Dynamically consistent analysis of Galactic WN4b stars

TL;DR

This work demonstrates that hydrodynamically consistent atmosphere modelling (PoWR-HD) combined with Gaia DR3 distances can resolve the long-standing WR radius problem for WN4b stars by showing winds launched deep in the atmosphere and a nearly uniform wind onset temperature around 140,000 K. The six Galactic WN4b targets exhibit a tight wind-structure profile with a velocity plateau near 85% of the terminal velocity and mass-loss rates of a few times 10^-5 solar masses per year, challenging traditional L-only mass-loss prescriptions and favoring L/M-dependent formulations in the thick-wind regime. Comparisons with GENEC/FRANEC evolutionary tracks reveal significant tensions: spectroscopic masses exceed track masses, and predicted surface abundances and mass-loss rates do not consistently match the observed WN4b properties, implying gaps in current single-star evolution models and potential roles for multiplicity or multi-D effects. Overall, the study highlights the critical need to incorporate dynamically consistent wind physics into massive-star evolution models to improve predictions for WR populations and their feedback in galaxies.

Abstract

Many Wolf-Rayet (WR) stars have optically thick winds that cloak the hydrostatic layers of the underlying star. In these cases, traditional spectral analysis methods are plagued by degeneracies that make it difficult to constrain parameters such as the stellar radius and the deeper density and velocity structure of the atmosphere. Focussing on the regime of nitrogen-rich WN4-stars with strong emission lines, we employ hydrodynamically-consistent modelling using the PoWR-HD code branch to perform a next generation spectral analysis. The inherent coupling of the stellar and wind parameters enables us to break parameter degeneracies, constrain the wind structure, and get a mass estimate. With this information, we can draw evolutionary implications and test current mass-loss descriptions for WR stars. We selected a sample of six Galactic WN4b stars. Applying updated parallaxes from Gaia DR3 and calculating PoWR-HD models that sufficiently resemble most of their spectral appearance, we obtain new values for the stellar and wind parameters of the WN4b sample. We compare our results to previous studies employing grid models with a beta = 1 velocity structure and cross-check our derived parameters with stellar structure predictions from GENEC and FRANEC evolution tracks. For all six targets, we obtain a narrow range of stellar temperatures T~140 kK, in contrast to previous grid-model analyses. We confirm the existence of WRs with luminosities as low as log L/Lsol = 5.0 and M~5 Msol. All derived velocity fields include a plateau feature at ~85% of the terminal velocity. Both the distance updates and the switch to dynamically-consistent atmospheres lead to substantial parameter adjustments compared to earlier grid-based studies. A comparison of the derived mass-loss rates favours a different description for the WN4b sample than for WN2 stars analysed with the same methodology.

Figures (18)

• Figure 1: HR-diagram of the WN4b sample with their parameters from hamann2019 and sander2019 (black circles), updated parameters using only corrected parallaxes from Gaia DR3 (grey circles), and parameters from modelling with PoWR$^\textsc{hd}$ in this work (red). Gaia DR3 corrections are depicted by arrows from black circles to grey circles. Differences in addition to the distance corrections are illustrated with grey arrows from the grey circles to the red data points. For comparison, we also show the stars from sander2025 (cyan) as well as the H-ZAMS from ekstroem2012 and the He-ZAMS from langer1989 as the grey and green dashed lines, respectively.
• Figure 2: Same as in Fig. \ref{['fig:hrd']}, but now with the effective temperature $T_{2/3}$ corresponding to an optical depth of $\tau=2/3$, on the x-axis.
• Figure 3: Velocity stratification $\varv(r)$ comparison of the best-fit WN4b $\beta=1$ grid model (hamann2019, black connected circles) and the PoWR$^\textsc{hd}$ models in this work (red connected circles) of WR1. The $\varv(r)$ of the remaining 5 targets in this work are shown in cyan for comparison. The range of photospheric radii of the PoWR$^\textsc{hd}$ models at optical depth $\tau=2/3$$R_{2/3}$ is shaded in grey. The difference in outermost radii between the grid models and the PoWR$^\textsc{hd}$ models is due to the former being traditionally modelled up to 1 000 $R_\odot$, with the latter up to 50 000 $R_\ast$.
• Figure 4: Acceleration stratification comparison between the WN4b $\beta=1$ grid model (hamann2019, black lines) and the PoWR$^\textsc{hd}$ model (red lines) of WR1. The main wind deceleration terms (gravity $g$ and wind inertia $a_\mathrm{mech}$) are solid lines, while the main acceleration terms (radiative force $a_\mathrm{rad}$ and pressure gradient $a_\mathrm{press}$) are dashed lines. The grey horizontal line depicts the accelerations equal to gravity. The reason for the difference in outermost radii is the same as in Fig. \ref{['fig:velo_comp']}.
• Figure 5: Mass-loss rate $\dot{M}$ comparison between the WN4b $\beta=1$ grid models (hamann2019 & sander2019, black circles) and the PoWR$^\textsc{hd}$ models in this work (red squares). In similar fashion to Figs. \ref{['fig:hrd']} and \ref{['fig:hrd_t23']}, the Gaia DR3 distance correction is applied on the grid models (grey circles) separate from the $\dot{M}$ updates in this work (see Subsect. \ref{['subsec:mdot']}). The two-step parameter difference is denoted with the grey arrows per target.