We investigate a model in which galactic nuclei form via the coalescence of pre-existing stellar systems containing supermassive black holes. Merger simulations are carried out using N-body algorithms that can follow the formation and decay of a black-hole binary and its effect on the surrounding stars down to sub-parsec scales. Our initial stellar systems have steep central density cusps similar to those in low-luminosity elliptical galaxies. Immediately following the merger, the density profile of the remnant is homologous with the initial density profile and the steep nuclear cusp is preserved. However the formation of a black-hole binary transfers energy to the stars and lowers the central density; continued decay of the binary creates a ρ~ r^−1 density cusp similar to those observed in bright elliptical galaxies, with a break radius that extends well beyond the sphere of gravitational influence of the black holes. Our simulations are the first to successfully produce shallow power-law cusps from mergers of galaxies with steep cusps, and our results support a picture in which the observed dependence of nuclear cusp slope on galaxy luminosity is a consequence of galaxy interactions. We discuss the implications of our results for the survivability of dark-matter cusps. We follow the decay of the black hole binary over a factor of ~20 in separation after formation of a hard binary, considerably farther than in previous simulations. We see almost no dependence of the binary’s decay rate on number of particles in the simulation, contrary to earlier studies in which a lower initial density of stars led to a more rapid depletion of the binary’s loss cone. We nevertheless argue that the decay of a black hole binary in a real galaxy would be expected to stall at separations of 0.01−1 pc unless some additional mechanism is able to extract energy from the binary (Refer to PDF file for exact formulas).

Publication Date



Also archived in: arXiv: astro-ph/0103350 v1 21 Mar 2001 and Rutgers Astrophysics Preprint Series No. 300 This work was supported by NSF grants AST 96- 17088 and 00-71099 and by NASA grants NAG5-6037 and NAG5-9046. We thank Sverre Aarseth, Marc Hemsendorf and Rainer Spurzem for their patient and expert guidance with the N-body codes NBODY6 and NBODY6++ , and for making their programs available to us in advance of general release. Without their generous help this project would not have been possible. We discussed the potential observability of binary supermassive black holes with E. Sadler and J. Wrobel. The pre-hard-binary merger simulations presented here (§2) were carried out by Fidel Cruz using the tree code GADGET on the Rutgers HPC-10000 supercomputer; we are indebted to him and to Volker Springel for advice about using this code. Dr. Cruz was supported by a fellowship from the Consejo Nacional de Ciencia y Tecnologia de Mexico. This work was partially supported by the National Computational Science Alliance under grant no. MCA00N010N and utilized the San Diego Supercomputer Center Cray T3E. We are also grateful to the Center for Advanced Information Processing at Rutgers University for their generous allocation of computer time on the HPC-10000.ISSN:1538-4357 Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works in February 2014.

Document Type


Department, Program, or Center

School of Physics and Astronomy (COS)


RIT – Main Campus