OPO
An optical parametric oscillator packaged like a laser: a pump beam goes in at the back, a tunable signal and an optional idler come out of the front.
Open in the canvas →In the real world
An optical parametric oscillator turns a pump beam into two longer wavelengths, the signal and the idler, whose photon energies add up to the pump's: 1/λp = 1/λs + 1/λi. A second-order nonlinear crystal inside a resonator provides parametric gain, and once that gain overcomes the cavity losses the resonant wave builds up from noise[1,2]. Tuning the phase matching — the crystal angle, its temperature or the poling period — tunes the signal, and the idler follows by energy conservation[2].
On a bench an OPO usually arrives as a closed instrument. Synchronously pumped femtosecond and picosecond OPOs are pumped by a mode-locked laser, often frequency-doubled, with the cavity round trip matched to the pump's repetition period, so the outputs stay locked to the pump's pulse train[3]. Their pump depletion builds up as the resonant signal is amplified over many round trips and can exceed 75 %[3]. A picosecond OPO pumped by the second harmonic of a femtosecond laser, spectrally compressed to a narrow green line, is one way to obtain two synchronised narrowband colours for stimulated Raman microscopy[4].
In OpticalSetup
Point a pump beam into the rear aperture. There is no pump wavelength to set: whatever arrives within 20° of the body axis is the pump, and the only condition on it is that the signal must be longer. Light arriving at a steeper angle stays inside the box, and the Oscillation readout says why nothing came out — no pump yet, a misaligned pump, a signal not longer than the pump, an empty tuning list or zero depletion. When it works, the readout names the pump it received and the signal and idler it made. The unconverted pump is always discarded inside.
The signal leaves the port marked S on the body axis, and the idler the port marked I, a fixed distance below it and parallel to it; both rotate with the body. Output idler switches that port off, which removes the idler's power from the bench rather than handing it to the signal. If the signal is set equal to the idler wavelength — degeneracy — the two leave together through the signal port. Each output's beam diameter is its full, unclipped envelope, whatever the pump's width, and a diameter of 0 draws it as a single line. A sized pump beam's samples keep their order across that envelope, so a pump clipped by the aperture loses the part that did not get in and its output covers only part of the diameter; a single-line pump is spread across the diameter in nine samples of equal power. The diameter is an authored drawing envelope, not a calculated cavity mode or a 1/e² Gaussian width. The ports and the body grow with the diameters so the two beams never overlap. The outputs take no path inside the box: their timing is referenced to the pump's arrival at the aperture, not to a cavity length.
The conversion is the crystal's OPO mode, shared rather than copied: Pump depletion up to 95 %, the Manley–Rowe split between signal and idler, the three output-linewidth choices, and the three Output pulses choices — transform-limited, duration set with spectral phase unknown, or duration set and positively chirped as an assumed Gaussian. Light the OPO generated is never converted by it again.
Signal tuning is Fixed, Sweep — from the first wavelength at the start of the animation to the second at half the period and back — or Steps, which holds each listed wavelength for the set time and then jumps to the next, in the order written, repeats included. Entries that are not wavelengths between 100 and 11000 nm are skipped and counted in the Tuning readout, and a list with none left produces no output; a valid wavelength that leaves no idler keeps its place in the sequence and reads as an invalid signal while it plays. The Oscillation readout gives the current step and the signal and idler actually generated, following the animation. Tuning runs on the motion clock, independently of pulse playback: Pause pulse animation stops the pulses, not the tuning, as with the other moving elements. A static SVG or PNG export shows the start of the tuning program, while animation frames show their own moment. The saved signal wavelength is never changed by tuning. A spectrometer reads the instantaneous line; it does not accumulate a sweep.
This is the same phenomenological model as the crystal's OPO mode, not a cavity simulation. There is no threshold, gain, build-up, saturation or back-conversion, and no dependence on pump power, cavity length or synchronisation: an accepted pump converts its authored depletion whatever its intensity, and a pump that could not reach threshold on a real bench still converts here. Phase matching is not calculated: the signal wavelength is set directly rather than by a crystal angle, temperature or poling period, and a sweep or a list of steps is an authored tuning program, not a prediction of how fast a real OPO can tune or whether it keeps oscillating across the range.
The ports, their spacing and the rear aperture are a packaging convention for this workbench, not the layout of any particular instrument, the output beam diameters are authored rather than calculated from a cavity mode, and the fixed 20° angular acceptance is a geometric rule rather than a calculation of mode matching or coupling efficiency. Any pump wavelength converts: phase matching, and whether a real OPO could be pumped at that wavelength at all, are not checked. No path inside the box is added to the outputs. Pulse durations, spectral widths and the chirp assumption follow the crystal's OPO mode and share its limitations.
Related components
References
- J.-M. Melkonian, J.-B. Dherbecourt, M. Raybaut, A. Godard, “Optical Parametric Oscillators,” Photoniques no. 110, 53–57 (2021)
- RP Photonics Encyclopedia — Optical Parametric Oscillators
- C. F. O’Donnell, S. Chaitanya Kumar, M. Ebrahim-Zadeh, “Enhancement of efficiency in femtosecond optical parametric oscillators using group-velocity-matching in long nonlinear crystals,” APL Photonics 4, 050801 (2019)
- L. Genchi, S. P. Laptenok, D. Gonzalez-Hernandez, J. Menzies, M. Aranda, C. Liberale, “Broadband background-free stimulated Raman scattering microspectroscopy with a novel frequency modulation scheme,” APL Photonics 9, 126112 (2024)