The winds of OBA hypergiants and luminous blue variables: Dynamically-consistent atmosphere models reveal multiple wind regimes

TL;DR

This work tackles how winds form and appear in OBA hypergiants and LBVs near the Eddington limit by employing hydrodynamically consistent PoWR atmosphere models across a wide $T_\mathrm{eff}$ range at $\Gamma_\mathrm{e} \sim 0.4$. The authors uncover a complex, temperature-dependent mass-loss pattern featuring two wind solutions, 'dense' and 'airy', and bi-stability jumps driven by Fe ion recombinations, with turbulent pressure playing a crucial role at cooler temperatures. They demonstrate that observable spectra resemble known hypergiants and LBVs across the regime, while current mass-loss prescriptions fail to capture the full behavior, especially the second bi-stability jump. The results imply that near-Eddington winds can switch between wind regimes only in a limited parameter space and emphasize the need for broader modeling, including radiatively-driven turbulence and metallicity variations, to predict feedback from these stars. Overall, the study provides a nuanced framework linking Fe ionization, turbulent pressure, and wind regime switches to the wind properties and spectra of the most luminous massive stars, with implications for their evolution and environmental impact.

Abstract

OBA hypergiants (OBAHGs) are evolved massive stars with notable wind features in their optical spectrum. Positioned at the cool edge of the line-driven wind regime, many are candidate luminous blue variables (LBVs) likely near the Eddington Limit. Although brief, this evolutionary stage deeply impacts their surroundings and subsequent evolution. We study the mechanisms behind OBAHG winds and spectra, covering the temperature range of non-eruptive LBVs. Using the PoWR atmosphere code, we compute models with an Eddington parameter Gamma_e ~ 0.4 and moderate turbulent pressure, typical for cool hypergiants, varying the effective temperature from ~12.5 to ~38.0 kK at solar metallicity. Our models show a complex temperature-dependent mass-loss pattern, with regions of higher/lower rates linked to two wind solutions: "dense" and "airy." Spectra of known OBAHGs and LBVs match models from all solution regions. We find bi-stability jumps -- with sharp mass-loss increases -- at temperatures where Fe IV recombines to Fe III (and Fe III to Fe II). "Drops" in mass-loss also occur when the leading Fe ion changes at wind onset, signaling a switch to airy solutions under insufficient driving opacity. The resulting velocity fields also reflect these different regimes: airy solutions match the empirical terminal velocity vs temperature relation, while dense ones deviate. Turbulent pressure is crucial for wind acceleration at cooler temperatures, especially in airy regimes. We demonstrate that the bi-stability jumps exist in OBAHGs but are part of a broader complex behavior not replicated by current mass-loss recipes. Combining our and other recent results, we suggest that the switch between airy and dense solutions only occurs within a certain proximity to the Eddington Limit. Testing this requires future models with broader parameters and advanced treatments of radiatively-driven turbulence.

Figures (23)

• Figure 1: HR diagram depicting our model sequence against LBVs (cyan squares), B[e]SG (blue triangles), OBA hypergiants/candidates (green circles), YSGs (small yellow pentagons), and YHGs (big gold pentagons) from the literature -- see Appendix Sect. \ref{['sec:distance-lum']}. The luminosities are updated to distances mostly based on Gaia DR3 BailerJones+2021GaiaCollab+2023. The black line indicates the luminosity of our sequence. The small diamonds below indicate the $T_*$ while the large hollow diamonds indicate the $T_\mathrm{2/3}$. The thin dotted lines indicate the galactic evolutionary tracks by Ekstroem+2012. The thick cyan line indicates the LBV instability strip Groh+2011 and the thick violet line indicates the HD limit Humphreys-Davidson1979.
• Figure 2: Effective temperature at different optical depths $\tau$, normalized by $T_\star = T_\mathrm{eff}(\tau=50)$ for each point of the sequence. The pink curve indicates the $T_{2/3}$ and the blue curve indicates the $T_\mathrm{eff}$ at the critical point.
• Figure 3: Non-tailored comparison between the output spectra of some of the models (red lines) and observed spectra (black lines) of OB supergiant/hypergiants and LBVs.
• Figure 4: Transformed radius ($R_\mathrm{t}$) versus $T_\ast$: Model with $R_\mathrm{t} < 200$ R$_\odot$ are in the "dense" wind regime -- i.e., with higher $\dot{M}$ and optical spectra crowded with emission/P Cygni features -- (shaded region). The colored texts in the lower part of the diagram name the wind-solution regions identified in the $T_\star$ sequence. The 1-BiSJ and the 2-BiSJ are indicated by the dotted and dashed lines in the transitions from a valley to a plateau (i.e., increase in $\dot{M}$) when moving towards lower $T_\star$.
• Figure 5: Contributions of different processes/elements/ions to the wind launching at the critical point (stacked area plot) and corresponding mass-loss rates (red line) for different inner boundary temperatures ($T_*$). Bottom-up, the light beige area indicate the electron scattering acceleration, the dark beige the pressure, and the smokewhite indicate HI continuum driving. The teal tones indicate the contribution of CNO and other metals while the shades of blue indicate the contributions of different Fe ions.