Recent studies of soliton formation in a Raman amplifier by a \ensuremath{\pi} phase shift have shown that the soliton will decay if the input Stokes beam is off resonance. Initial experiments indicated that this soliton decay is related to frequency shifts of the input Stokes beam that result from quantum fluctuations. In this paper we use the quantum statistics of the Raman spectral fluctuations to compare the predicted decay statistics of the soliton decay with the experimentally measured distribution. We find that, as conjectured, the soliton decay is directly related to the statistics of the initiating quantum fluctuations.

Measurements of the single-shot power spectrum of stimulated Raman scattering from an ${\mathrm{H}}_{2}$ Raman generator at 10 and 32 atm using a Fabry-Perot interferometer are presented. The results show that the single-shot linewidth can be much narrower than the ensemble average predicted by quantum mechanics. However, when the spectra of numerous shots are averaged together the resulting linewidth is in close agreement with the predicted result. In addition, the power spectrum exhibits large shot-to-shot fluctuations within the gain-narrowed profile which may be related to observed soliton decay in stimulated Raman scattering.

Measurements on soliton decay in stimulated Raman scattering have led to the study of frequency fluctuations in the single shot Stokes linewidth. These spectral measurements show that the spectrum of a single shot can be considerably narrower than the gain narrowed profile predicted by quantum mechanics for the ensemble average. However when spectra of many shots are averaged together the predicted result is recovered. The fluctuations seen are thought to be related to the quantum fluctuations associated with the initiating spontaneous emission. Theoretical results based on coherent mode theory are shown to be in agreement with the data.

It has been predicted that quantum fluctuations will lead to the formation of spontaneous solitons in stimulated Raman scattering. We present experimental data in support of this prediction. From an ensemble of 1000 shots we found solitonlike pulses in 101 shots. We have measured the distribution of soliton heights and delay times.

In experimental studies of stimulated Raman scattering in a multipass cell (MPC) with a frequency-doubled single-mode Nd:YAG laser at 532 nm, it is found that only forward Stokes generation is observed and that the growth and saturation of the Stokes beam are in agreement with a simple plane wave, transient theory scaled to account for the focusing, and the multiple passes. The absence of second S...

In our studies of soliton formation in stimulated Raman scattering (SRS), we have discovered that a fully quantum mechanical theory predicts large shot-to-shot fluctuations in the SRS power spectrum, which leads to soliton decay. This theory, which includes a Langevin operator to simulate collision-induced fluctuations and maintain operator consistency, can be used to generate the statistics of the macroscopic SRS output. We have also studied a simpler model which uses a random vacuum field as an input to the semiclassical SRS equations. The vacuum field was numerically generated by summing a set of electromagnetic field modes that spanned the collisional Raman linewidth. Each mode was separated in frequency by less than the resolution of the system and given a random phase for each shot. Despite the lack of Langevin terms, the shot-to-shot variations and two-frequency correlation function from the simple approach were essentially indistinguishable from the fully quantum mechanical theory. This suggests that a simulation of SRS from spontaneous emission does not always require Langevin terms in the quantum noise.

We present a comparison of the theoretical predictions of a multimode broadband model with the experimentally measured gain enhancement in a Raman amplifier. The results show that the multimode theory with fixed and totally random phases is in agreement with the data obtained from an excimer laser pumped Raman amplifier. Additionally, this theory indicates that the correlated gain can be larger than the gain for a monochromatic laser, as might be expected for a model with amplitude modulation.

We used the Stokes output from an H2 Raman generator at 10 atm to seed a Raman amplifier. Before entering the amplifier a phase shift was electrooptically placed in the Stokes seed. Although a solitonlike pulse was regularly produced in the pump beam, its amplitude varied greatly. Subsequent studies of the Stokes seed’s power spectrum revealed large shot-to-shot fluctuations. The power spectrum normally consisted of a single near-transform-limited spike whose location jumped from shot to shot and for some shots consisted of two well-resolved spikes. Druhl found theoretically that shifting the Stokes seed slightly off resonance caused the soliton pulse to decay. These frequency fluctuations may thus explain soliton decay. It has also been predicted that solitons will be produced in the output of the Raman generator due to phase shifts resulting from the quantum noise which initiates the Stokes pulse.

A comparison of the theoretical predictions of a multimode broadband model with the experimentally measured gain enhancement in a Raman amplifier is presented. The results show that the multimode theory with fixed and totally random phases is in agreement with the data obtained from an excimer-laser-pumped Raman amplifier. Additionally, this theory indicates that the correlated gain can be larger than the gain for a monochromatic laser, as might be expected for a model with amplitude modulation.

At the present time several groups are studying the use of optical phase shifts to initiate soliton formation in transient stimulated Raman scattering. 1 Our experiments use a generator-amplifier Raman laser which is pumped by a frequency-doubled Nd:YAG laser at 532 nm. The Raman medium is hydrogen at 1-100 atm. We are studying the use of an electrooptic crystal to place a π phase shift in the Stokes beam between the generator and the amplifier. This shift momentarily causes energy transfer from the Stokes to the pump beam. Surprisingly, numerical calculations indicate that in one configuration the speed of the electrooptical switch need not be faster than the coherence time of the medium to generate a soliton pulse considerably shorter than the coherence time. Thus it appears that this technique can be used to generate pulses considerably shorter than those produced using standard electro-optical switching techniques. Numerical calculations and experimental results are presented.

The spatial mode structure of the stimulated Stokes emission from a H2 Raman generator has been accurately determined by using a linear photodiode array and imaging techniques. The diffraction-limited pump beam was obtained by spatially clipping the output of an injection-locked XeCl laser. The spatial profiles of the stimulated Stokes beam are compared with the predictions of Yariv's theory of propagation in a quadratic gain medium applied to Raman scattering [ A. Yariv , Quantum Electronics ( Wiley, New York, 1975)]. While the theory compares well with the experimental data at low intensities, the spatial profile develops features at higher intensities that are not predicted by the simple theory.

Recently there has been considerable interest in the formation of solitons in three level systems, 1 such as in Raman scattering. 2-5 Solitons in stimulated Raman scattering were first predicted by Chu and Scott, 6 but their experimental realization seemed remote because of the apparent necessity for particular pulse shapes for the polarization and fields. It is now clear that solitons do indeed form 7 in SRS and that transiency plays a major role in their formation.