Why Do Red Supergiants Lose So Much Mass?

Red supergiants are colossal stars, hundreds of times larger than our Sun. Despite their beauty, they lead short, turbulent lives because they shed mass at astonishing rates. This loss is tied to gravity itself: the force that holds the star together also drives a stellar wind that pushes away outer layers. Recent studies reveal that, contrary to earlier beliefs, a star’s metal content (like iron) hardly affects the mass‑loss rate, although dust formation becomes harder in low‑metallicity environments.\n\nA common issue is that many measurement methods fail to properly adjust scaling factors, thus underestimating weaker outflows. Conversely, a dense, extended chromosphere doesn’t automatically mean a strong wind, and the role of stochastic loss is sometimes overstated.\n\nOn a population level, stars in binary systems experience interactions that can trigger sudden mass‑loss events, crucial for their final fate. Low‑mass red supergiants can’t shed enough material before core collapse, while their massive counterparts eject copious dusty winds, heat up, and turn into hotter stars before exploding as hydrogen‑rich Type II supernovae.\n\nSupernovae provide a unique window to study these processes, but it’s essential to distinguish the star’s atmosphere from the genuine wind. New techniques, such as high‑resolution infrared observations and 3‑D modeling, promise to finally uncover the “holy grail” of red supergiant mass loss.

Red supergiants (RSG) occupy the post‑main‑sequence phase of stars with initial masses between 8 and 30 M☉. Their structure features a compact helium or carbon core surrounded by an extremely extended convective envelope, with surface temperatures around 3 500 K. Mass loss occurs via a low‑velocity (10–30 km s⁻¹) but high‑density stellar wind, driven by acoustic pressure waves and radiation on silicate or carbon dust grains.\n\nThe most widely used empirical prescriptions, such as those by de Jager et al. (1988) and van Loon et al. (2005), relate the mass‑loss rate (Ṁ) to luminosity (L) and effective temperature (T_eff) through power‑law relations. However, recent radiative‑hydrodynamic models indicate a weak dependence on metallicity (Z), because radiation pressure on dust is compensated by thermal driving and convective turbulence.\n\nK‑band interferometric observations have revealed extended chromospheres spanning 2–5 R⋆, yet the correlation with Ṁ remains ambiguous: some RSG display dense chromospheres without a corresponding increase in Ṁ, suggesting the chromosphere acts more as a thermal “cushion” than a mass‑carrying agent.\n\nBinary interactions, especially Roche‑lobe overflow and common‑envelope phases, can produce Ṁ spikes up to 10⁻³ M☉ yr⁻¹, sufficient to strip the hydrogen envelope before core collapse. These scenarios explain progenitors of Type II‑b and II‑L supernovae, which exhibit signs of a dense circumstellar medium (CSM).\n\nFuture facilities like JWST and the ELT will combine high‑resolution spectroscopy with high‑contrast imaging, enabling detailed mapping of dust and gas around RSG. Moreover, 3‑D simulations using MESA‑STAR and CO5BOLD are integrating convective dynamics with dust formation, paving the way for more accurate predictions of mass‑loss rates as functions of Z, rotation, and binarity.