[en] A relatively simple setup for collection and detection of light emitted from isolated photo-excited molecular ions has been constructed. It benefits from a high collection efficiency of photons, which is accomplished by using a cylindrical ion trap where one end-cap electrode is a mesh grid combined with an aspheric condenser lens. The geometry permits nearly 10% of the emitted light to be collected and, after transmission losses, approximately 5% to be delivered to the entrance of a grating spectrometer equipped with a detector array. The high collection efficiency enables the use of pulsed tunable lasers with low repetition rates (e.g., 20 Hz) instead of continuous wave (cw) lasers or very high repetition rate (e.g., MHz) lasers that are typically used as light sources for gas-phase fluorescence experiments on molecular ions. A hole has been drilled in the cylinder electrode so that a light pulse can interact with the ion cloud in the center of the trap. Simulations indicate that these modifications to the trap do not significantly affect the storage capability and the overall shape of the ion cloud. The overlap between the ion cloud and the laser light is basically 100%, and experimentally >50% of negatively charged chromophore ions are routinely photodepleted. The performance of the setup is illustrated based on fluorescence spectra of several laser dyes, and the quality of these spectra is comparable to those reported by other groups. Finally, by replacing the optical system with a channeltron detector, we demonstrate that the setup can also be used for gas-phase action spectroscopy where either depletion or fragmentation is monitored to provide an indirect measurement on the absorption spectrum of the ion.

[en] We present a detailed experimental study of the velocity distribution of atoms cooled in an optical lattice. Our results are supported by full-quantum numerical simulations. Even though the Sisyphus effect, the responsible cooling mechanism, has been used extensively in many cold atom experiments, no detailed study of the velocity distribution has been reported previously. For the experimental as well as for the numerical investigation, it turns out that a Gaussian function is not the one that best reproduces the data for all parameters. We also fit the data to alternative functions, such as Lorentzians, Tsallis functions, and double Gaussians. In particular, a double Gaussian provides a more precise fitting to our results