Long-term GRMHD Simulations of Neutron Star Merger Accretion Disks: Implications for Electromagnetic Counterparts

TL;DR

This study presents long-term 3D GRMHD simulations of neutron star merger accretion disks, revealing mass ejection mechanisms, jet formation, and implications for electromagnetic counterparts like kilonovae and gamma-ray bursts.

Contribution

First long-duration 3D GRMHD simulation of neutron star merger disks showing detailed mass ejection and jet formation, improving understanding of electromagnetic signals.

Findings

Ejected mass is twice that in hydrodynamic models.

Ejecta velocity averages 0.1c with broad electron fraction distribution.

Simulation predicts potential kilonova precursor from neutron freezeout.

Abstract

We investigate the long-term evolution of black hole accretion disks formed in neutron star mergers. These disks expel matter that contributes to an $r$-process kilonova, and can produce relativistic jets powering short gamma-ray bursts. Here we report the results of a three-dimensional, general-relativistic magnetohydrodynamic (GRMHD) simulation of such a disk which is evolved for long enough ($\sim 9$s, or $\sim 6\times 10^5 r_{\rm g}/c$) to achieve completion of mass ejection far from the disk. Our model starts with a poloidal field, and fully resolves the most unstable mode of the magnetorotational instability. We parameterize the dominant microphysics and neutrino cooling effects, and compare with axisymmetric hydrodynamic models with shear viscosity. The GRMHD model ejects mass in two ways: a prompt MHD-mediated outflow and a late-time, thermally-driven wind once the disk becomes advective. The total amount of unbound mass ejected ($0.013M_\odot$, or $\simeq 40\%$ of the initial torus mass) is twice as much as in hydrodynamic models, with higher average velocity ($0.1c$) and a broad electron fraction distribution with a lower average value ($0.16$). Scaling the ejected fractions to a disk mass of $\sim 0.1M_\odot$ can account for the red kilonova from GW170817 but underpredicts the blue component. About $\sim 10^{-3}M_\odot$ of material should undergo neutron freezout and could produce a bright kilonova precursor in the first few hours after the merger. With our idealized initial magnetic field configuration, we obtain a robust jet and sufficient ejecta with Lorentz factor $\sim 1-10$ to (over)produce the non-thermal emission from GW1708107.