Etalon (Fabry–Pérot)
Two closely spaced, matched partially reflective coatings. Multi-beam interference gives narrow, periodic high-transmission peaks (the free spectral range) rather than a flat partial reflection — light off-resonance reflects, light at a resonance transmits at up to the coating-limited peak. Rotating the element shifts the resonance wavelength, like tilting a real etalon.
Open in the canvas →In the real world
A Fabry–Pérot etalon is just two closely spaced, parallel, partially reflective surfaces — but unlike a single partial mirror, light inside that gap bounces back and forth indefinitely, and every one of those internal reflections leaks a little light out and interferes with all the others. Sum that infinite series of multiply-reflected beams and, at most wavelengths, the interference is destructive enough that the etalon simply reflects, behaving like an ordinary partial mirror. But at a resonance — where the round-trip phase is a multiple of 2π — every reflected component cancels almost perfectly, and transmission surges to a coating-limited peak that can approach 100% even through two mirrors that are individually 99% reflective. That counterintuitive buildup, not a simple partial transmission, is the entire operating principle.
Resonances repeat periodically in wavelength at the free spectral range (FSR), and how sharp each resonance is — how far you can detune before transmission collapses back toward zero — is set by the finesse, which climbs steeply as the mirror reflectivity approaches 1.
Because the round-trip phase δ depends on the incidence angle through cos θ, tilting an etalon shifts its resonance wavelength without changing the mirrors at all — a standard tuning technique in real optical systems, alongside temperature tuning of the spacing itself. Etalons are used intracavity in lasers to force single-longitudinal-mode operation, and standalone as narrowband spectral filters and scanning spectrum analyzers.
In OpticalSetup
The Etalon is specified the way a real one is speced on a datasheet — center wavelength, transmission bandwidth (FWHM), free spectral range, and peak transmission — rather than by the raw mirror spacing and reflectivity the Airy function actually needs. Those spectral targets are inverted internally into the matched-mirror reflectivity R and cavity spacing that produce them, then the exact closed-form Airy function above is evaluated at every ray's real incidence angle: off-resonance light reflects, on-resonance light transmits up to the configured peak, and rotating the element on the canvas shifts the resonance exactly like tilting a real etalon — because the tracer uses the ray's actual hit angle, not a separately stored tilt parameter.
This is one of only two elements in the library implementing genuine multi-beam interference rather than an idealized on/off band — the app's ray tracer otherwise never tracks phase, so the etalon is special-cased as a single surface driven by the closed-form Airy result instead of actually summing repeated internal bounces. There's no mirror-parallelism defect (wedge), no temperature drift of the spacing, and peak transmission below 100% is reached with a single lumped loss term rather than a modeled absorption or scatter mechanism on each coating.