Power Scaling Limits of Optical Parametric Amplifiers

Posted on 2006-09-21. Permanent link: http://www.rp-photonics.com/spotlight_2006_09_21.html

Ref.: G. Arisholm et al., "Limits to the power scalability of high-gain optical parametric oscillators and amplifiers", J. Opt. Soc. Am. B 21 (3), 578 (2004); encyclopedia articles on gain guiding, nonlinear frequency conversion, power scaling of lasers

Most laser architectures are not truly power-scalable, because increasing powers lead to more and more detrimental thermal effects, e.g. thermal lensing with strong aberrations, affecting the beam quality and possibly also the conversion efficiency. Realizing that parametric devices such as e.g. optical parametric amplifiers or optical parametric generators do in principle not have to dissipate any energy in the amplifying element (a nonlinear crystal), one might expect that at least this category of active devices should allow for easy power scaling. The envisaged scaling law is rather simple: for doubling the output power, just double the mode areas in the crystal, so that the optical intensities remain unchanged, and we won't run into damage problems.

A first problem encountered here is that of parasitic absorption in nonlinear crystals. Although the dissipated power in a nonlinear device is usually small compared to that in a laser amplifier, thermal effects can be significant, because phase matching is a delicate issue: sometimes, temperature differences of well below one Kelvin are enough to significantly disturb the phase matching and thus degrade the conversion efficiency and/or the beam quality.

Another practical limitation may arise from the aperture of available nonlinear crystals. However, some materials such as e.g. KDP are available in very large sizes, and in most cases the crystal size is not limiting the achievable power.

A more fundamental and rather involved issue is that of gain guiding. In a high-gain parametric amplifier operated with Gaussian beams, gain guiding can be so strong that the beam radius of the amplified signal becomes rather small toward the end of the crystal – with little impact of changes of the input signal beam profile. This leads to problems just in that last section, where significant power conversion occurs. If the signal beam becomes significantly smaller than the pump beam, and if it becomes just strong enough to well deplete the pump beam in its center, a lot of pump power in the outer regions remains unconverted. If one tries to more strongly saturate the amplifier (e.g. with a stronger signal input), the beam quality breaks down.

This problem is actually not severe for relatively low-power devices, operating with tightly focused beams. This is because diffraction prevents gain guiding from having a strong effect. However, if one tries to scale up the power by scaling up the mode areas, diffraction loses its influence, and the detrimental effect of gain guiding becomes very strong. Interestingly, the resulting power scaling limits are less tight for crystals with a relatively low nonlinearity, which are operated with more tightly focused beams. So a crystal material with high nonlinearity is not necessarily most suitable for highest powers.

Note that these power scaling limits apply only to high-gain devices. This means that the problem may be solved by splitting an amplifier device into two stages: a high-gain preamplifier with moderate pump depletion (and moderate power efficiency) and a relatively low-gain power amplifier with a moderate gain of only a few decibels. Another possibility is to operate the nonlinear device with top-hat beams rather than with Gaussian beams.