We have investigated the noise characteristics and the low-light-level imaging capabilities of a stimulated Raman amplifier. Diffraction-limited Stokes beams as well as Stokes beams carrying spatial structure have been amplified with a collimated pump with a Fresnel number in the range 3–17. Both direct imaging and Fourier-transform imaging of the Stokes signal through the Raman amplifier were used. Results indicate that approximately 1–4 photons per spatial mode are needed to produce images with a signal-to-noise ratio of 1. Amplification of the Fourier transform of the object effectively provides spatial filtering of the noise and gives the best signal-to-noise ratio at the lowest input levels for our experiments.
We have investigated the noise characteristics and the low-light-level imaging capabilities of a stimulated Raman amplifier. Diffraction-limited Stokes beams as well as Stokes beams carrying spatial structure have been amplified with a collimated pump with a Fresnel number in the range 3–17. Both direct imaging and Fourier-transform imaging of the Stokes signal through the Raman amplifier were used. Results indicate that approximately 1–4 photons per spatial mode are needed to produce images with a signal-to-noise ratio of 1. Amplification of the Fourier transform of the object effectively provides spatial filtering of the noise and gives the best signal-to-noise ratio at the lowest input levels for our experiments.
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%.
We have studied the transverse modes of a transient stimulated Raman signal generated in unsaturated Raman amplifiers with Fresnel numbers from one to seven, both with and without a seed Stokes pulse. In the absence of a seed Stokes pulse and for a pump Fresnel number greater than 1.5, the spatial intensity pattern of the stimulated Stokes signal is highly variable from shot to shot, reflecting the random nature of the quantum-noise source. However, when a separately generated and sufficiently strong seed Stokes pulse with a Gaussian spatial profile is injected into the amplifier, the spatial mode of the output Stokes beam assumes a smooth Gaussian spatial pattern that is significantly smaller in diameter than either the pump or the seed Stokes beam. This spatial mode control, which is due to the injected Stokes signal, persists down to a level of approximately 300 seed Stokes photons per spatial mode and, surprisingly, is observed to suppress the noise-initiated signal by at least a factor of 10 in the outer regions of the amplified Stokes beam. We have also studied the Stokes pulse-energy statistics over a range of Fresnel numbers. Data taken with systems having Fresnel numbers between 4 and 7 follow the same pattern as the spatial-mode-control results in that the statistical character of the Stokes pulse energies changes rapidly from that characteristic of a quantum-noise-initiated process to one characteristic of a stable source as the level of the injected seed Stokes is increased. However, for interaction regions with Fresnel numbers of 1 or smaller, the Stokes pulse-energy statistics show a gradual change as the level of the seed Stokes signal is increased and do not become stabilized until the seed Stokes is more than 100 times the total level needed to control the spatial character of the output Stokes in the larger-Fresnel-number cases. This slow change in the statistics for the Fresnel 1 system, and the suppression of the quantum-noise-initiated signal in the larger-Fresnel-number systems, is not fully understood at this time.
Stimulated Raman scattering (SRS) microscopy has emerged as a powerful technique for probing the spatiotemporal dynamics of molecular bonds with exceptional sensitivity, resolution, and speed. However, classically, its performance remains fundamentally constrained by optical shot noise, which imposes a strict limit on detection sensitivity and speed. Here, we demonstrate a quantum-enhanced SRS microscopy platform that circumvents this barrier by harnessing amplitude-squeezed light. Specifically, we generate a Stokes beam with 5.2dB of amplitude squeezing using traveling-wave optical parametric amplification in second-order nonlinear waveguides, and combine it with a tunable coherent pump to access vibrational modes spanning from 1000 to 3100cm−1. Applied to quantum imaging of metabolites in biological tissue (pork muscle), our quantum-enhanced Raman microscope achieves an average noise suppression of 3.6dB and a 51% enhancement in signal-to-noise ratio (SNR) -- to the best of our knowledge, the largest improvement reported to date in quantum-enhanced SRS microscopy of biological samples.
Measurements of the output of a Raman amplifier when there is no input Stokes seed are presented.It is shown that a fully quantum mechanical scaled planewave theory describes the growth from spontaneous emission well in the high gain limit.However, in the low gain regime there is a significant increase in the amount of Stokes output which is not predicted by the plane wave theory.To account for the deviation from planewave growth observed at low gain, a fully quantum mechanical theory of amplification in a focused gain geometry is presented.The theory is based on a non-orthogonal mode expansion of the amplified field and incorporates the effects of both gain-guiding and diffraction.In addition to accounting for the observed Stokes output at low gain the non-orthogonal mode theory predicts that the Stokes field cart experience enhanced gain as well as have excess noise.Results from an experiment are presented which verify the existence of enhanced gain in a multipass Raman amplifier.Though the experiment does not give a direct measure of the excess noise factor the results
We have studied the phase pulling and locking of an injected Stokes pulse in a transient Raman amplifier both experimentally and theoretically. In addition, we have investigated the pulse energy statistics and related spatial modes of a high gain amplified Stokes pulse. Experiments reported here show that in the unsaturated case, the phase of the amplified Stokes is pulled almost instantaneously to that of the pump, as soon as integrated gain levels of slightly greater than unity are reached. For higher input seed-Stokes levels, such that the amplified Stokes energy is saturated, the phase locking is completed early, and the phase of the amplified Stokes then follows the phase excursion of the pump. The phase pulling has been modelled theoretically for the specific parameters of our experiment. We have also used the pulse energy statistics to study the seed-Stokes control of a high gain Raman amplifier. For an input seed-Stokes beam to control a transient, high gain Raman amplifier, operated near threshold, it must dominate the self-generated Raman which builds from quantum noise. For a near-unity-Fresnel-number Raman amplifier we have measured the number of input seed-Stokes photons needed to produce an unchanged statistical signature on the amplified Stokes pulse and have found that a surprisingly large number (109) were needed. For a larger Fresnel number Raman amplifier, a near-field camera was used to study the level at which the seed-Stokes pulse controlled the spatial output of the amplifier. Control of the spatial modes seems to have a sharp threshold with input Stokes energy and was fully established with an input of 2x103 photons per mode.
Probing the quantum noise level with respect to a coherent input has recently received much attention. Several researchers have probed this question by examining the level of external optical input needed to influence the output field of a quantum noise amplifier. For example, it has been found that it takes, on average, about one photon per (temporal) mode to control the phase of the output field of a Raman amplifier.1 A similar result has been seen in optical parametric amplification.2 In contrast to using light to seed the amplification, we have investigated the level of initial molecular coherence needed to control the phase of the output field of a Raman amplifier. Two laser pulses, separated in time, were used to pump a common set of hydrogen molecules producing two Stokes pulses via stimulated Raman scattering. The relative phase of the two Stokes pulses was measured interferometrically. The variance of this relative phase is a direct measure of the memory of the molecular coherence that was induced by the first scattering process. Correlations between the two Stokes phases were observed even after nine molecular dephasing times, when the level of coherence had decayed by a factor of e−9. We show that this extraordinarily long-lived phase memory comes about because a molecular coherence corresponding to order ten vibrational photons per mode is sufficient to influence the phase of the output field.