Abundances of light elements in metal-poor stars - II. Non-LTE abundance corrections

Citation
Rg. Gratton et al., Abundances of light elements in metal-poor stars - II. Non-LTE abundance corrections, ASTRON ASTR, 350(3), 1999, pp. 955-969
Citations number
95
Categorie Soggetti
Space Sciences
Journal title
ASTRONOMY AND ASTROPHYSICS
ISSN journal
00046361 → ACNP
Volume
350
Issue
3
Year of publication
1999
Pages
955 - 969
Database
ISI
SICI code
0004-6361(199910)350:3<955:AOLEIM>2.0.ZU;2-5
Abstract
We present non-LTE corrections to abundances of Fe, O, Na, and Mg derived f rom LTE analyses of F-K stars over a broad range of gravities and metal abu ndances; they were obtained using statistical equilibrium calculations and new model atoms. Line opacity was considered by means of an empirical proce dure where it was attributed to a veil of weak Fe I lines; in the case of s olar-type dwarfs, results were compared with those obtained using (LTE) mea n intensities computed from OSMARCS models. We think that the empirical pro cedure produces better results for metal-poor stars, while mean intensities should perhaps be preferred for the Sun (where departures from LTE are any way not very large). Collisions with both electrons and H I atoms were cons idered. Since cross sections for this second mechanism are very poorly know n, lye calibrated them empirically by matching observations of RR Lyrae var iables at minimum light (discussed in Clementini et al. 1995). These stars were selected because non-LTE effects are expected to be larger in these st ars than in those usually considered in the study of the chemical evolution of the Galaxy (cool main sequence and red giant branch stars). We found th at different non-LTE mechanisms are important for the different species and transitions considered; on the whole, our calculations yielded moderate co rrections to LTE abundances for high excitation O lines in warm dwarfs and giants, Na and Mg lines in giants and supergiants, and Fe I lines in F-supe rgiants (where corrections becomes very large for IR O lines). Non-LTE corr ections were found to be negligible in the other cases studied.