VIPA (Virtually Imaged Phased Array)
A tilted plate with a high-reflectivity entrance face (except a small entrance window) and a partially transmitting output face. Produces a fan of spatially offset leakage beams from repeated internal bounces — the geometric walk-off a real VIPA disperser relies on. Specified by center wavelength, spectral resolution and free spectral range, the same as the Etalon.
Open in the canvas →In the real world
A VIPA is, at heart, the same tilted Fabry–Pérot cavity as an etalon — two closely spaced reflective coatings — but illuminated and read out completely differently. Light enters through a small uncoated window in an otherwise near-perfectly reflective front face, focused to a line inside the cavity. Because the plate is tilted relative to that incoming beam, each internal bounce off the partially transmitting back face leaks light out at a slightly different lateral position instead of retracing the same path — producing a fan of many spatially offset, mutually coherent beams that interfere in the far field exactly like light emerging from a real phased array of point sources, except every one of those virtual sources is actually a single physical cavity imaged multiple times[1]. That's the "virtually imaged" half of the name.
The result is angular dispersion 10–20× higher than an ordinary diffraction grating in a device a few millimeters thick, at the cost of a much smaller free spectral range — which is why VIPAs are typically paired with a grating in a cross-dispersed configuration (the grating separates orders that would otherwise overlap) in high-resolution spectrometers, optical coherence tomography systems, and dense wavelength-division-multiplexing demultiplexers.
In OpticalSetup
Because the walk-off between successive leaked beams is a purely geometric
consequence of the tilt — each bounce genuinely exits at a different point along
the plate — OpticalSetup traces it directly as repeated ordinary mirror
reflections rather than borrowing the Etalon's closed-form Airy transmission: an
entrance window in the front coating lets rays in, and each subsequent bounce off
the partially reflective rear face spawns both a continuing internal ray and a
leaked output ray, exactly reproducing the fan of offset beams a real VIPA
produces. Only the output face's reflectivity needs the Fabry–Pérot mathematics,
and it's derived the same way the Etalon derives its mirror reflectivity: you
specify center wavelength, resolution (FWHM), and free spectral range, and
resolveVipaPhysical() solves for the plate spacing and coating
reflectivity that would actually produce them — sharing its solver with the Etalon
element, since spectrally the two are the same cavity.
The fan of leaked beams is genuine ray-traced geometry, but each individual leaked ray still carries only the ordinary (incoherent) intensity propagated by the rest of the tracer — the far-field interference between those beams that a real VIPA relies on to build its angular dispersion pattern isn't computed; what you see is the correct geometric walk-off, not a simulated diffraction pattern. There's also no modeled anti-reflection coating on the entrance window, no cylindrical input-lens focusing, and no cross-dispersing grating stage — this element models the VIPA plate alone.