Isotope shift in the electron affinity of sulfur: observation and theory

Full text: Photodetachment microscopy was performed on a beam of S- generated by a hot cathode discharge in a mixture of 98 % Ar and 2 % CS2, with the sulfur isotopes in natural abundances. Isotope 34 was selected by a Wien velocity filter. Laser excitation was provided by a CW ring laser operating with the Rhodamine 590 dye. The laser wavenumber was measured by an Angstroem WS-U lambda meter, with an accuracy better than 10-3 cm-1. Subtracting the photoelectron energy found by analyzing the electron interferogram from the photon energy, one can determine the electron affinity eA. The result for eA(34S) is 16 752.978(10) cm-1, to be compared to the previously measured eA(32S)=16 752.976(4) cm-1. Technical correlations between the two measurements lets the isotope shift Δexp = eA(34S) - eA(32S) be a little more accurate than the more imprecise electron affinity. Numerically Δexp = + 0.002(8) cm-1, in which the (2σ) error bars leave room for a normal or anomalous result. Ab initio calculations of the isotope shift on the electron affinity from the infinite-mass systems S-/S were carried out, adopting the multiconfiguration Hartree-Fock (MCHF) approach using the ATSP2K package. Our model includes in a systematic way valence correlation, limiting the core to the n=2 shell. The one-electron orbitals are optimized using single- and double- multi-reference expansions. Configuration-interaction (CI) calculations including up to 6 · 105 configuration state functions were performed in order to complete the convergence patterns of the S- energy, resulting in a unextrapolated nonrelativistic electron affinity of eA(∞S) 16 987(44)cm-1. The theoretical isotope shift value Δtheor = eA(34S) - eA(32S) = -0.0022(2)cm-1 is found to be rather small but definitely negative. The analysis of the various contributions reveals a very large specific mass shift that counterbalances the normal mass shift, while the positive field shift is smaller than the total mass contribution by one order of magnitude. (author)