The Physics, Observational Signatures, and Consequences of Galactic Winds Driven by Active Galactic Nuclei

Observations of AGN feedback have been revolutionized in recent years by the direct detection of galaxy-scale outflows driven by supermassive black holes. These AGN-driven galactic winds are now observed in several observational windows, including emission and absorption in the UV, optical, X-ray, radio, and far IR. The outflows are now routinely observed in several molecules, indicating that even the cold and dense gas that can form stars is directly affected by AGN feedback. However, the basic physics of these outflows and how to interpret their observational signatures remain poorly understood. Building on analytic models that we recently developed, we will use two complementary types of numerical simulations to model the physics and observational signatures of AGN-driven galactic winds. First, we will use controlled numerical experiments in simplified geometry (but fully 3D) to develop a robust understanding of different physical processes. For the first time, we will simulate the full time-dependent chemistry at wind shocks, enabling us to determine how molecules (including H_2, CO, and OH) can survive and/or form in AGN-driven outflows. Our simulations will follow the hydrodynamic and gravitational instabilities of gas swept up in the outflows and allow us to study whether stars can form in AGN outflows. We will study the effects of thermal conduction and magnetic fields on AGN wind dynamics by including them in a subset of our simulations. In stellar wind bubbles, thermal conduction mixes swept-up cool gas with the hot shocked wind and determines the evolution of wind bubbles. However this effect has so far been neglected in models of AGN-driven galactic winds, potentially affecting the momentum of galaxy-scale outflows at the order-of-magnitude level. We will also predict the X-ray emission from AGN winds with and without conduction, and how the X-ray morphology relates to the molecular emission. Second, we will use simulations of realistic galaxies with a multiphase interstellar medium shaped by stellar feedback to make detailed observational predictions. In these galaxy simulations, a nuclear wind modeled after observed accretion disk winds will interact self-consistently with stellar feedback and enable us to quantify how the mass and energy of AGN-driven outflows is partitioned between different phases and observational tracers. These realistic galaxy simulations will include the same physical processes as our controlled numerical experiments. We will process our galaxy simulations with the Monte Carlo radiative transfer code RADMC-3D to produce emission line maps and spectra in the CO, OH, and [CII] cold gas tracers, which we will compare with observations from Herschel, ALMA, and other ground-based radio observatories. Our simulation suite will include galaxies of several types in which AGN play an important role, including quasars fueled by galaxy mergers, lower-luminosity Seyferts fueled by secular processes, and gas-rich high-redshift star-forming galaxies.