A benchmark for binary star interaction with a supermassive black hole in general relativity

TL;DR

This study compares different numerical schemes for simulating binary star interactions with supermassive black holes in general relativity, revealing significant differences in their predictions especially near black holes.

Contribution

It introduces and compares multiple PN and perturbation schemes for three-body simulations involving SMBHs, highlighting their relative reliability and limitations.

Findings

Pair-wise PN method consistently decreases binary separation at pericentre.

Higher order PN and metric-perturbation schemes agree for stellar-mass black holes.

Discrepancies increase around billion-solar-mass black holes, affecting simulation accuracy.

Abstract

Most galaxies have supermassive black holes (SMBH) at their centres, surrounded by stars with binary systems also present in this environment. We use two schemes - post-Newtonian (PN) and a scalar perturbation to a background metric to numerically solve the three-body problem of a binary with a SMBH. We test three different PN formulations for the PN scheme: The Einstein-Infeld-Hoffman equation, pair-wise implementation of two-body PN-terms for three bodies and the Arnowitt-Deser-Misner Hamiltonian. We compare these approaches for one million solar mass and one billion solar mass black holes, and find a statistical match between the two approximations for stellar mass binary interacting with a million solar mass black hole. We also perform a statistical study for encounters with this black hole, and find that the higher order PN formulation matches with metric-with-perturbation scheme. However, we find a decrease in separation of the binary, and eccentricity variations between different schemes around the billion solar mass black hole. This behaviour is not present if binary has a large separation or is further away from the black hole due to decreased general-relativistic effects. We find that the pair-wise PN method results in a decrease in separation at pericentre in all test cases irrespective of the distance from the black hole or mass of the black hole, making this the least reliable method for solving this problem. Our work highlights the need for caution when interpreting the results in different formulations around SMBHs. This also shows that when understanding extreme mass ratio inspirals (EMRIs) using simulations, one should beware as the binary gets closer to the black hole.