Stellar oxygen abundances. 4: Systematic effects on oxygen abundances derived from the 6300 A (O I) and 7774 A O I lines
We consider potential systematic effects on oxygen abundances derived from the 6300 A (OI) line and the 7774 A OI triplet. Our solar intensity spectra of the 7774 A triplet confirm previous results which indicate a discrepancy between the observed equivalent widths of Altrock (1968) and the values predicted by Local Thermodynamic Equilibrium (LTE) and recent NLTE models. However, this disagreement (at low micron(s)) does not seem to affect the solar O abundance as derived from flux spectra. We derive O abundances for a selection of relatively metal-rich F and G dwarfs from both the 6300 A (OI) line and 7774 A OI triplet and detail the various uncertainties which enter into the analyses. Minimizing possible systematic effects to the extent possible, we find for T(sub eff) approximately less than 6200-6300 K no systematic difference between the 6300 and 7774 A abundances. For T(sub eff) approximately greater than 6200-6300 K, however 7774 A abundances are substantially larger than the 6300 A abundances. This agreement in O abundances from the two features at cooler T(sub eff) conflicts with that of others and we suggest that the discrepancy may be due to the different model atmospheres utilized. If recently proposed, hotter T(sub eff) values for metal-poor dwarfs are correct, then there appears to be no discrepancy between the 6300 A abundances of metal-poor giants are correct, then there appears to be no discrepancy between the 6300 A abundances of metal-poor giants or dwarfs and the 7774 A abundances for dwarfs. This would seem to rule out substantial LTE departures or atmospheric inhomogeneity effects skewing metal-poor O abundances from the 7774 A triplet (for cooler stars having low metallicity anyway). Given the repeated inability of authors to reproduce each others' O abundances from the 6300 A (OI) line and the uncertainties in the solar equivalent width, we question the usual assumption that the 6300 A (OI) line and the uncertainties in the solar equivalent width, we question the usual assumption that the 6300 A (OI) abundances are more reliable than those from the permitted triplet.