Journal of Chemical Physics, Vol.119, No.6, 3174-3186, 2003
Velocity redistribution of excited atoms by radiative excitation transfer. I. Experimental demonstration by photodissociation of Na-2 and field-free imaging
Photodissociation of state-selected sodium molecules, Na-2(X (1)Sigma(g)(+),v('))+hnu-->Na-2(*)(B (1)Pi(u))-->Na-*(3p(3/2))+Na(3s(1/2)) has been studied theoretically and experimentally using a novel "field-free" ion imaging design. The experiment uses a supersonic Na/Na-2 beam in combination with the stimulated Raman adiabatic passage technique to prepare Na-2 molecules in selected rovibronic levels of the electronic ground state. The Na(3p(3/2)) fragments are photoionized (or excited to high Rydberg states) in a permanently field-free reaction zone. The fragments enter the ion optics because of the flow velocity of the beam and are focused onto a position sensitive detector, which provides an energy resolution of about 50 meV. The measured anisotropic photofragment angular distributions reflect the alignment of the molecules prior to dissociation and are well explained by the anisotropic nature of the photodissociation by polarized laser light. The measured images show not only the expected relatively fast photodissociation fragments, but also the efficient formation of slow Na(3p(3/2)) atoms. Fast and slow refer to the atomic velocity relative to the center of-mass of the dissociating molecule. The ratio of the numbers of slow atoms and fast photofragments is 0.16 and 0.22 for the dissociation of Na-2 from levels v(')=17 and v(')=23, respectively. Several models are analyzed to explain the observations. Calculations show that the dramatic velocity redistribution is caused by radiation trapping: the excitation is efficiently radiatively transferred from the fast Na(3p) photofragments to the abundant Na(3s) atoms from the primary beam, whereby the hyperfine splitting of the 3s state must be taken into account. Analytical formulas describing this mechanism show a ratio of slow to fast Na(3p) atoms of 0.13 for v(')=17 and 0.19 for v(')=23, which is in very good agreement with the experimental observations. (C) 2003 American Institute of Physics.