[Paper Review] Mass loss from late-type WN stars and its Z-dependence: very massive stars approaching the Eddington limit

G. Graefener, W.‐R. Hamann|ArXiv.org|Mar 6, 2008

Gamma-ray bursts and supernovae78 references117 citations

TL;DR

This paper presents the first self-consistent hydrodynamic non-LTE model atmospheres for late-type WN (WNL) stars, demonstrating that their strong mass loss is driven by radiative acceleration in optically thick winds near the Eddington limit. The models reveal a strong dependence of mass loss on metallicity (Z), but also on the Eddington factor Γₑ and stellar temperature, showing that even at very low Z, high-mass stars can sustain WR-type winds—especially if enriched with primary nitrogen, which may have played a key role in early interstellar nitrogen enrichment.

ABSTRACT

The mass loss from Wolf-Rayet (WR) stars is of fundamental importance for the final fate of massive stars and their chemical yields. Its Z-dependence is discussed in relation to the formation of long-duration Gamma Ray Bursts (GRBs) and the yields from early stellar generations. However, the mechanism of formation of WR-type stellar winds is still under debate. We present the first fully self-consistent atmosphere/wind models for late-type WN stars. We investigate the mechanisms leading to their strong mass loss, and examine the dependence on stellar parameters, in particular on the metallicity Z. We identify WNL stars as very massive stars close to the Eddington limit, potentially still in the phase of central H-burning. Due to their high L/M ratios, these stars develop optically thick, radiatively driven winds. These winds show qualitatively different properties than the thin winds of OB stars. The resultant mass loss depends strongly on Z, but also on the Eddington factor, and the stellar temperature. We combine our results in a parametrized mass loss recipe for WNL stars. According to our present model computations, stars close to the Eddington limit tend to form strong WR-type winds, even at very low Z. Our models thus predict an efficient mass loss mechanism for low metallicity stars. For extremely metal-poor stars, we find that the self-enrichment with primary nitrogen can drive WR-type mass loss. These first WN stars might play an important role in the enrichment of the early ISM with freshly produced nitrogen.

Motivation & Objective

• To understand the physical mechanism behind the strong mass loss in late-type WN (WNL) stars, which are critical for massive star evolution and chemical yields.
• To resolve the long-standing debate on whether WR winds are radiatively driven, especially given the high wind performance numbers (η ≈ 1–10) that challenge standard radiation pressure theory.
• To investigate the Z-dependence of mass loss in WNL stars, particularly in the context of early massive stars and the formation of long-duration Gamma Ray Bursts.
• To determine whether extremely metal-poor stars, including those enriched with primary nitrogen, can still drive strong winds despite low metallicity.
• To develop a physically grounded, parametrized mass loss recipe for WNL stars based on self-consistent modeling of wind hydrodynamics and line-blanketing.

Proposed method

• The study employs a new generation of Potsdam Wolf-Rayet (PoWR) non-LTE model atmospheres with self-consistent hydrodynamic wind solutions, solving the full radiative transfer in the expanding wind.
• The models include Fe-group line-blanketing and wind clumping, enabling realistic treatment of radiative acceleration in optically thick, radiatively driven winds.
• A systematic parameter study is performed across stellar parameters, including luminosity-to-mass ratio, effective temperature, metallicity (Z), and Eddington factor (Γₑ), to isolate dependencies.
• The models are validated by direct comparison with observed WNL spectra, particularly for the spectroscopic binary WR 22.
• The wind acceleration is computed from deep atmospheric layers, ensuring the sonic point is self-consistently located where the Fe-peak opacity enhances radiative driving.
• An effective metallicity (Z_eff) is proposed to account for the differing efficiency of CNO elements versus Fe-group elements in driving winds, using a weighted formula based on elemental contributions.

Experimental results

Research questions

• RQ1 What physical mechanism drives the strong, radiatively accelerated winds observed in late-type WN (WNL) stars, especially given their high wind performance numbers (η ≈ 1–10)?
• RQ2 How does the mass loss rate from WNL stars depend on metallicity (Z), and can strong winds persist at very low metallicities?
• RQ3 To what extent does the Eddington factor (Γₑ) influence mass loss, and how does it modify the Z-dependence of mass loss in WNL stars?
• RQ4 Can extremely metal-poor stars enriched with primary nitrogen (produced via CNO cycling) still drive strong WR-type winds, and what is the role of such stars in early interstellar medium enrichment?
• RQ5 What is the role of Fe-group elements versus CNO elements in wind driving, and how should metallicity be effectively parameterized in mass loss recipes for WNL stars?

Key findings

• WNL stars are very massive, luminous stars near the Eddington limit, still in central hydrogen-burning, with L/M ratios that drive optically thick, radiatively accelerated winds.
• The models reproduce the observed WNL spectral sequence from WN6 to WN9 qualitatively, with wind densities consistent with galactic WNL stars except for the extreme WN8 subtype.
• Mass loss in WNL stars depends strongly on both metallicity (Z) and the Eddington factor (Γₑ), with the Z-dependence becoming flatter at higher Γₑ, allowing strong mass loss even at very low Z.
• For extremely metal-poor stars (Z < 10⁻⁴ Z☉), self-enrichment with primary nitrogen can drive WR-type winds, enabling efficient mass loss and nitrogen enrichment of the early interstellar medium.
• Iron-group elements are far more effective in driving winds than CNO elements; thus, an effective metallicity Z_eff must weight contributions from different elements, with a proposed formula Z_eff ≈ Z☉·(1/50·X_CNO/X_CNO,☉ + X_Fe/X_Fe,☉).
• The models confirm that WNL stars can maintain high mass loss rates at low Z, supporting the idea that such stars could be progenitors of long-duration GRBs and key contributors to early chemical enrichment.