Cooling flows and quasars. II. Detailed models of feedback-modulated accretion flows

Most elliptical galaxies contain central black holes (BHs), and most also contain significant amounts of hot gas capable of accreting on to the central BH as a result of having cooling times that are short compared to the Hubble time. Why, therefore, do we not see active galactic nuclei (AGNs) at the center of most elliptical galaxies rather than in only (at most) a few percent of them? We here propose the simple idea that feedback from accretion events heats the ambient gas, retarding subsequent infall, in a follow-up to papers by Binney & Tabor and Ciotti & Ostriker. Even small amounts of accretion on a central BH can cause the release of enough energy to reverse the central inflow, when the Compton temperature (TX) of the emitted radiation is higher than the mean galactic gas temperature, the basic assumption of this paper. Well-observed nearby AGNs (3C 273, 3C 279), having TX near 5 × 108 K, amply satisfy this requirement. In this context, we present a new class of one-dimensional hydrodynamical evolutionary sequences for the gas flows in elliptical galaxies with massive central BHs. The model galaxies are constrained to lie on the fundamental plane of elliptical galaxies, and are surrounded by variable amounts of dark matter. Two source terms operate: mass loss from evolving stars, and a secularly declining heating by Type Ia supernovae (SNe Ia). Like the previous models investigated by Ciotti et al., these new models can evolve up to three consecutive evolutionary stages: the wind, outflow, and inflow phases. At this point the presence of the BH dramatically alters the subsequent evolution, because of the energy emitted by the accreting gas flow. The effects of Compton heating and cooling, hydrogen and helium photoionization heating, and bremsstrahlung recycling on the gas flow are investigated by numerical integration of the nonstationary equations of hydrodynamics, in the simplifying assumption of spherical symmetry, and for various values of the accretion efficiency and supernova rates. The resulting evolution is characterized by strong oscillations, in which very fast and energetic bursts from the BH are followed by longer periods during which the X-ray galaxy emission comes from the coronal gas. For a fixed galaxy total mass and structure, the length and the intensity of the bursts depend sensitively on the accretion efficiency and the SN Ia rate. For high efficiency and SN Ia rate values, short and strong bursts are followed by a degassing of the galaxy, with a consequent shut-off of the BH followed by a long period when the mass loss from the stellar population replenishes the galaxy, and after which a new cooling catastrophe another accretion event takes place. In this case, high accretion rates characterize the BH evolution, but the total mass accreted by the BH is very small. For low efficiency and SN Ia rate values, the luminosity evolution is still characterized by strong intermittencies, but the number of global degassing events is considerably reduced, and for very low efficiency values it completely disappears. The remaining instability is then concentrated in the central galactic regions. We also allow for departures from spherical symmetry by examining scenarios in which the central engine is either an advection-dominated accretion flow or a more conventional accretion disk that is optically thick except for a polar region. The general property of highly unstable accretion remains true, with central BHs growing episodically to the mass range 108-109 M ⊙ (in contrast to ΔMBH ≳ 1010-1011 M ⊙, if feedback is ignored). In all cases the duty cycle (fraction of the time that the system will be seen as an AGN) is quite small and in the range fBH ≃ 10-4-2 x 10-3. Thus, for any reasonable value of the efficiency, the presence of a massive BH at the center of a galaxy seems to be incompatible with the presence of a long-lived cooling flow.