ 193-nm radiation. The industry had planned to go to 157-nm source lithography and, eventually, to the extreme-UV. The former, however, seems dead, overcome by doubts about its economic viability -- in particular the cost of manufacturing calcium-fluoride optical elements. The limited supply and expense of those components have led to the conclusion, widely shared in the industry, that 157 nm is finished.
But time, tide and transistor size wait for no one, so the semiconductor industry has scrambled to find a way to plug the looming capability gap. One possibility is immersion lithography, which uses water or another suitable liquid as the final optical element to extend the imaging capabilities of 193-nm sources. That allows for higher numerical aperture optics and the imaging of finer linewidths. One of the implementations is known as "splash and flash" because a thin layer of water is injected onto the wafer, followed by exposure and removal of the water. This is repeated in steps across the wafer and is based on the stepper imaging tools currently used in semiconductor manufacturing.
Behind the brewing battle over the wisdom of getting wet lie money, photonics and Moore's law. The latter is the observation that circuit transistor counts double every few years, which has led to ever-cheaper computers and electronics. This drives the multibillion-dollar semiconductor and electronics industries, but the pace can continue only if circuit feature sizes constantly shrink. That, in turn, has traditionally meant a switch to ever-shorter wavelength sources, with all that that entails in terms of changing optics, lithography tools and the photoresists used in imaging.
Go to extremes?
Immersion lithography has been demonstrated and is the focus of active development. Its kinship to known 193-nm techniques makes it a favorite to win the technology derby. Nevertheless, there are those who believe that the approach may turn out to be all wet.
Peter J. Silverman, director of lithography capital equipment development at Intel Corp., thinks that immersion lithography will work, but he noted that the cost of the tools has been characterized as undetermined. "That's something of an understatement," he said. "No one really knows what immersion tools will cost. That's an issue."
Intel is a leading semiconductor manufacturer, and its concerns about the economic viability of 193-nm immersion are leading it to hedge its bets. As the lithography forum got under way, the company announced that it had given deep-UV illumination supplier Cymer Inc. $20 million in funding to develop an extreme-UV source with an eye toward manufacturing in 2009.
"The exotic nature of [extreme-ultraviolet] lithography makes it inherently much more costly if it ever becomes a manufacturing technology," countered Burn Lin, senior director of the micropatterning division at Taiwan Semiconductor Manufacturing Co. Ltd., a major dedicated semiconductor foundry. The only problem he foresees with 193-nm immersion will come in 2009, when the technology attempts to produce 45-nm half-pitch features. Numerical aperture and other aspects of the optics then may make the tools expensive. He predicted, however, that the numerical aperture required for the 45-nm half-pitch node will be kept at 1.3, which is below the point at which costs escalate.
That said, extreme-UV is making significant strides. In February, a Sematech-sponsored workshop on the process revealed progress in many areas. Vivek Bakshi, Sematech's extreme-ultraviolet source project leader, was chairman of the workshop. He said that the meeting gave the participants many reasons to be optimistic about the potential of the technology for high-volume manufacturing.
MORE FROM PHOTONICS MEDIA