Bridging Scales in Black Hole Accretion and Feedback: Relativistic Jet linking the Horizon to the Host Galaxy (2507.17818v1)
Abstract: Simulating black hole (BH) accretion and feedback from the horizon to galactic scales is extremely challenging, as it involves a vast range of scales. Recently, our multizone method has successfully achieved global dynamical steady-states of hot accretion flows in three-dimensional general relativistic magnetohydrodynamic (GRMHD) simulations by tracking the bidirectional interaction between a non-spinning BH and its host galaxy. In this paper, we present technical improvements to the method and apply it to spin $a_=0.9$ BHs, which power relativistic jets. We first test the new multizone set-up with a smaller Bondi radius, $R_B\approx400\,r_g$, where $r_g$ is the gravitational radius. The strongly magnetized accretion launches a relativistic jet with an intermediate feedback efficiency $\eta\sim30\,\%$, in between that of a prograde ($\eta\sim100\,\%$) and retrograde ($\eta\sim 10\,\%$) torus. Interestingly, both prograde and retrograde simulations also eventually converge to the same intermediate efficiency when evolved long enough, as accumulated magnetic fields remove gas rotation. We then extend strongly magnetized simulations to larger Bondi radii, $R_B\approx 2\times103,~2\times 104,~2\times 105\,r_g$. We find that the BH accretion rate $\dot{M}$ is suppressed with respect to the Bondi rate $\dot{M}B$ as $\dot{M}/\dot{M}_B\propto R_B{-1/2}$. However, despite some variability, the time-averaged feedback efficiency is $\eta\sim30\,\%$, independent of $R_B$. This suggests that BH feedback efficiency in hot accretion flows is mainly governed by the BH spin ($a$) rather than by the galactic properties ($R_B$). From these first-principles simulations, we provide a feedback subgrid prescription for cosmological simulations: $\dot{E}{\rm fb}=2\times10{-3}[R_B/(2\times105\,r_g)]{-1/2}\dot{M}_Bc2$ for BH spin $a*=0.9$.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Collections
Sign up for free to add this paper to one or more collections.