EUV Lithography
# EUV Projection Optics Carbon Contamination and Hydrogen Cleaning
The projection optics mirrors that relay the mask pattern down to the wafer face a contamination problem distinct from the collector mirror's tin debris: trace hydrocarbon molecules in the vacuum chamber, under constant EUV photon bombardment, crack and deposit as a thin carbon film directly onto the mirror surface, gradually degrading the reflectivity of optics that have nothing to do with the plasma source itself. Even in a carefully maintained vacuum, residual hydrocarbon molecules — outgassed from seals, lubricants, or other chamber materials — are never fully eliminated. EUV photons carry enough energy to break those hydrocarbon molecules apart when they strike a mirror surface, and the resulting carbon fragments bond to the mirror, slowly building up a carbon layer. Unlike collector-mirror tin deposition, which comes from the plasma source itself, carbon contamination on projection optics originates from the ambient vacuum chemistry and photon-driven chemistry at the mirror surface.
The reason even a very thin carbon layer matters is that projection optics mirrors operate near their maximum achievable reflectivity already — EUV mirrors reflect only a fraction of incident light even when perfectly clean, since no material reflects 13.5nm light with the efficiency visible-light mirrors achieve — so any additional absorption from a carbon layer removes light the system can't afford to lose, and a scanner's optical train includes multiple mirrors in series, each one multiplying this loss. With six or more mirrors between source and wafer in a typical EUV projection system, even a small percentage reflectivity loss per mirror compounds significantly across the full optical path, directly reducing how much light reaches the wafer and therefore how fast the scanner can expose each field.
Scanners address carbon buildup with the same hydrogen-radical cleaning chemistry used on the collector mirror — hydrogen reacts with deposited carbon to form volatile methane, which pumps away — running continuously during operation to hold reflectivity losses within an acceptable range across the scanner's operating life. Because carbon deposition is a continuous process driven by residual chamber chemistry and ongoing photon exposure, cleaning has to be continuous too; there's no practical way to eliminate hydrocarbon traces from the vacuum entirely, so the engineering approach is the same balance point seen with collector-mirror tin removal — deposition rate versus removal rate, held in steady-state equilibrium rather than solved outright.
The collector mirror and the projection optics face contamination from entirely different sources — plasma debris versus ambient hydrocarbon chemistry — but both rely on the same underlying hydrogen-radical cleaning principle, and both ultimately trace back to the same consequence: every percentage point of lost mirror reflectivity is a percentage point of lost scanner throughput, which is why EUV's practical economics run through contamination chemistry at every single mirror in the optical path, not just the source.