We develop analytic and numerical models of the properties of super-Eddington stellar winds, motivated by phases in stellar evolution when super-Eddington energy deposition (via, e.g. unstable fusion, wave heating, or a binary companion) heats a region near the stellar surface. This appears to occur in the giant eruptions of luminous blue variables (LBVs), Type IIn supernovae progenitors, classical novae, and X-ray bursts. We show that when the wind kinetic power exceeds Eddington, the photons are trapped and behave like a fluid. Convection does not play a significant role in the wind energy transport. The wind properties depend on the ratio of a characteristic speed in the problem <f>v\ crit\~ (̇ G)1/5</f> (where <f>̇</f> is the heating rate) to the stellar escape speed near the heating region <it>v</it><inf>esc</inf>(<it>r</it><inf>h</inf>). For <it>v</it><inf>crit</inf> ≳ <it>v</it><inf>esc</inf>(<it>r</it><inf>h</inf>), the wind kinetic power at large radii <f>̇\ w \~ ̇</f>. For <it>v</it><inf>crit</inf> ≲ <it>v</it><inf>esc</inf>(<it>r</it><inf>h</inf>), most of the energy is used to unbind the wind material and thus <f>̇\ w \ ̇</f>. Multidimensional hydrodynamic simulations without radiation diffusion using flash and one-dimensional hydrodynamic simulations with radiation diffusion using mesa are in good agreement with the analytic predictions. The photon luminosity from the wind is itself super-Eddington but in many cases the photon luminosity is likely dominated by ‘internal shocks’ in the wind. We discuss the application of our models to eruptive mass-loss from massive stars and argue that the wind models described here can account for the broad properties of LBV outflows and the enhanced mass-loss in the years prior to Type IIn core-collapse supernovae.
No takes yet. Share an insight, caveat, or question.
Quataert et al. (2016) studied this question.