2018A&A...615A...4C

Context. The determination of chemical abundances constitutes a fundamental requirement for obtaining a complete picture of a star. Particularly in massive stars, CNO abundances are of prime interest, due to the nuclear CNO-cycle, and various mixing processes which bring these elements to the surface. The precise determination of carbon abundances, together with N and O, is thus a key ingredient for understanding the different phases of stellar evolution.
Aims. We aim to enable a reliable carbon spectroscopy for our unified non-LTE atmosphere codeFASTWIND.
Methods. We have developed a new carbon model atom including CII/III/IV/V, and we discuss specific problems related to carbon spectroscopy in O-type stars. We describe different tests we have performed to examine the reliability of our implementation, and investigate which mechanisms influence the carbon ionization balance. By comparing with high-resolution spectra from six O-type stars, we verified to what extent observational constraints can be reproduced by our new carbon line synthesis.
Results. Carbon lines are even more sensitive to a variation of T_eff, logg, and M, than hydrogen and helium lines. We are able to reproduce most of the observed lines from our stellar sample, and to estimate those specific carbon abundances which bring the lines from different ions into agreement (three stages in parallel for cool objects, two for intermediate O-types). For hot dwarfs and supergiants earlier than O7, X-rays from wind-embedded shocks can have an impact on the synthesized line strengths, particularly for CIV, potentially affecting the abundance determination. Dielectronic recombination has a significant impact on the ionization balance in the wind.
Conclusions. We demonstrate our capability to derive realistic carbon abundances by means ofFASTWIND, using our recently developed model atom. We find that complex effects can have a strong influence on the carbon ionization balance in hot stars. For a further understanding, the UV range needs to be explored as well. By means of detailed and available nitrogen and oxygen model atoms, we will be able to perform a complete CNO abundance analysis for larger samples of massive stars, and to provide constraints on corresponding evolutionary models and aspects.