Black hole mergers in the Universe

Mergers of black hole binaries are expected to release large amounts of energy in the form of gravitational radiation. However, binary evolution models predict merger rates that are too low to be of observational interest. In this Letter, we explore the possibility that black holes become members of close binaries via dynamical interactions with other stars in dense stellar systems. In star clusters, black holes become the most massive objects within a few tens of millions of years; dynamical relaxation then causes them to sink to the cluster core, where they form binaries. These black hole binaries become more tightly bound by superelastic encounters with other cluster members and are ultimately ejected from the cluster. The majority of escaping black hole binaries have orbital periods short enough and eccentricities high enough that the emission of gravitational radiation causes them to coalesce within a few billion years. We predict a black hole merger rate of about 1.6 x 10(-7) yr(-1) Mpc(-3), implying gravity-wave detection rates substantially greater than the corresponding rates from neutron star mergers. For the first-generation Laser Interferometer Gravitational-Wave Observatory (LIGO-I), we expect about one detection during the first 2 years of operation. For its successor LIGO-II, the rate rises to roughly one detection per day. The uncertainties in these numbers are large. Event rates may drop by about an order of magnitude if the most massive clusters eject their black hole binaries early in their evolution.