ON THE DETERMINATION OF THE SOLAR IRON ABUNDANCE USING FE-I LINES

There presently exist two main-stream solar iron abundances determined using Fe I lines; a 'high' value of which 7.67 +/- 0.03 (Blackwell et al. 1984) is typical, and a 'low' value typified by 7.50 +/- 0.07 (Ho lweger et al. 1991). The 'low' abundance agrees with determinations us ing Fe II lines, and with the meteoritic value. The two Fe I values di ffer by 48 per cent. An in-depth discussion of the two results, and po ssible reasons for their difference, is given in the present paper. It is concluded that the Fe I lines with excitation energy less than 2.6 eV do indeed show a 'high' abundance when interpreted using the Holwe ger-Muller empirical model atmosphere. The adoption of less accurate o scillator strengths and equivalent widths, and less suitable damping c onstants, all account, in our opinion, for the 'low' value found by Ho lweger et al. (1991). However, analyses identical to those of Blackwel l et al. (1984) which use the Kurucz (ATLAS9) and the new MARCS model atmospheres instead of the Holweger-Muller empirical atmosphere, show a 'low' abundance in reasonable agreement with the results of Fe II an alyses and the meteoritic determination. These numerical models, howev er, ate less successful at reproducing observed limb-darkening than th e Holweger-Muller empirical model, although the ATLAS9 model reproduce s the ultraviolet continuous flux better than the Holweger-Muller atmo sphere. Existing numerical and empirical solar model atmospheres would therefore appear to be inadequate; they cannot both match the observe d limb-darkening and emergent flux, and give a consistent iron abundan ce determination based on Fe I and Fe II lines of all excitations.