2D unified atmosphere and wind simulations for a grid of O-type stars

TL;DR

This study presents a grid of 2D radiation-hydrodynamic simulations of O-star atmospheres and winds that self-consistently generate subphotospheric turbulence. It demonstrates that the maximum subphotospheric turbulent velocity scales approximately as the square of the classical Eddington parameter, $v_{ m turb} \propto \Gamma_{ m e}^2$, and correlates linearly with broadening of photospheric absorption lines, offering a physical basis for macro-turbulence. Turbulent pressure significantly affects the atmospheric structure, inflating the photosphere and altering $T_{ m eff}$ and $R_{ m ph}$ relative to 1D models, with stronger effects at higher $\Gamma_{ m e}$. Mass-loss rates and wind structure are compared with standard 1D prescriptions, revealing underestimations by several recipes for the most luminous stars and highlighting the need to incorporate multi-D turbulence effects into 1D atmosphere and evolution codes.

Abstract

The atmospheres of massive O-type stars (O stars) are dynamic, turbulent environments resulting from radiatively driven instabilities over the iron bump, located slightly beneath the stellar surface. Here, complex radiation hydrodynamic processes affect the structure of the atmosphere as well as the formation of spectral lines. In quantitative spectroscopic analysis, the effects of these processes are often parametrized with ad hoc techniques and values. This work is aimed at exploring how variation of basic atmospheric parameters affects the dynamics within the subsurface turbulent zone. We also explore how this turbulence relates to absorption lines formed in the photosphere for a broad range of O stars at solar metallically. The work in this paper centers around a grid of 2D, radiation-hydrodynamic O-star atmosphere and wind simulations, where the turbulent region is an emergent property of the simulation. For each of the 36 models in the grid, we derived the turbulent properties and correlated them to an estimate of turbulent line broadening imposed by the models' velocity fields. Our work suggests that the subphotospheric turbulent velocity in O-stars scales approximately with the square of the Eddington arameter, $Γ_{\rm e}$. We also find a linear correlation between subphotospheric turbulent velocity and the line broadening of several synthetic photospheric absorption lines. Radiation carries more energy than advection throughout the atmosphere for all models in the grid; however, for O-type supergiants, the latter can account for up to 30 \% of the total flux at the peak of the iron bump.

Figures (15)

• Figure 1: HRD positions after model relaxation for the entire grid. The gray lines indicate MIST Dotter2016 evolution tracks for non-rotating massive stars. The effective temperatures shown here are emergent properties of the simulation. Due to statistical scatter in very turbulent models, they do not line up in perfect sequences.
• Figure 2: Density maps of a relatively steady atmosphere (left), corresponding to a dwarf model D1 and a relatively turbulent atmosphere (right), corresponding to a supergiant model S12. The velocity field is indicated by the black arrows. The density is given in units of the lower boundary density (see Table \ref{['table: input']}), and the magnitude of the velocity peaks at $300 \, \rm km/s$
• Figure 3: Value for the turbulent velocity as a function of radius for model G1. The vertical black line indicates the average model photosphere. The brown and pink lines indicate the radial and lateral velocity dispersions . The black dashed line represents the average photosphere.
• Figure 4: Average radial velocity profile as a function of radius (blue), the average density weighted radial velocity profile (orange), and the best fit $\beta$-velocity law (green) for the G1 model. The black dashed and dash-dotted lines represent the average photosphere and sonic radius, respectively.
• Figure 5: Maximum value of the turbulent velocity below the photosphere, as a function of the classical Eddington factor for the entire grid of models. Red squares indicate the series of dwarf models, blue circles the series of giants, and black triangles the series of supergiants.