Accretion Driven Evolution of Quasars and Black Holes: Theoretical Models

Abstract

We present a flexible framework for constructing physical models of quasar evolution that can incorporate a variety of observational constraints, such as multi-wavelength luminosity functions, estimated masses and accretion rates of active black holes, space densities of quasar hosts, and the local black hole mass function. We focus on the accretion rate distribution p(mdot|M,z), the probability that a black hole of mass M at redshift z accretes at a rate mdot in Eddington units. Given the radiative efficiency as a function of mdot, the quasar luminosity function (QLF) is determined by a convolution of p(mdot|M,z) with the black hole mass function n(M,z). In the absence of mergers p(mdot|M,z) also determines the full evolution of n(M,z), given a "boundary value" of n(M) at some redshift. Matching the observed decline of the QLF break luminosity at z<2 requires either a shift towards lower characteristic accretion rates or an evolving mass dependence of p(mdot) that preferentially shuts off accretion onto high mass black holes. These two scenarios make different predictions for the masses and accretion rates of active black holes. If the first mechanism dominates, then the QLF changes character between z=2 and z=0, shifting from a sequence of black hole mass towards a sequence of L/Ledd. We construct and compare five models that illustrate different assumptions about the quasar population: short and long lifetime models dominated by unobscured thin-disk accretion, a model with a high fraction of obscured quasars, and models in which mergers or ADAF accretion produce substantial black hole growth at low redshift. We discuss the observational advances that would be most valuable for distinguishing such models and for pinning down the physics that drives black hole and quasar evolution. (Abridged)