The merger of binary white dwarf-neutron stars: Simulations in full general relativity

TL;DR

This study uses full general relativistic simulations to analyze the merger of white dwarf-neutron star binaries, revealing that the remnants are spinning objects supported by differential rotation, which may eventually collapse into black holes.

Contribution

First fully general relativistic simulations of white dwarf-neutron star mergers showing the formation of spinning Thorne-Zytkow-like objects with potential delayed collapse.

Findings

Remnants are spinning TZlOs surrounded by a massive disk.

Cooling induces collapse in head-on collision remnants but not in circular-orbit mergers.

Differential rotation supports the remnant against immediate collapse.

Abstract

We present fully general relativistic (GR) simulations of binary white dwarf-neutron star (WDNS) inspiral and merger. The initial binary is in a circular orbit at the Roche critical separation. The goal is to determine the ultimate fate of such systems. We focus on binaries whose total mass exceeds the maximum mass (Mmax) a cold, degenerate EOS can support against gravitational collapse. The time and length scales span many orders of magnitude, making fully general relativistic hydrodynamic (GRHD) simulations computationally prohibitive. For this reason, we model the WD as a "pseudo-white dwarf" (pWD) as in our binary WDNS head-on collisions study [PRD83:064002,2011]. Our GRHD simulations of a pWDNS system with a 0.98-solar-mass WD and a 1.4-solar-mass NS show that the merger remnant is a spinning Thorne-Zytkow-like Object (TZlO) surrounded by a massive disk. The final total rest mass exceeds Mmax, but the remnant does not collapse promptly. To assess whether the object will ultimately collapse after cooling, we introduce radiative thermal cooling. We first apply our cooling algorithm to TZlOs formed in WDNS head-on collisions, and show that these objects collapse and form black holes on the cooling time scale, as expected. However, when we cool the spinning TZlO formed in the merger of a circular-orbit WDNS binary, the remnant does not collapse, demonstrating that differential rotational support is sufficient to prevent collapse. Given that the final total mass exceeds Mmax, magnetic fields and/or viscosity may redistribute angular momentum and ultimately lead to delayed collapse to a BH. We infer that the merger of realistic massive WDNS binaries likely will lead to the formation of spinning TZlOs that undergo delayed collapse.