AGN outflows in different hosts: when does it eject or compress gas?

Observed AGN outflows correlate with the luminosity of the driving AGN, but show order-of-magnitude scatter around the relation. One source of the scatter is the diversity of AGN host galaxy properties: identical AGN, powered by identical SMBHs, can reside in galaxies with significantly different masses, effective radii, stellar velocity dispersions and gas fractions. We ran a suite of 50 purpose-built SPH simulations of AGN outflows propagating in spherical gaseous bulges with realistic clumpy interstellar medium properties. The simulations covered a range of SMBH masses between 10^7-3×10^8 Msun, with velocity dispersions and bulge masses spanning the observed scatter in the M-sigma and MBH-Mbulge relations. We tested the evolution of outflows and their impact on the host galaxy gas when driven by constant-luminosity AGN with Eddington ratios of 0.1-1. I will present results showing how galaxies with deeper gravitational potentials and higher gas fractions efficiently confine outflows, leading to more efficient cooling and reduced outflow energy. This produces distinct outcomes: gas compression and enhanced star formation in compact galaxies, versus gas ejection and quenching in diffuse ones. I will discuss how IFU observations of host galaxy structure and outflow kinematics can test these predictions and clarify the role of AGN feedback on star formation over multi-Myr timescales.