EUV's dirty secret is that the light is hard to make. You fire a powerful CO2 laser at microscopic tin droplets tens of thousands of times a second; each hit creates a plasma that radiates 13.5 nm light. The brighter and more stable that source, the more wafers the multimillion-dollar scanner prints per hour - and throughput is the entire economic case for EUV.
TSMC's grant US11150559B2 (issued October 19, 2021; TAIWAN SEMICONDUCTOR MANUFACTURING CO., LTD.; CPC G03F 7/70033 EUV source/illumination, G03F 7/702) claims controlling laser interference fringes to boost source brightness and throughput. The G03F 7/70033 classification is the EUV-source illumination bucket.
Interference fringes are pattern artifacts that arise when coherent laser light overlaps with itself; uncontrolled, they waste energy and destabilize the droplet-laser interaction. Controlling them means more of the laser's power converts into usable EUV and less into noise - a direct lever on how bright and steady the source runs.
Note who is filing. ASML builds the scanner and the source subsystem, but here TSMC - the customer running the tool - is patenting source-throughput improvements. That is what it looks like when the chokepoint is so valuable that even the buyer invests its own IP into wringing more wafers per hour out of it.
Throughput, not resolution, is often the binding EUV constraint at this point. The optics can resolve the features; the question is whether the source is bright enough to expose wafers fast enough to justify the capital. A fringe-control grant is squarely aimed at that economic edge.
For the equipment reader, the rule is to watch the source, not just the lens. TSMC's 2021 grant shows the leading-edge foundry treating EUV light generation as a throughput problem worth its own patents - because at EUV prices, every extra wafer per hour is real money.