2005MNRAS.363L..91C

Relativistic outflows (mainly observed in the radio) are a characteristic feature of both Galactic stellar-mass black holes and supermassive black holes (SMBHs). Simultaneous radio and X-ray observations of Galactic sources have shown that the outflow is strong at low accretion rates, but it weakens dramatically or disappears completely at high accretion rates, manifesting structural changes in the accretion flow. It is reasonable to assume that SMBHs follow the same trend. For low-luminosity SMBHs in nearby elliptical galaxies and clusters, recent observations strongly suggest that the outflows play the central role in keeping the gas hot (mechanical feedback). If the outflow is quenched in SMBHs at high accretion rates similarly to the behaviour of Galactic sources, then the straightforward consequence is a relatively weak feedback of rapidly accreting SMBHs. We argue that elliptical galaxies and their central engines should then evolve through two stages. Early on, the central SMBH rapidly grows by accreting cooling gas at a near-Eddington rate with high radiative efficiency but with weak feedback on the infalling gas. This stage terminates when the black hole has grown to a sufficiently large mass that its feedback (radiative and/or mechanical), despite the low gas heating efficiency, is able to suppress gas cooling. After that the system switches to a stable state corresponding to passively evolving ellipticals, when the accretion rate and radiative efficiency are very low, but the gas heating efficiency is high and energy input from the relativistic outflow keeps the gas hot.