Abstract

Mergers of massive black hole binaries (BHBs) are expected to be a key source of gravitational waves (GWs). However, the expected number and frequency of such mergers remains highly uncertain. This owes to a long-standing puzzle known as the final parsec problem. Gravitational N-body simulations that assume a spherical distribution of stars surrounding the BHB find that the binary stalls when all stars on intersecting orbits have been ejected. This is inconsistent with available observations: every massive galaxy observed to date appears to host a single supermassive black hole, and only a handful of BHB candidates exist.   I will review the final parsec problem and present a new approach to determine whether collisionless repopulation of the binary’s losscone in non-spherical nuclei is a viable mechanism to drive binary coalescence. I will present results of direct summation N-body simulations of galaxy mergers, running on multiple GPUs. These allow to measure the binary hardening rate and determine a new proxy that depends only on the stars’ angular momentum. I will then present collisionless simulations of spherical and non-spherical isolated models to increase the number of particles sufficiently that collisional effects become unimportant and collisionless losscone refilling can be measured.