Quiescent mass loss during the red supergiant (RSG) phase has been shown to be far lower than prescriptions typically employed in single-star evolutionary models. Importantly, RSG winds are too weak to drive the production of Wolf-Rayet (WR) stars and stripped-envelope supernovae (SE-SNe) at initial masses of roughly 20–40 M ⊙. If single stars are to make WR stars and SE-SNe, this shifts the burden of mass loss to rare dust-enshrouded RSGs (DE-RSGs), objects claimed to represent a short-lived,

more »

... igh-mass-loss phase. Here, we take a fresh look at the purported DE-RSGs. By modeling the mid-IR excesses of the full sample of RSGs in the Large Magellanic Cloud, we find that only one RSG has both a high mass-loss rate ( M ̇ ≥ 10−4 M ⊙ yr−1) and a high optical circumstellar dust extinction (7.92 mag). This RSG is WOH G64, and it is the only one of the 14 originally proposed DE-RSGs that is actually dust enshrouded. The rest appear to be either normal RSGs without strong IR excess, or lower-mass asymptotic giant branch stars. Only one additional object in the full catalog of RSGs (not previously identified as a DE-RSG) shows strong mid-IR excess. We conclude that if DE-RSGs do represent a pre-SN phase of enhanced M ̇ in single stars, it is extremely short-lived, only capable of removing ≤2 M ⊙ of material. This rules out the single-star post-RSG pathway for the production of WR stars, luminous blue variables, and SE-SNe. Single-star models should not employ M ̇ -prescriptions based on these extreme objects for any significant fraction of the RSG phase.