From 3D hydrodynamic simulations of common-envelope interaction to gravitational-wave mergers

Abstract

Modelling the evolution of progenitors of gravitational-wave merger events in binary stars faces two major uncertainties: The common-envelope phase and supernova kicks. These two processes are critical for the final orbital configuration of double compact-object systems with neutron stars and black holes. Predictive one-dimensional models of common-envelope interaction are lacking and multidimensional simulations are challenged by the vast range of relevant spatial and temporal scales. Here, we present three-dimensional, magnetohydrodynamic simulations of the common-envelope interaction of an initially 10'M red supergiant primary star with a black-hole and a neutron-star companion. Employing the moving-mesh code AREPO and replacing the core of the primary star and the companion with point masses, we show that the high-mass regime is accessible to full ab initio simulations. About half of the common envelope is dynamically ejected at the end of our simulations and the ejecta mass fraction keeps growing. Almost complete envelope ejection seems possible if all ionised gas left over at the end of our simulation eventually recombines and the released energy continues to help unbind the envelope. We find that the dynamical plunge-in of both companions terminates at orbital separations that are too wide for gravitational waves to merge the systems in a Hubble time. However, the orbital separations at the end of our simulations are still decreasing such that the true final value at the end of the common-envelope phase remains uncertain. We discuss the further evolution of the system based on analytical estimates. A subsequent mass-transfer episode from the remaining 3'M core of the supergiant to the compact companion does not shrink the orbit sufficiently either. A neutron-star'neutron-star and neutron-star'black-hole merger is still expected for a fraction of the systems if the supernova kick aligns favourably with the orbital motion. For double neutron star (neutron-star'black-hole) systems we estimate mergers in about 9% (1%) of cases while about 77% (94%) of binaries are disrupted; that is, supernova kicks actually enable gravitational-wave mergers in the binary systems studied here. Assuming orbits smaller by one-third after the common-envelope phase enhances the merger rates by about a factor of two. However, the large post-common-envelope orbital separations found in our simulations mean that a reduction in predicted gravitational-wave merger events appears possible.