Studies of atomic collision processes using molecular photodissociation

We have produced fast, excited Cs atoms by photodissociation of Cs2 with a pulsed dye laser. These excited atoms in the 5 D state possess velocities that are much greater than typical thermal velocities associated with the vapor temperature. Because the photons used to break apart the molecules have well defined wavelengths (and energies), the excited atom speed distribution is fairly well defined. We detected these fast, excited state atoms through the increased Doppler broadening of the 5D3/2 → 5F5/2 absorption line observed using a narrow band cw probe laser.; A model was developed in order to predict the absorption lineshape of these fast atoms in the early time after they are produced. Over time, velocity changing collisions with ground state cesium atoms cause the excited atom velocity distribution to thermalize. Using the strong collision model, we fit our data to a linear combination of "fast" and "thermalized" atom lineshapes to yield effective thermalization rates, Gammatherm . The rate coefficient for velocity changing collisions of fast Cs(5 D3/2) atoms with ground state Cs atoms was found to be kVCC = 6.5 +/- 1.2 x 10-10 cm3s-1, corresponding to a velocity changing collision cross section of sigmaVCC =1.3 +/- 0.2 x 10-14 cm2.; In addition, preliminary work was performed on an experiment to study the velocity dependence of an energy pooling cross section. Photodissociation was used to create a population of fast 5D atoms and a second seed laser was used to create a population of 6P atoms. Weak 7D → 6P fluorescence resulting from the energy pooling process Cs(5D) + Cs(6P) → Cs(7D) + Cs(6S) was detected. However, the signal to noise ratio in these measurements was too low for us to obtain quantitative results for the energy pooling rate coefficient as a function of 5 D atom speed. A calibration experiment was performed by pumping the cesium 6S → 5D quadrupole transitions to directly produce large numbers of 5D atoms. It was determined that we must either increase the number of 5D atoms produced by photodissociation or increase our signal to noise ratio by a factor of ∼20 in order for us to successfully measure the energy pooling collision rate coefficient vs. 5D atom speed. By pumping the 6S 1/2 → 5D5/2 quadrupole line and scanning the probe laser over the 5D5/2 → 5FJ transition and measuring the ratio of fluorescence to the different 5D fine structure levels (5F 5/2,7/2 → 5DJ) a measurement of the 5F mixing rate was also obtained. We found that the fine structure mixing rate coefficient is given by kFSC5F7/2→ 5F5/2 = (1.79 +/- 0.42) x 10-8 cm 3s-1.