The effect of AGN feedback on the migration timescale of supermassive black holes binaries

TL;DR

This study investigates how active galactic nucleus (AGN) feedback influences the migration timescale of supermassive black hole binaries, revealing that AGN winds can stall binary coalescence in rapid migration phases.

Contribution

The paper introduces a novel implementation of AGN feedback in SPH simulations and demonstrates its significant impact on SMBHB evolution during fast migration.

Findings

AGN feedback negligible in slow migration regime

AGN winds create feedback cavities in fast migration regime

Feedback can stall SMBHB orbital decay

Abstract

The gravitational interaction at parsec to sub-parsec scales between a circumbinary gas disc and a super massive black hole binary (SMBHB) is a promising mechanism to drive the migration of SMBHBs toward coalescence. The typical dynamical evolution can be separated in two main regimes: I) Slow migration ($T_{\rm mig}$ $\sim$ $10^{3-4}\times T_{\rm orb}$), where viscous torques are not efficient enough to redistribute the extracted angular momentum from the binary, leading to the formation of a low density cavity around the binary. II) Fast migration ($T_{\rm mig}$ $\sim$ $10^{1-2} \times T_{\rm orb}$), in which the redistribution of angular momentum is efficient and no low density cavity is formed in the circumbinary disc. Using N-Body/SPH simulations we study the effect of AGN feedback in this phase of a SMBHB evolution. We implement an AGN feedback model in the SPH code Gadget-3 that includes momentum feedback from winds, X-ray heating/radial-momentum and Eddington force. Using this implementation we run a set of simulations of SMBHB+disc in the two main shrinking regimes. From these simulations we conclude that the effect of the AGN mechanical feedback is negligible for SMBHBs in the slowly shrinking regime. However, in the fast shrinking regime the AGN wind excavate a "feedback cavity" leaving the SMBHB almost naked, thus stalling the orbital decay of the binary.