Abstract

We study the formation of a supermassive black hole (SMBH) binary and the shrinking of the separation of the two holes to sub-pc scales starting from a realistic major merger between two gas-rich spiral galaxies with mass comparable to our Milky Way. The simulations, carried out with the Adaptive Mesh Refinement (AMR) code RAMSES, are capable of resolving separations as small as 0.1 pc. The collision of the two galaxies produces a gravo-turbulent rotating nuclear disk with mass (10^9 Msun) and size (60 pc) in excellent agreement with previous SPH simulations with particle splitting that used a similar setup (Mayer et al. 2007) but were limited to separations of a few parsecs. The AMR results confirm that the two black holes sink rapidly as a result of dynamical friction onto the gaseous background, reaching a separation of 1 pc in less than 10^7 yr. We show that the dynamical friction wake is well resolved by our model and we find good agreement with analytical predictions of the drag force as a function of the Mach number. Below 1 pc, black hole pairing slows down significantly, as the relative velocity between the sinking SMBH becomes highly subsonic and the mass contained within their orbit falls below the mass of the binary itself, rendering dynamical friction ineffective. In this final stage, the black holes have not opened a gap as the gaseous background is highly pressurized in the center. Non-axisymmetric gas torques do not arise to restart sinking in absence of efficient dynamical friction, at variance with previous calculations using idealized equilibrium nuclear disk models. (abridged)

Highlights

  • The coalescence of two supermassive black holes (SMBHs), namely black holes in the mass range 105–109 M⊙, would produce the loudest gravitational wave signals in the Universe, and is the main target of planned low-frequency gravitational wave experiments, such as space-based laser interferometers and pulsar timing arrays (Vecchio 2004; Sesana, Volonteri & Haardt 2009)

  • We study the formation of a supermassive black hole (SMBH) binary and the shrinking of the separation of the two holes to sub-parsec scales starting from a realistic major merger between two gas-rich spiral galaxies with mass comparable to our Milky Way

  • Using a different thermodynamical model than the previous smoothed particle hydrodynamics (SPH) simulations results in a denser, more stable nuclear disc, but, as we show in more details, this leads to inefficient hydrodynamical friction and failure of the model to harden the binary system down to sub-parsec scales in the centre of this nuclear disc

Read more

Summary

INTRODUCTION

The coalescence of two supermassive black holes (SMBHs), namely black holes in the mass range 105–109 M⊙, would produce the loudest gravitational wave signals in the Universe, and is the main target of planned low-frequency gravitational wave experiments, such as space-based laser interferometers and pulsar timing arrays (Vecchio 2004; Sesana, Volonteri & Haardt 2009). Loss of orbital energy can occur via dynamical friction on to the stellar background (Milosavljevic & Merritt 2001) or due to the gas drag (Escala et al 2004) Both mechanisms are relevant since SMBHs are inferred to exist at the centre of both gas-rich spirals and gas-poor ellipticals/S0s (Volonteri, Haardt & Gultekin 2008). The two black holes spiral down to parsec scales in a gaseous, dense nuclear disc a hundred pc in size formed by the dramatic gas inflow in the merger. The configuration of the host on which the two black holes are found at less than parsec scales is not really known because no computation exists that can reach such a stage starting from a realistic merger, and the decay in such regime is not yet explored by three-dimensional simulations.

HYDRODYNAMICAL SIMULATION PA RAMETERS
Gas physics and equation of state
RESULTS
Nuclear disc formation and turbulence dissipation
Dynamical friction in a gaseous medium
A transition from fast to slow orbital decay
Comparison with previous smoothed particle hydrodynamics simulations
Resolving the drag force and the associated wake
Is there a last parsec problem?
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call