Extreme ultraviolet (EUV) lithography, which has an exposure (actinic) wavelength of 13.5 nm, has been chosen by the computer chipmaking industry as the path forward for extending minimum chip feature sizes downward beyond the capabilities of today’s 193 nm deep-UV lithography. The results of a gigantic decades-long effort has led to the commercial introduction of EUV lithographic scanners by ASML (Veldhoven, Netherlands), which rely on a light source consisting of a large pulsed carbon dioxide (CO2) laser whose beam is focused one-by-one on a stream of liquid tin droplets to form a light-emitting plasma. The hugely complex light source requires tight alignment of the laser beams to the tin droplets and also produces optics-damaging debris from the process.
A group from the Ioffe Institute (St. Petersburg, Russia) is proposing an alternate solution: the laser-based production of a plasma from xenon (Xe) gas to produce light at a wavelength of 11.2 nm. Interestingly, Xe plasma also can produce light at the existing EUV wavelength of 13.5 nm (which was selected because molybdenum/silicon multilayer mirror coatings work well at 13.5 nm), but its low efficiency drove the industry to select tin as a plasma source instead. However, the Russian researchers have experimentally developed a process to make Xe produce light at 11.2 nm much more efficiently than at 13.5 nm; the technique includes irradiating the target with a defocused beam, raising EUV output by an order of magnitude to a high conversion efficiency of 3.9%. Along with the hypothesis that adding some beryllium to the molybdenum/silicon mirror structure can boost reflectivity at 11.2 nm, the use of Xe for a plasma source could make possible a more trouble-free light source for EUV chipmaking. Reference: S. G. Kalmykov et al., J. Appl. Phys.,126, 103301 (2019); https://doi.org/10.1063/1.5115785.
