Verification of atomic data for solar oxygen abundance models

Abstract

ABSTRACT Recently, Bergemann et al. reported a re-analysis of O i lines, 777 and 630 nm, using their new atomic models of O and Ni, and presented solar photospheric oxygen abundance differing from others. We discuss the accuracy of atomic data employed in their models, and demonstrate that the photoionization cross-sections (σPI) of the two levels ${\rm1s^22s^22p^33s (^5S^o_2)}$ and 1s22s22p33p(5P1), which they present as illustrations of their data in the non-local-thermodynamic-equilibrium (NLTE) model, are incorrect by large factors. For example, their σPI at the threshold of level ${\rm ^5S^o_2}$ is larger by a factor of 1000 than that of Nahar and the Opacity Project data base, TOPbase, and has an incorrect energy behaviour. They included a blending of the 630 nm line with N i for which they computed σPI in the hydrogenic approximation. The approximation does not include electron–electron interaction and hence is not valid for Ni i. They noted that the use of oscillator strengths of O i by Hibbert et al., which agree very well with TOPbase and Nahar, gave an abundance close to the existing values, but reported using averaged values of two sources. Therefore, their oxygen abundance from the new atomic models is likely to be in question.