The frequency-dispersed birefringence and dichroism of an electronically nonresonant liquid excited and probed by ultrafast pulses in an optically heterodyned detected configuration is reported. The nominally putative dichroic response of a transparent sample is shown to result from π out-of-phase contributions of Stokes and anti-Stokes third-order polarization components on, respectively, the red and blue sides of the probe pulse spectrum and are derived from CSRS and CARS type resonances. The strong corresponding nondispersed birefringent response, in contrast, results from the in-phase combination of these Stokes and anti-Stokes polarization components. Thus, by observing the dispersed probe pulse to the red or blue of the central carrier frequency, the various density matrix components, or pathways in time evolution history, contributing to either a dichroic or birefringent measurement may be more selectively viewed. The contribution of a particular nuclear response is enhanced when the observed frequency within the probe pulse spectrum is tuned to be either one quantum to the red or blue of the carrier frequency. This frequency-filtering technique can be used to enhance weak features in the total response, allow the determination of isotropic and anisotropic contributions to a nuclear response, and help probe the homogeneous and inhomogeneous character of the low-frequency Raman active density of states commonly observed in these two-pulse responses of nonresonant liquids. These effects are illustrated by the OHD birefringence and dichroism of CHCl3.

The depolarized reduced Raman and corresponding optical Kerr effect (OKE) spectral density of ambient CS2 have been calculated by way of time correlation function (TCF) and instantaneous normal mode (INM) methods and compared with experimental OKE data. When compared in the reduced Raman spectrum form, where the INM spectrum is proportional to the squared polarizability derivative weighted density of states (DOS), the INM results agree nearly quantitatively (at all but the lowest frequencies) with the TCF results. Both are in excellent agreement with experimental measurements. The INM signal has a significant contribution from the imaginary INMs. Within our INM theory of spectroscopy the imaginary INMs contribute like the real modes, at the magnitude of their imaginary frequency. When only the real modes are allowed to contribute, and the spectrum is rescaled to account for the missing degrees of freedom, the results are much poorer, as has been observed previously. When the spectra are compared in their OKE form, the INM spectrum is found to lack the low-frequency spike which is associated with long time scale rotational diffusion, and it is not surprising that an INM theory would not capture such a feature. The results demonstrate that while the OKE and spontaneous depolarized Raman spectrum contain the same information, they clearly highlight different dynamical time scales. At higher frequencies (ω>25 cm−1) the INM OKE results are in excellent agreement with TCF and experimental results. The TCF results capture the low-frequency spike and are in agreement with experiment everywhere within the precision of the present calculations. The molecular contributions to the OKE signal are analyzed using INM methods.

No abstract

No abstract