Free software

About OpticalSetup

Copyright © 2026 Luca Genchi and contributors.

OpticalSetup is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

It is distributed in the hope that it will be useful, but without any warranty; without even the implied warranty of merchantability or fitness for a particular purpose. See the GNU General Public License for more details.

The numbers OpticalSetup reports come from models with stated limits; read each component's page in the wiki before relying on one.

If you use OpticalSetup in your research, teaching or publications, please cite the software:

Bertoncini, A., & Genchi, L. (2026). OpticalSetup - the optical workbench software [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.23191537

Examples / Ultrashort Pulses / Prism compressor

Prism compressor

A 31 fs pulse stretched to 122 fs by 30 mm of N-BK7, and brought back to 31 fs by four Brewster prisms whose GDD comes from the traced ray paths.

Open in the canvas →

A live trace of the setup above, shown as a picture. Open it in the canvas to take it apart, retune it, and save or export it as your own.

Background

Glass gives a pulse positive group delay dispersion: the red end of its spectrum travels faster than the blue, so the pulse leaves longer than it arrived. A prism pair gives the opposite sign without needing a lens or a grating. The first prism fans the colours out at slightly different angles; the second, cut and turned the same way but inverted, catches each colour where it has arrived and sends it on parallel to the others.

Each colour leaves with the phase φ(ω) = ωP(ω)/c of its optical path P; its group delay is dφ/dω and the GDD is d²φ/dω². What makes the pair's GDD negative is how that phase changes with frequency, not which colour's path happens to be longer: through the pair the blue colours have the shorter group delay and arrive before the red ones, the opposite of what glass does[1].

Fork, Martinez and Gordon showed that the optical path between the wavefronts through the two apices is simply P = l cos β, with l the distance between the apices and β the angle between that line and each colour's ray. The GDD that follows from ωP/c is negative and grows with l, so the spacing within a pair sets it; pushing a prism further into the beam adds glass, and with it positive GDD, so insertion fine-tunes it. A second, mirror-image pair (or a mirror sending the light back through the first) gathers the colours back into one beam; between the two pairs the colours travel parallel, so that gap adds no GDD.

What this setup demonstrates

Three identical 31 fs, 800 nm transform-limited Gaussian sources, each measured by its own autocorrelator wired to a screen.

The first arm is the reference. The second passes through 30 mm of N-BK7, about +1 340 fs², and the pulse stretches to about 122 fs. The third adds four N-SF10 prisms cut for Brewster's angle at 800 nm, in two mirror-image pairs; within each pair the central ray crosses 161 mm of air between the prisms, and their apices are 165.3 mm apart. Nothing in the scene states their GDD: each colour of the pulse is traced through the glass and the air between, and the autocorrelator reads the GDD from the phases of those paths. Here they cancel the rod to within a fraction of a fs², and the pulse is back to 31 fs.

Select a prism and move it. Spreading the two prisms of a pair further apart along the beam (keeping every colour on their faces) adds negative GDD; moving the second pair as a whole further from the first changes nothing, since the colours are parallel there; pushing a prism into the beam adds the glass back. The screen of the third arm follows, and its inspector names where the GDD came from.

What you won't see

The duration is computed from the GDD alone: the prisms' third-order dispersion, which limits real compressors for very short pulses, is in the traced phase but not in the duration (the duration model line says so when it is large). The layout is two-dimensional, so the folded double pass of a lab compressor, where a mirror sends the beam back slightly lower and a D-shaped mirror picks it off, is drawn here as the equivalent four-prism line. Surface losses are a fixed fraction per face rather than the Fresnel loss at the actual angle and polarization. The pulse packets drawn along the beam still use each colour's glass GDD alone: after the prisms the drawn packet looks like a pulse of about 260 fs, while the autocorrelator, which reads the GDD from the ray paths, gives 31 fs. The recompression shows in the detector reading, not in the drawn packets.

Related components

References

  1. R. L. Fork, O. E. Martinez and J. P. Gordon, “Negative dispersion using pairs of prisms,” Opt. Lett. 9, 150–152 (1984)

Further reading