Atmospheric NLTE models for the spectroscopic analysis of blue stars with winds

Abstract

Context. Extreme ultraviolet (EUV) and X-ray radiation emitted from wind-embedded shocks in hot, massive stars can affect the ionization balance in their outer atmospheres and can be the mechanism responsible for producing highly ionized atomic species detected in stellar wind UV spectra. Aims. To allow for these processes in the context of spectral analysis, we have implemented the emission from wind-embedded shocks and related physics into our unified, NLTE model atmosphere/spectrum synthesis code FASTWIND. Methods. The shock structure and corresponding emission is calculated as a function of user-supplied parameters (volume filling factor, radial stratification of shock strength, and radial onset of emission). We account for a temperature and density stratification inside the postshock cooling zones, calculated for radiative and adiabatic cooling in the inner and outer wind, respectively. The highenergy absorption of the cool wind is considered by adding important K-shell opacities, and corresponding Auger ionization rates have been included in the NLTE network. To test our implementation and to check the resulting effects, we calculated a comprehensive model grid with a variety of X-ray emission parameters. Results. We tested and verified our implementation carefully against corresponding results from various alternative model atmosphere codes, and studied the effects from shock emission for important ions from He, C, N, O, Si, and P. Surprisingly, dielectronic recombination turned out to play an essential role for the ionization balance of O iv/O v (particularly in dwarfs with T eff ∼ 45 000 K).Finally, we investigated the frequency dependence and radial behavior of the mass absorption coefficient, κ ν (r), which is important in the context of X-ray line formation in massive star winds. Conclusions. In almost all of the cases considered, direct ionization is of major influence because of the enhanced EUV radiation field, and Auger ionization only affects N vi and O vi significantly. The approximation of a radially constant κ ν is justified for r 1.2 R * and λ 18 Å and also for many models at longer wavelengths. To estimate the actual value of this quantity, however, the He ii opacities need to be calculated from detailed NLTE modeling, at least for wavelengths longer than 18 to 20 Å, and information on the individual CNO abundances has to be present.

Comoving frame models of hot star winds

“…The difference in opacities may be caused either by inaccurate level populations or by missing opacities. However, we found reasonable agreement with ionization fractions calculated by Carneiro et al (2016). Our tests using line data from different sources have not revealed any significant opacity gap for lighter elements with Z ≤ 20.…”

“…5 we plot the ionization fraction of selected ions at radius where v ≈ v ∞ /2 as a function of the stellar effective temperature. In our calculations we do not consider any source of enhanced X-ray radiation, which may cause X-ray ionization (Cassinelli & Olson 1979;MacFarlane et al 1994;Pauldrach et al 2001;Krtička & Kubát 2009;Carneiro et al 2016) and we also do not consider clumping (Hamann et al 2008;Muijres et al 2011). These effects may influence the comparison of our results with observations.…”

Clues on the Origin and Evolution of Massive Contact Binaries: Atmosphere Analysis of VFTS 352

“…• X-rays -The inclusion of X-rays in the fastwind calculations (Carneiro et al 2016) affects the ionization balance, driving species to higher ionization states. This then affects the ratios between lines of different ionization states for the same element.…”

Characterizing the turbulent porosity of stellar wind structure generated by the line-deshadowing instability

“…Specifically, the above analysis indicates that LDI-generated 7 Note here that because of the high-mass loss rates from OB supergiant structure has a turbulent porosity length, H ≈ ∆X/4 = R * /40; for winds with characteristic total continuum optical depth τ * ≡ κṀ /4πv∞R * , the associated porosity thickness would be thus be of order τH ≈ τ * H/R * ≈ τ * /40. Although the opacity from such b-f continuum X-ray absorption has a broad wavelength dependence, with sharp changes across ionization edges, the characteristic global optical depth τ * is still only of order a few at all X-ray wavelengths Carneiro et al 2016). (In particular, see, e.g., figure 10 of Cohen et al 2010, for the case of the OB supergiant ζ Puppis).…”