Carbon abundances in O-type stars: Completing the picture of massive star evolution

Abstract

Massive stars are key agents on the chemical evolution of their parent galaxy. Their extreme conditions reached at their cores along their various stages of evolution allow them to produce elements as heavy as iron that will be released into the interstellar medium after their dramatic deaths in the form of supernova explosions. Despite their importance as chemical engines of the Universe, the current picture of massive star evolution has many layers of complexity which are still far from being fully understood: recent studies have shown that rotation, multiplicity and their strong mass-loss impact significantly their evolutionary paths. An empirical evidence of the products of nuclear fusion is the study of chemical surface abundances. Particularly, carbon, nitrogen and oxygen are key elements since they are involved in the CNO cycle, the main hydrogen fusion mechanism of massive stars. In this work we performed a quantitative spectroscopic analysis on a sample of O-type stars to determine their carbon surface abundances in order to compare them with the prediction of the evolutionary models. Our sample consists of 104 stars with spectra belonging to the IACOB spectroscopic database that had their parameters (effective temperature, surface gravity, projected rotational velocity and wind parameters) previously determined. Additionally, prior to the spectroscopic analysis, we performed a detailed investigation of all the available carbon diagnostic lines in the optical spectra and their suitability to perform a reliable estimation of the surface carbon abundance. We ended with set of 14 carbon spectral lines from three different ions that proved suitable for the abundance analysis. Our abundance analysis relies on a methodology based on the measurement of equivalent widths (EWs). The abundances were determined by comparing the measured EWs with the predictions of FASTWIND stellar atmosphere models and after a χ2 minimization, the best pair of abundance and microturbulence (ξ) was adopted. We obtained a carbon abundance baseline of ϵC = log(C/H) = 8.37 ± 0.15(dex), which is in agreement with previous results from the literature. At last, we found that when comparing the determined abundances with parameters related to stellar evolution (rotational velocity, helium abundance and surface gravity), most of the stars agree with the predictions of single stellar evolutionary models with moderate rotation, with 20% of outliers which we postulate that could be explained as products of binary interaction.