The solar iron abundance: not the last word

Abstract

Determinations of the solar iron abundance have converged to the meteoritic value with the Fe II studies of Holweger et al. (1990), Biémont et al. (1991) and Hannaford et al. (1992) and the Fe i results of Holweger et al. (1991). However, the latter authors pointed out that Blackwell et al. (1984) obtained a discordant result from similar oscillator strengths. A recent debate on this lingering discrepancy by the Oxford and Kiel contenders themselves has not clarified the issue. We do so here by showing that it stems from systematic differences between equivalent widths and oscillator strengths which masquerade as difference in fitted damping enhancement factors. We first discuss the various error sources in classical abundance determination and then emulate both sides of the debate with abundance fits of our own. Our emulation of the Oxford side shows that the abundance anomaly claimed by Blackwell et al. (1984) for solar Fe i 2.2 eV lines vanishes when equivalent width measurements from other authors are combined with better evaluation of the collisional damping parameter. On the Kiel side, we find that the oscillator strengths of Bard et al. (1991) used by Holweger et al. (1991) produce a suspicious trend when used to fuit solar Fe I lines, whereas comparable application of oscillator strengths from Oxford does not. The trend is mainly set by categories of Fe i lines not measured at Oxford; for lines of overlap the two sets agree and deliver the iron abundance value A Fe = 7.62 ± 0.04 which exceeds the meteorite value. The dissimilar lines may suffer from solar line-formation effects. We conclude that the issue of the solar iron abundance remains open. Definitive oscillator strengths are still needed, as well as verification of classical abundance determination by more realistic representations of the solar photosphere and of photospheric line formation.