Loss-cone dynamics

Abstract

Supermassive black holes can capture or disrupt stars that come sufficiently close. This paper reviews the dynamical processes by which stars or stellar remnants are placed onto loss-cone orbits and the implications for feeding rates. The capture rate is well defined for spherical galaxies with nuclear relaxation times that are shorter than the galaxy's age. However, even the dense nucleus of the Milky Way may be less than one relaxation time old, and this is certainly the case for more massive galaxies; the capture rate in such galaxies is an initial-value problem with poorly-known initial conditions and the rate can be much higher, or much lower, than the rate in a collisionally-relaxed nucleus. In nonspherical (axisymmetric, triaxial) galaxies, torquing of orbits by the mean field can dominate perturbations due to random encounters, leading to much higher capture rates than in the spherical geometry, particularly in (massive) galaxies with long central relaxation times. Relativistic precession plays a crucial role in mediating the capture of compact remnants from regions very near to the black hole, by destroying the orbital correlations that would otherwise dominate the torques. The complex dynamics of relativistic loss cones are not yet well enough understood for accurate estimates of compact-object (‘EMRI’) capture rates to be made.