Statistical study of beam replication in stimulated Raman scattering

Abstract

Experimental work on Raman beam cleanup and beam combining has long been confounded by the generation of parasitic Stokes side beams. It has been found experimentally that when the pump beams are moderately aberrated, Stokes beams which propagate along the pump beams are generated. This effect is referred to as side beam replication. By contrast, when the pump beams are strongly aberrated, 24 × diffraction limited, no replication is observed. We can attribute this phenomenon to scintillation of the pump intensity. When the diffraction scale length is large compared to the gain length then the Stokes noise propagating col- linearly with an aberrated pump beam will experience larger gain than the central, seeded Stokes beam in the regions where the pump beam intensity is higher than average. As the aberrations increase, the gain of the side beams increases at first, maximizing at χ ~ 4 (χ diffraction limited), and then decreases again due to the decrease in diffraction length. The output energy of the central Stokes decreases because diffraction causes the pump beams to spread, so that they become less efficient at amplifying the Stokes beams. Replication is strongest for moderate aberrations, χ ~ 4, in which the scintillation mechanism is predominant. In this case, we observe a spread of the output side beam intensities of over four orders of magnitude, due to the sensitive dependence of the amplification on the detailed pump beam scintillation pattern. For strong aberrations, the total side beam intensity and its variation decreases strongly. At aberration numbers χ > 20, intensity mixing due to diffraction effectively suppresses beam replication.