EUV Sources – IEEE Nano Council
November 15, 2011, Nanotechnology Council Symposium, Santa Clara, CA--Hakaru Mizoguchi from Gigaphoton talked about efforts to produce a manufacturing-grade EUV source. This photon source will be needed for the next-generation processes below 20 nm.
Some of the shortcomings for EUV sources are low power and short lifetimes. To date, the industry has promised production-quality sources in a couple of years, and it remains at a couple of years. The first step is to get a 10W source, next 100 W, and eventually 200 W.
After years of research, they have settled on CO2 lasers and liquid tin droplets as the most likely combination for EUV generation. They double pulse the laser to first, vaporize the droplets, and then ignite the vapor into a plasma. A magnetic field confines the plasma, which helps to minimize the metal migration to the mirrors. This configuration is showing promise to have a life-cycle of several thousand hours.
The latest version envisions a 13 kW compound CO2 laser with 20 nS pulses at a 100kHz rate to get an output of 100 W. The CO2 laser would be comprised of 2 continuous laser amplifiers going though a shutter to converge on 60 µ tin droplets. The test bed currently uses a 3.6 kW laser and 30 µ tin droplets to get 20 W output power for about 7 hours total.
Experiments with a laser pulse at 10 Hz and 20 µ droplets are in progress to work on conversion efficiency, debris mitigation, and laser power and duty cycle issues. The current pre-pulse to vaporize the tin is about 1 µS and the CO2 laser pulse is about 10 µS. results indicate that a 20 µ droplet is completely vaporized into a sub-micron cloud of particles. The test platform is getting 3.8 percent conversion efficiency, so scaling the pulse power should result in a 200 W output.
One big issue is the mirrors. If the droplets are not fully vaporized and ionized, then some metal mist deposits on the mirrors. The high magnetic containment field traps 99 percent of the ions in the plasma to create a EUV beam. Nevertheless, the mirrors have to be chemically periodically cleaned to maintain operational functionality.
The ongoing results project a 250 W system will be working some time in the second half of '12. The basic machine will produce 50 W of clean power. Upgrades will increase the power to 250 W, and to 350 W in '13, and possibly to 500W by '14.
There are still some engineering challenges on droplet stability and variability as well as other configuration problems before the 50 W system is ready next year. These experimental results now make it possible to say that a EUV source is no longer about 2 years away from production, but will be here next year.