Dr. Merritt and his collaborators will use the special-purpose hardware GRAPE to run direct N-body simulations with roughly a million particles, to study the motion of stars and of less massive black holes in a cusp around the Milky Way's central black hole, a few million times as massive as the Sun. This will be the first set of simulations that can represent each star near the Galactic center by a single particle. The simulations will use a new regularization scheme to follow stars that make close approaches to each other or to the black hole; this is adapted to handle extreme mass ratios between the approaching particles, and includes some relativistic effects. The code will be used to investigate how an intermediate-mass black hole of about 10,000 times the Sun's mass would affect the orbits of nuclear stars, and the build-up of the central supermassive black hole, as it spiraled inwards. The simulations will be used to test models for the origin and dynamical evolution of the observed stellar disks on scales below a parsec. Finally, the team will model relaxation of the stellar orbits in the vicinity of the central black hole, which controls how fast stars are nudged onto orbits from which the black hole will swallow them.
The work will be carried out with international collaborators in Europe and Israel, as well as colleagues and students at the Rochester Institute of Technology (RIT). The 'GRAPE group' of undergraduates and master's students from computer science and other disciplines will work on the visualization software. Dr Merritt will also work with the National Technical Institute for the Deaf on the RIT campus to attract deaf students into the GRAPE group.
Understanding the distribution of stars near the centers of galaxies that contain supermassive black holes (SMBHs) is critically important for many problems in astrophysics. These include: predicting the rate at which stars are tidally disrupted, producing luminous flares; and predicting the rate at which gravitationally-compact objects -- neutron stars and stellar-mass black holes -- are captured by SMBHs; the latter events would constitute a primary source of gravitational waves detectable by space-based interferometers. The nucleus of the Milky Way is considered to be a prototype for such systems, but recent observations of the galactic center have revealed a number of puzzling and unexpected features. For instance, instead of a steeply-rising density of old stars near the SMBH, as most theoretical models predicted, there is instead a relatively low-density region, or "core". These observations suggest that our understanding of the dynamical evolution of galactic nuclei like the Milky Way's is far from complete. The research carried out under this grant contributed in several ways to a deeper understanding of these questions. Using both analytic techniques, and large-scale N-body simulations, the formation and persistence of "unrelaxed" stellar distributions like that observed at the center of the Milky Way were investigated. The research was successful, in the sense of providing self-consistent, time-dependent models that "look" like the Milky Way nucleus and that do not require ad hoc initial conditions. If these models are correct, the dynamical state of galactic nuclei like the Milky Way's is likely to be very different than usually assumed, with a much lower density of stars and stellar remnants. Another major lacuna in our understanding of galactic nuclei is the role of relativity in the evolution of dense clusters of stars and stellar remnants. Relativistic corrections to the gravitational force law cause the orbits of stars around a SMBH to deviate from the usual Keplerian form; but the consequences of these differences have not been well understood. The research carried out under this grant revealed a new and unexpected phenomenon, dubbed the "Schwarzschild barrier". When a stellar remnant -- e.g., a stellar BH -- finds itself on a very eccentric orbit around a SMBH, the relativistic terms cause its orbit to rapidly precess. It turns out that this relativistic ("Schwarzschild") precession can effectively quench the effects of Newtonian perturbations from other stars, thus limiting how eccentric the BH orbit can become, and making it much harder for the BH to be captured by the SMBH. This means that past calculations of the rate of compact-object inspiral (the "EMRI", or "extreme-mass-ratio inspiral", problem) have substantially over-estimated the capture rate. This new result opens the door to a deeper understanding of the "relativistic loss cone" problem, and should motivate a reanalysis of the expected rate of EMRI events detectable by proposed space-based gravitational wave detectors. Other outcomes from this research included the dissemination of a sophisticated N-body code, phiGRAPE, that makes uses of special-purpose hardware (GPUs) on parallel supercomputers to accelerate N-body calculations.
