Testing the reliability of X-rays as a tool for constraining mass-loss rates of hot stars

Abstract

ABSTRACT We fit a new line shape model to Chandra X-ray spectra of the O supergiant ζ Puppis to test the robustness of mass-loss rates derived from X-ray wind line profiles against different assumed heating models. Our goal is to track the hot gas by replacing the common assumption that it is proportional to the cool gas emission measure. Instead of assuming a turn-on radius for the hot gas (as appropriate for the line-deshadowing instability internal to the wind), we parametrize the hot gas in terms of a mean-free path for accelerated low-density gas to encounter slower high-density material. This alternative model is equally successful as previous approaches at fitting X-ray spectral lines in the 5–17 Å wavelength range. We find that the characteristic radius where the hottest gas appears is inversely proportional to line-formation temperature, suggesting that stronger shocks appear generally closer to the surface. This picture is more consistent with pockets of low density, rapid acceleration at the lower boundary than with an internally generated wind instability. We also infer an overall wind mass-loss rate from the profile shapes with a technique used previously in the literature. In doing so, we find evidence that the mass-loss rate derived from X-ray wind line profiles is not robust with respect to changes in the specific heating picture used.