Picosecond time-resolved transmission Raman data were acquired for 1 mm thick powder samples of trans-stilbene, and a Monte Carlo model was developed that can successfully model the laser and Raman pulse profiles. Photon migration broadened the incident (approximately 1 ps) probe pulse by two orders of magnitude. As expected from previous studies of Raman photon migration in backscattering mode, the transmitted Raman pulse was broader than the transmitted laser pulse and took longer to propagate through the sample. The late-arriving photons followed tortuous flight paths in excess of 50 mm on traversing the 1 mm sample. The Monte Carlo code was also used to study the spatial resolution (lateral and depth) of steady-state transmission Raman spectroscopy in the diffusion regime by examining the distribution of Raman generation positions as a function of incident beam size, sample thickness, and transport length. It was predicted that the lateral resolution should worsen linearly with sample thickness (typically the resolution was about 50% of the sample thickness), and this is an inevitable consequence of operating in the diffusion regime. The lateral resolution was better at the sample surface (essentially determined by the probe beam diameter or the collection aperture) than for buried objects, but transmission sampling was shown to be biased towards the mid-point of thick samples. Time-resolved transmission experiments should improve the lateral resolution by preferentially detecting snake photons, subject to constraints of signal-to-noise ratio.
Read full abstract