Claudio Mendoza (IVIC/CeCalCULA)
The difficulties to obtain accurate atomic data for astrophysical applications are well known, and perhaps reinforced in the many entries in this forum. But we have not come across discrepancies of six orders of magnitude in atomic data until the recent paper “Collision rates and the determination of atmospheric parameters” by A. Spielfiedel, N. Feautrier, M. Guitou & A. K. Belyaev, presented at the Annual Meeting of the Sociéte Francaise d’Astronomie et d’Astrophysique (SF2A). Interestingly enough, this is an interdisciplinary work where quantum chemistry, collision physics and astrophysical modeling are combined to study the collisional excitation of Mg and O by H atoms, which could have an impact on NLTE atmosphere modeling. The resulting cross sections and rate coefficients appear to indicate that usual approximate formulae such as Drawin and Kaulakys lead to errors up to a factor of 106. Regarding abundance determinations, collisions with H atoms could be important in low metallicity stars when the diagnostics involve lines between excited states.