Envelope inflation in massive stars near the Eddington limit

Abstract

Massive stars, i.e. those stars with masses more than about 10 times that of the Sun, are important agents for the chemical and dynamical evolution of galaxies. Because of the steep mass-luminosity relation, massive stars are extremely luminous objects. Therefore, increasingly massive stars approach their so-called Eddington luminosity, i.e., the maximum luminosity they can radiate at. It is usually believed that when the Eddington limit is reached, strong outflows are initiated.<br /> We present state-of-the-art stellar evolutionary models of massive stars, and the detailed analyses of their interior structures. After investigating model grids with five different initial chemical compositions, we find that stars reach the Eddington limit in their envelope already at masses of ~30 Solar masses and above. Furthermore, instead of showing any violent behaviour upon reaching the Eddington limit, the models develop inflated envelopes, which are extended low-density regions beneath the surface. This phenomenon is mediated by the opacity and convective energy transport in the models.<br /> Luminous Blue Variables (LBVs) like S Doradus which show strong variability on a timescale of a decades have been previously suggested to be connected to the Eddington limit, although little was known about their evolutionary stage or interior structure. We find that the hot edge of the S Doradus variability strip coincides with a line beyond which our models show strong envelope inflation, indicating a possible connection between the two. Furthermore, the inflated envelope mass in our coolest models reach several Solar masses. They provide the first physical model which could explain the large mass ejected by LBVs during the so-called Giant Eruptions, like the one observed for Eta Carinae in the 19th century.<br /> To further explore the observational consequences of envelope inflation, we follow the evolution of massive hydrogen-free models through the post main-sequence phase and predict that when these inflated stars explode as supernovae, it will lead to extended rise times of the shock breakout signal. Our model closely matches the properties of SN 2008D, the only observation of a shock breakout from a supernova so far.<br /> We conclude that envelope inflation affects the evolution of massive stars from the zero-age main-sequence up to the point of explosion, and may invoke observational instabilities in the envelope that might manifest themselves as pulsations and atmospheric macroturbulence.