Wavelength-Tunable Lasers: Does the Tuner Degrade the Power Efficiency?

Posted on 2008-10-03 as a part of the Photonics Spotlight.

Permanent link: http://www.rp-photonics.com/spotlight_2008_10_03.html

Author: Dr. Rüdiger Paschotta, RP Photonics Consulting GmbH

Ref.: encyclopedia articles on tunable lasers, wavelength tuning, optical filters

In a wavelength-tunable laser, some kind of adjustable optical bandpass filter such as a birefringent tuner is often used for adjusting the emission wavelength. As such a tuner introduces wavelength-dependent power losses, one may wonder to what extent it reduces the power efficiency of the laser, and on which factors this power reduction depends.

A tunable continuous-wave laser usually operates with a relatively narrow emission bandwidth. The question is then to what extent the filter introduces losses in that narrow wavelength region. In principle, it could have virtually zero losses in some wavelength interval and strong losses outside that interval. One might then expect that the power efficiency of the laser is not affected at all, simply because there is no loss for the circulating light, having the right wavelength to be well transmitted.

Of course, the power may be reduced when forcing the laser to operate at some wavelength with low gain. Even when not enforcing oscillation at an “inconvenient” wavelength, the power efficiency may be reduced if the laser doesn't “like” to emit with a narrow bandwidth. For example, spatial hole burning in the laser crystal may reduce the power efficiency when single-frequency operation is enforced. Other kinds of inhomogeneous saturation may also play a role, even without single-frequency operation.

Another problem arises when the filter has no steep edge in its transmission spectrum. For example, consider the use of an etalon as a tuner. The etalon shouldn't be too thick in order to select one particular transmission window, but that window is then relatively broad. A similar situation occurs when a prism pair is used to spatially disperse different wavelengths in some region of the laser resonator, and a knife edge is inserted in that region. The oscillation wavelength of the laser will then be a kind of compromise between minimum filter loss and maximum gain. Therefore, it does introduce significant loss for the oscillating wavelength, even though there would be some other wavelength nearby with even smaller losses. It can then be helpful to use a filter with a steeper spectral transmission function. It is even possible that the insertion of such a narrowband filter in addition to the original tuner increases the output power by forcing the laser to operate where the loss is lower. Of course, one would then remove the first tuner, as it is not more required – except if the steeper filter is a thicker etalon, and the thinner one is required to select the right transmission window.