Paper detail

BAL QSOs and Extreme UFOs: the Eddington connection

We suggest a common physical origin connecting the fast, highly ionized winds (UFOs) seen in nearby AGN, and the slower and less ionized winds of BAL QSOs. The primary difference is the mass loss rate in the wind, which is ultimately determined by the rate at which mass is fed towards the central supermassive black hole (SMBH) on large scales. This is below the Eddington accretion rate in most UFOs, and slightly super-Eddington in extreme UFOs such as PG1211+143, but ranges up to $\sim 10-50$ times this in BAL QSOs. For UFOs this implies black hole accretion rates and wind mass loss rates which are at most comparable to Eddington, giving fast, highly-ionized winds. In contrast BAL QSO black holes have mildly super-Eddington accretion rates, and drive winds whose mass loss rates are significantly super-Eddington, and so are slower and less ionized. This picture correctly predicts the velocities and ionization states of the observed winds, including the recently-discovered one in SDSS J1106+1939. We suggest that luminous AGN may evolve through a sequence from BAL QSO through LoBAL to UFO-producing Seyfert or quasar as their Eddington factors drop during the decay of a bright accretion event. LoBALs correspond to a short-lived stage in which the AGN radiation pressure largely evacuates the ionization cone, but before the large-scale accretion rate has dropped to the Eddington value. We show that sub-Eddington wind rates would produce an $M - σ$ relation lying above that observed. We conclude that significant SMBH mass growth must occur in super-Eddington phases, either as BAL QSOs, extreme UFOs, or obscured from direct observation.