Why binary interaction does not necessarily dominate the formation of Wolf-Rayet stars at low metallicity

Abstract

Context. Classical Wolf-Rayet (WR) stars are massive, hydrogen depleted, post main-sequence stars that exhibit emission-line dominated spectra. For a given metallicity Z, stars exceeding a certain initial mass M WR single (Z) can reach the WR phase through intrinsic mass-loss or eruptions (single-star channel). In principle, stars of lower masses can reach the WR phase via stripping through binary interactions (binary channel). Because winds become weaker at low Z, it is commonly assumed that the binary channel dominates the formation of WR stars in environments with low metallicity such as the Small and Large Magellanic Clouds (SMC, LMC). However, the reported WR binary fractions of 30 − 40% in the SMC (Z = 0.002) and LMC (Z = 0.006) are comparable to that of the Galaxy (Z = 0.014), and no evidence for the dominance of the binary channel at low Z could be identified observationally. Here, we explain this apparent contradiction by considering the minimum initial mass M WR spec (Z) needed for the stripped product to appear as a WR star. Aims. By constraining M WR spec (Z) and M WR single (Z), we estimate the importance of binaries in forming WR stars as a function of Z. Methods. We calibrate M WR spec (Z) using the lowest-luminosity WR stars in the Magellanic Clouds and the Galaxy. A range of M WR single values are explored using various evolution codes. We estimate the additional contribution of the binary channel by considering the interval [M WR spec (Z), M WR single (Z)], which characterizes the initial-mass range in which the binary channel can form additional WR stars. Results. The WR-phenomenon ceases below luminosities of log L≈4. 9, 5.25, and 5.6 [L ] in the Galaxy, the LMC, and the SMC, respectively, which translates to minimum He-star masses of 7.5, 11, 17 M and minimum initial masses of M WR spec = 18, 23, 37 M . Stripped stars with lower initial masses in the respective galaxies would tend to not appear as WR stars. The minimum mass necessary for self-stripping, M WR single (Z), is strongly model dependent, but lies in the range 20 − 30, 30 − 60, and 40 M for the Galaxy, LMC, and SMC, respectively. We find that that the additional contribution of the binary channel is a non-trivial and model-dependent function of Z that cannot be conclusively claimed to be monotonically increasing with decreasing Z. Conclusions. The WR spectral appearance arises from the presence of strong winds. Therefore, both M WR spec and M WR single increase with decreasing metallicity. Considering this, we show that one should not a-priori expect that binary interactions become increasingly important in forming WR stars at low Z, or that the WR binary fraction grows with decreasing Z.

Supernovae Ib and Ic from the explosion of helium stars

“…Our models were evolved at solar metallicity. Metallicity might play a role, though probably secondary, in the evolution of binaries, and thus in the formation of stripped-envelope stars (Gilkis et al 2019;Laplace et al 2020;Shenar et al 2020). However, metallicity will probably influence primarily the WR mass loss rate so that higher metallicities will favor the production of SNe Ic relative to Ib (Georgy et al 2009).…”

BAT99 126: A multiple Wolf-Rayet system in the Large Magellanic Cloud with a massive near-contact binary

“…The relative importance of either channel remains however an ongoing debate (e.g. Vanbeveren et al 1998;Neugent & Massey 2014;Shenar et al 2020). In this context, establishing the evolutionary status of WRs in binary systems is of particular interest as it may offer ways to better constrain the impact of multiplicity on its evolutionary history and, from there, refine our physical understanding of binary evolution.…”

Binary fraction indicators in resolved stellar populations and supernova-type ratios

“…The extant observational data cannot distinguish between WR definitions in either the WR/O or the WR/RSG ratio, but hints that the revised luminosity limit suggested by Shenar et al (2020) cannot reproduce the trend in WC/WN ratio with luminosity, for which our original log(L/L ⊙ )=4.9 luminosity limit, independent of metallicity, provides a good match. If there is indeed a strong metallicity dependence in the luminosity limit for Wolf-Rayet spectroscopic identification, then the apparent discrepancy between the data and these predictions would suggest that the mass-loss rates, and especially their scaling with metallicity, in the BPASS stellar evolution models need to be revised.…”

“…Adopting the metallicity dependence suggested by Shenar et al (2020) for the minimum luminosity of classical Wolf-Rayet stars significantly changes both the metallicity and binary fraction dependence of Wolf-Rayet number type ratios. Both the WR/O and WR/RSG ratios become more strongly metallicity dependent, while the WC/WN ratio becomes less so, in mild conflict with recent observational evidence.…”

“…Recent work (Shenar et al 2020) has argued on both observational and theoretical grounds that this simple constraint is insuffi- Figure 1. Multiple fractions tested in an experimental grid to examine possible observable signatures for binary populations.…”