On January 5, 2025, Lawrence Livermore National Laboratory (LLNL) is developing a thulium-based petawatt-class laser technology that is expected to replace the carbon dioxide lasers used in current extreme ultraviolet lithography (EUV) tools and increase source efficiency by a factor of about ten. This breakthrough could pave the way for a new generation of "beyond EUV" lithography systems that can fabricate chips faster and with less energy.
Currently, the energy consumption of EUV lithography systems is a major concern. Low-NA and high-NA EUV lithography systems, for example, consume as much as 1,170 kW and 1,400 kW, respectively. This high energy consumption stems from the principle of the EUV system: High-energy laser pulses evaporate a tin droplet (500,000 degrees Celsius) tens of thousands of times per second to form a plasma and emit light at a wavelength of 13.5 nanometers. Not only does this process require an extensive laser infrastructure and cooling system, it also needs to be performed in a vacuum environment to prevent EUV light from being absorbed by the air. In addition, the advanced mirrors in EUV tools only reflect a portion of the EUV light, so more powerful lasers are needed to increase throughput.
LLNL's leading "large aperture thulium laser" (BAT) technology is designed to address these issues. Unlike carbon dioxide lasers, which have a wavelength of about 10 microns, the BAT laser operates at a wavelength of 2 microns, which theoretically improves the plasma-to-EUV conversion efficiency of the tin droplets as they interact with the laser. In addition, the BAT system utilizes diode-pumped solid-state technology, which provides higher overall electrical efficiency and better thermal management than gas CO2 lasers.
Initially, the LLNL research team planned to combine the compact, high-repetition-rate BAT laser with an EUV light source system to test its interaction with tin droplets at a wavelength of 2 microns," said LLNL laser physicist Brendan Reagan. "Over the past five years, we have completed theoretical plasma simulations and proof-of-concept experiments over the past five years, laying the groundwork for this project. Our work has already made a significant impact in the field of EUV lithography, and we are now looking forward to the next step in our research."
However, applying BAT technology to semiconductor production still requires overcoming the challenges of major infrastructure modifications. Current EUV systems have taken decades to mature, so practical application of BAT technology may take longer.
Industry analyst firm TechInsights predicts that by 2030, semiconductor manufacturing plants will consume 54,000 gigawatts (GW) of electricity annually, more than the annual electricity use of Singapore or Greece. The energy consumption problem could be further exacerbated if the next generation of Hyper-Numerical Aperture (Hyper-NA) EUV lithography technology comes to market. As a result, the industry's need for more efficient, energy-saving EUV machine technology will continue to grow, and LLNL's BAT laser technology certainly opens up new possibilities for this goal.
