X-ray lithography (XRL) uses short-wavelength X-rays (typically 0.4–5 nm, soft X-ray to hard X-ray) to expose patterns through a thin transmission mask (typically a silicon nitride or silicon carbide membrane with absorber features in gold or tungsten) onto a resist-coated substrate held in proximity (~10–50 µm gap). The short wavelength — 10 to 100 times shorter than DUV ArF (193 nm) — eliminates optical diffraction as the resolution limiter, enabling in principle sub-20 nm and even sub-10 nm feature resolution without the multilayer mirror optics EUV requires.
The core physics advantage is significant: X-rays expose through modest proximity gaps without the extreme vacuum, multilayer reflective optics, and plasma light sources that make EUV systems >$200M per tool. However, XRL faces three structural problems that have kept it at research scale for four decades. First, mask fabrication: thin membranes with sub-20 nm absorber features at <1 nm placement accuracy are extraordinarily difficult to manufacture and inspect at volume. Second, source brightness: synchrotron sources provide adequate flux but are facility-scale instruments; compact X-ray sources (laser-plasma, compact storage rings) have not achieved the brightness needed for production throughput. Third, limited depth of focus and lack of reduction optics (XRL is typically 1:1, not the 4:1 or 8:1 reduction of optical steppers).
Frontier
- Does any vendor break ground on a high-volume X-ray lithography facility by 2028?
- Can compact synchrotron X-ray sources reach sufficient brightness and cost for commercial semiconductor use outside specialist MEMS/LIGA applications?
- Is X-ray lithography fundable at pre-seed/seed — what beachhead exists that EUV or e-beam does not already serve?