To appreciate the coherent beam combining problem, it is instructive to assess the physical limits of beam combining dictated by the radiance theorem. Specifically, the radiance theorem implies that, given N uncorrelated modes, the power per mode cannot be increased by any optical system that is in thermal equilibrium. Since the fields from mutually incoherent lasers are uncorrelated, the light from an ensemble of these beams can be described as containing multiple modes, one from each laser. From the radiance theorem, the only way to increase the radiance beyond that of a single source is to reduce the number of modes in the system. The radiance can therefore be maximized by forcing all the individual modes into a single coherent state by establishing a fixed (non-time-varying) phase relationship between each of the laser sources.
Montoya et al. have used a Dammann grating resonator to couple light between individual lasers in the array. In a manner similar to mechanical oscillators that are coupled together by interconnecting springs, this coupled optical resonator oscillates in multiple “supermodes”, each with its own characteristic loss. By ensuring that the losses to all but one supermode are large enough to prevent lasing, the resonator can be forced to oscillate in a single supermode with all lasers in the array contributing to its power. The Dammann grating resonator has two advantages in this regard. First, this resonator is inherently “single mode”, meaning that all supermodes save one suffer complete attenuation in one cavity round-trip. Thus, coherence between lasers is assured. Second, the intensity of this single supermode is the same at each laser in the array, meaning that the power can be extracted from all the lasers in an optimal manner.
Properly designed coupled cavities can ensure that the power per mode is enhanced. However, this does not necessarily mean that the power per unit area per unit solid angle (the traditional definition of radiance) is optimal in any sense. To achieve this, the mutually coherent light from the various lasers in the array must be converted into a plane wave or Gaussian beam by an appropriate optical system. In the case of the Dammann grating cavity, the grating itself performs a superposition of the fields, producing a beam shape that is similar to a single laser in the array. The grating effectively acts as a multi-port beam splitter, efficiently coupling the coherent state from the multiple arms of the individual lasers to a single output port.
For proper operation, one more important requirement must be met: each laser must be adjusted to have the proper phase relationship required by the coherent state. Incorrect laser phases both increase the loss to the desired supermode (making the resonator less efficient) and prevent the grating from properly channeling light into a single output port. In some coherent beam combining systems, the lasers’ ability to change longitudinal modes can partially compensate for phase errors. However, when this is not the case, some form of active phase control is required. Montoya et al. have employed current tuning to adjust the phases of each laser. Changes in laser current modify its temperature and electron density, both of which can modify the phase of the individual laser beam. Each laser phase is then adjusted by a control system using the stochastic parallel gradient descent algorithm to optimize the output power.
The combination of an active control system and state-of-the-art sources consisting of slab-coupled optical amplifiers arrays (SCOWA) has led to truly outstanding performance for coherent beam combining of semiconductor lasers. The authors report coherent beam combining of a 21 element SCOWA device with a non-optimized combining efficiency of 81 % and a total output power of 1.2 watts. They also note that, although active phase control was employed, system stability was sufficient to permit high efficiency operation for up to an hour between phase updates. Indeed, coherent beam combining of semiconductor lasers with external resonant cavities appears to have finally come of age.
You must log in to add comments.