Immersion Tech Likely to See the Light
By Jeff Chappell -- Electronic News, 1/29/2004
LOS ANGELES – It would seem work on immersion lithography has moved from proof of concept to being readied for production in less than two years.
Here at R&D consortium International Sematech's third and final workshop on immersion technology, all three of the major litho tool vendors affirmed their immersion tool roadmaps, with production tools slated for 2006.
On top of that, in presentation after presentation on various aspects of the technology, from resists to fluid delivery to tool manufacturing, the conclusions were the same: Additional engineering work remains, but there doesn't appear to be any reason why the technology can't be ready for production at the 65nm node.
Simply put, immersion technology involves placing a layer of fluid between the wafer and the lens of an exposure tool. Aside from the number of engineering challenges this introduces, the technology can potentially provide a number of benefits, such as effectively decreasing the wavelength of the exposure light, and providing a considerable boost to the depth of focus.
But perhaps more importantly, it can preserve the industry's investment in 193nm lithography infrastructure for one or more generations.
Immersion technology has existed for considerably longer than the semiconductor, but only in the last couple of years has the chip industry seriously considered it as a means of extending optical technology once again.
"We have made a tremendous amount of progress," said Sematech immersion litho program chair, Walt Trybula. "We would be wrong to go back and think everything is solid," he said at the conclusion of the workshop. "We have to convert this [engineering solution] into manufacturing. But everything is solid."
Japanese supplier Nikon Corp. confirmed once again that it would be shipping an engineering evaluation tool in Q3 of this year, a full-field tool with a numerical aperture (NA) of 0.85. It plans to have a production-worthy tool with an NA of 0.92 by 2005, and a tool with an NA of 1 by 2006.
Fellow Japanese exposure toolmaker Canon Corp., meanwhile, said that it began building two engineering evaluation systems last year, with its eyes on 2006 for production-ready 193nm immersion tools. The use of immersion technology with 193nm lithography currently appears to have a cost of ownership advantage over the next generation of optical lithography, 157nm technology, said Hideo Hata, a member of Canon's R&D division.
Canon envisions standard 193nm lithography being used in conjunction with reticle enhancement techniques for production at the 90nm node, and 193nm immersion technology used in production at the 65nm and 45nm nodes, Hata explained.
Dutch tool vendor ASML also reiterated that it would be shipping an immersion tool to Taiwanese foundry giant Taiwan Semiconductor Manufacturing Co. in Q3 of this year. Based on the 1250 Twinscan and dubbed the 1250i, ASML's Donis Flagello emphasized that the tool was not an engineering evaluation tool, but a manufacturing tool.
The next step after shipping the first 1250i will be to produce production-ready tools, Flagello concluded.
Don't Count Your Immersion Tools Before They Hatch
While researchers and tool vendors gathered here agree that there are no apparent roadblocks or showstoppers to implementing immersion technology, its adoption is still not a foregone conclusion. As a number of researchers observed here at the workshop, the industry won't have a full grasp of the challenges of implementing the technology in production until wafers are rolling off pilot lines with full-field immersion exposure tools.
Issues with resists and the use of water as an immersion fluid seem to be largely under control, but such things as water purity, lens coating, wafer edge exposures, scanning speeds and bubbles may not throw up technological roadblocks, but could still prove to be challenging headaches.
While other types of liquid besides water or additives to water may be used beyond the 65nm node to increase the refractive index, it appears as if water will be the immersion fluid of choice for the technology's first implementation in production. But it's not just water. Massachusetts Institute of Technology/Lincoln Labs researcher Mike Switkes noted that deionized water doesn't work, nor does buying and storing ultrapure water from an outside source, from a defect standpoint -- substances tend to leach into the water, reducing its effectiveness at transmitting exposure light.
In conducting its immersion R&D, Lincoln Labs had to build its own on-site water purification system. "It was a non-trivial exercise for us, and took some time and some doing," Switkes said.
As for bubbles, if immersion ever proves to have an Achilles heel, it would most likely be because of bubbles in the immersion fluid, which could potentially act as particle defects in terms of exposing a reticle pattern on a wafer.
Bubbles can be introduced during a variety of steps in the immersion lithography process. They can come from air trapped in wafer topography, for instance, or arise from the thermal effect created by the tool's laser pulse. Or they can occur during the injection or removal of the fluid from the space between lens and wafer.
So far, much of immersion research has centered around bubbles, and will continue to do so in the near future. While all of the researchers here concluded that bubbles aren't show stoppers, and indeed may prove to be less of a problem than once thought in some instances, as ASML's Flagello put it, it's a "non-trivial" issue. He identified defect formation along with the high numerical apertures required at the 65nm node and beyond.
While current cutting edge 193nm production tools have lenses with NAs of 0.85, the first production immersion tools, slated for the 65nm node, will have NAs around 1, according to toolmakers roadmaps. The NA will go higher for the 45nm node. But the higher NAs will require more complex lenses and lens designs. Again, this may not a showstopper, per se, but it is another non-trivial issue. |
