The formation of broad-line clouds in the accretion shocks of active galactic nuclei

Recent work on the gas dynamics in the Galactic center has improved our understanding of the accretion processes in galactic nuclei, particularly with regard to properties such as the specific angular momentum distribution, density, and temperature of the inflowing plasma. With the appropriate extrapolation of the physical conditions, this information can be valuable in trying to determine the origin of the broad-line region (BLR) in active galactic nuclei (AGNs). In this paper, we explore various scenarios for cloud formation based on the underlying principle that the source of plasma is ultimately that portion of the gas trapped by the central black hole from the interstellar medium. Based on what we know about the Galactic center, it is likely that in highly dynamic environments such as this, the supply of matter is due mostly to stellar winds from the central cluster. Winds accreting onto a central black hole are subjected to several disturbances capable of producing shocks, including a Bondi-Hoyle flow, stellar wind-wind collisions, and turbulence. Shocked gas is initially compressed and heated out of thermal equilibrium with the ambient radiation field; a cooling instability sets in as the gas is cooled via inverse-Compton and bremsstrahlung processes. If the cooling time is less than the dynamical flow time through the shock region, the gas may clump to form the clouds responsible for broad-line emission seen in many AGN spectra. Clouds produced by this process display the correct range of densities and velocity fields seen in broad emission lines. Very importantly, the cloud distribution agrees with the results of reverberation studies, in which it is seen that the central line peak (due to infalling gas at large radii) responds more slowly to continuum changes than the line wings, which originate in the faster moving, circularized clouds at smaller radii. Finally, we provide an example of fitting an observed line profile using the parameters specified by our model.