micralign projection lithography cross section

# The Working Gap That Ended Mask-Wafer Contact

Perkin-Elmer's Micralign (1973) — what the gap looks like, and what it bought in yield and mask life

CONTACT vs. NON-CONTACT, TO SCALE Contact printing (1957-era) mask trapped particle resist wafer zero gap: mask touches resist directly — the particle scratches both, every single exposure Micralign projection (1973) mask mirror mirror catadioptric: reflective, not refractive controlled working gap — zero contact, ever wafer 1:1 imaging, scanning slit sweeps the field mirrors reflect every wavelength at the same angle — far less chromatic blur than a refractive lens stack the same working gap that stops scratches also sets the diffraction-limited resolution floor

One contact event compounds: a scratched mask prints the same flaw into every wafer that follows it. Removing the contact removed the defect path at its root, not just its symptom.

Resolution stayed bounded by wavelength and numerical aperture either way — reduction steppers later succeeded Micralign for finer resolution, keeping its non-contact principle intact.

WHAT ZERO CONTACT ACTUALLY BOUGHT defect-driven yield loss and mask exposures before replacement printed-defect rate (relative) → contact scratch-and-transfer defects accumulate every exposure projection no contact event, so no scratch-transfer path at all mask exposures before replacement (relative) → contact every exposure is an abrasion event projection mask never touches anything — wear mechanism is gone ↑ bar length is schematic, not measured — the direction of each gap is the historical point fabs adopted Micralign on cost-per-good-die data throughput + mask lifetime + yield, compounded — not novelty

lineage, named not retold: Contact Printing (1957) → Projection Lithography (this series, 1973) → modern reduction steppers / EUV

Go deeper with CFSGPT

Get AI-powered deep-dives, save terms, and run advanced simulations — free account.

Create Free Account