Abstract

We have previously reported experimental results of Raman amplification of a low-light-level seed-Stokes beam in the transient limit.¹ Here we present results of measurements which used a 10-nsec seed-injected Nd:YAG laser. The 532-nm frequency-doubled output had a duration of 8.5 ns and was used both to generate a 4.5-ns duration seed-Stokes pulse at 683 nm and to pump the 1-m long Raman amplifier. The seed-Stokes beam was spatially filtered, apertured at the Airy minimum, and collimated before the amplifier at a diameter of 5.2 mm. Likewise, the spatial mode of the pump was optimized to be as smooth as possible and was collimated into the Raman amplifier at a diameter of 3.4 mm. With no seed Stokes incident in the amplifier, random spatial patterns characteristic of the quantum-noise initiated, self­generated Stokes light were observed over a spatially gain-narrowed beam size supporting ~ 12 spatial modes. For the Raman gains of ~10⁵ used in these experiments the number of temporal modes was 1.8. The seed-Stokes light was used to illuminate a bar chart of line-width 0.315 mm, and either a direct image of the bar chart, or its Fourier transform, was amplified in the Raman amplifier. The original bar chart image was then observed using an intensified camera. As the level of seed-Stokes was reduced to that approaching one photon per spatial-temporal mode, we were able to observe intensity modulations approaching 100% within the amplified bar chart images. Modulation levels of a few percent persist out to input levels of ~10³ photons per mode. This observation is consistent with an interpretation of interference between electric field of the seed-Stokes beam and the vacuum fluctuations present in the amplifier but is inconsistent with a photon number description where the maximum predicted modulation would be only 50%.

© 1991 Optical Society of America