A physical model of mass ejection in failed supernovae

During the core collapse of massive stars, the formation of the protoneutron star is accompanied by the emission of a significant amount of mass-energy ($\sim 0.3 \, M_{\odot}$) in the form of neutrinos. This mass-energy loss generates an outward-propagating pressure wave that steepens into a shock near the stellar surface, potentially powering a weak transient associated with an otherwise-failed supernova. We analytically investigate this mass-loss-induced wave generation and propagation. Heuristic arguments provide an accurate estimate of the amount of energy contained in the outgoing sound pulse. We then develop a general formalism for analyzing the response of the star to centrally concentrated mass loss in linear perturbation theory. To build intuition, we apply this formalism to polytropic stellar models, finding qualitative and quantitative agreement with simulations and heuristic arguments. We also apply our results to realistic pre-collapse massive star progenitors (both giants and compact stars). Our analytic results for the sound pulse energy, excitation radius, and steepening in the stellar envelope are in good agreement with full time-dependent hydrodynamic simulations. We show that {prior} to the sound pulses arrival at the stellar photosphere, the photosphere has already reached velocities $\sim 20-100 \%$ of the local sound speed, thus likely modestly decreasing the stellar effective temperature prior to the star disappearing. Our results provide important constraints on the physical properties and observational appearance of failed supernovae.
15 pages, 10 figures, submitted to MNRAS