To clump or not to clump

Abstract

Context. Winds of massive stars have density inhomogeneities (clumping) that may affect the formation of spectral lines in different ways, depending on their formation region. Most of previous and current spectroscopic analyses have been performed in the optical or ultraviolet domain. However, massive stars are often hidden behind dense clouds rendering near-infrared observations necessary. It is thus inevitable to compare the results of such analyses and the effects of clumping in the optical and the near-infrared, where lines share most of the line formation region. Aims. Our objective is to investigate whether a spectroscopic analysis using either optical or infrared observations results in the same stellar parameters with comparable accuracy, and whether clumping affects them in different ways. Methods. We analyzed optical and near-infrared observations of a set of massive O stars with spectral types O4-O9.5 and all luminosity classes. We used Fastwind model atmospheres with and without optically thin clumping. We first studied the differences in the stellar parameters derived from the optical and the infrared using unclumped models. Based on a coarse model grid, different clumping stratifications were tested. A subset of four linear clumping laws was selected to study the differences in the stellar parameters derived from clumped and unclumped models, and from the optical and the infrared wavelength regions. Results. We obtain similar stellar parameters in the optical and the infrared, although with larger uncertainties in the near-infrared, both with and without clumping, albeit with some individual deviating cases. We find that the inclusion of clumping improves the fit to Hα or He II 4686 in the optical for supergiants, as well as that of Brγ in the near-infrared, but it sometimes worsens the fit to He II 2.18 μm. Globally, there are no significant differences when using the clumping laws tested in this work. We also find that the high-lying Br lines in the infrared should be studied in more detail in the future. Conclusions. The infrared can be used for spectroscopic analyses, giving similar parameters as from the optical, though with larger uncertainties. The best fits to different lines are obtained with different (linear) clumping laws, indicating that the wind structure may be more complex than adopted in the present work. No clumping law results in a better global fit, or improves the consistency between optical and infrared stellar parameters. Our work shows that the optical and infrared lines are not sufficient to break the dichotomy between the mass-loss rate and clumping factor.