A hybrid active galactic nucleus feedback model with spinning black holes, winds and jets

Abstract

We present a hybrid active galactic nucleus (AGN) feedback model that features three accretion disc states (the thick, thin, and slim discs at low, moderate, and super-Eddington accretion rates, respectively), and two feedback modes: thermal isotropic and kinetic jets. The model includes black hole (BH) spin evolution due to gas accretion, BH mergers, jet spindown, and Lense-Thirring torques. The BH spin determines the jet directions and affects the feedback efficiencies. The model is implemented in the SWIFT code and coupled with the COLIBRE galaxy formation model. We present the first results from hybrid AGN feedback simulations run as part of the COLIBRE suite, focusing on the impact of new parameters and calibration efforts. Using the new hybrid AGN feedback model, we find that AGN feedback affects not just massive galaxies, but all galaxies down to $M_*\approx10^8$ $\mathrm{M}_\odot$. BH spins are predicted to be near-maximal for intermediate-mass BHs ($M_\mathrm{BH}\in[10^6,10^8]$ $\mathrm{M}_\odot$), and lower for other BH masses. These trends are in good agreement with observations. The intergalactic medium is hotter and impacted on larger scales in the hybrid AGN feedback simulations compared to those using purely thermal feedback. In the hybrid AGN simulations, we predict that half of the cumulative injected AGN energy is in thermal and the other half in jet form, broadly independent of BH mass and redshift. Jet feedback is important at all redshifts and dominates over thermal feedback at $z<0.5$ and $z>1.5$, but only mildly.

Figures (17)

• Figure 1: Visualization of jet activity in the L100m6h simulation at $z=0.2$ in proper coordinates. The background image shows the cosmic web using the gas surface density, while particles kicked into jets by black holes are displayed using both information on the elapsed time since when they were kicked into jets (colour) and their surface density (opacity). Side panels zoom in on individual jets of interest, with further zoom-ins onto their host galaxies (except in the case of a cluster and its BCG) displayed using luminosities in the SDSS i, r and g bands assigned to RGB colours.
• Figure 2: The three different accretion disc states assumed in our model, corresponding to different Eddington ratio regimes, from bottom to top displaying the thick, thin and slim discs, with the thin disc component present at large radii. The accretion disc state changes depending on the disc-scale Eddington ratio, $f_\mathrm{Edd,d}$, defined in § \ref{['sec:accretion_states']}, at two critical values of $0.01$ and $1$. The left and right illustrations show edge-on and inclined views of the accretion disc, respectively. Radiation and winds are assumed to originate on accretion disc scales, and the effects of both are implemented using thermal isotropic feedback in the simulations. Jets, which consist purely of Poynting flux on these accretion disc scales, are implemented as bipolar velocity kicks.
• Figure 3: The fraction of BHs occupying different accretion disc states in the L200m7h simulation: left: thick disc, middle: thin disc, right: slim disc, multiplied by a factor of 10 for visual comparison purposes. Different colours indicate different redshifts. The shaded regions show uncertainties computed assuming binomial proportion confidence intervals.
• Figure 4: The assumed accretion efficiencies as a function of the disc-scale Eddington ratio. The accretion efficiency connects the accretion rate onto a subgrid accretion disc (Eqn. \ref{['eq:net_accr_rate']}) with the accretion rate into the BH itself, and it represents the effects of disc winds that take away most of the mass. In the thin disc we assume a $100$ per cent accretion efficiency, while in the slim disc we assume 1 per cent. In the thick disc, we combine results of GRMHD simulations and analytical calculations (Eqns. \ref{['eq:acc_eff']} and \ref{['eq:R_th']}). For the free parameter $r_\mathrm{tr,0}$, we take the value $10^4$, corresponding to the value we found through calibration on the bolometric AGN luminosity function (see § \ref{['sec:calibration_first_results']}). The accretion efficiency in the thick disc is limited by the ratio of the Bondi radius and the gravitational radius of the BH, which depends only on the gas sound speed. Using different line styles, we show the accretion efficiency for several values of the sound speed, ranging from values appropriate for the hot ICM in galaxy clusters to those appropriate for cold, dense gas in the ISM.
• Figure 5: The dependence of feedback efficiencies used in the hybrid AGN feedback model on the accretion disc-scale Eddington ratio $f_\mathrm{Edd,d}$ for the three accretion and feedback states. Solid lines show the jet efficiencies (see § \ref{['sec:jet_efficiencies']}), and dashed the thermal efficiencies (see § \ref{['sec:thermal_eff']}), assuming a coupling efficiency $\epsilon_\mathrm{f}=0.025$. We show these for several values of BH spin using different colours (see legend); note that the jet efficiency is $\epsilon_\mathrm{jet}=0$ for $a=0$. In the left panel we show the pure feedback efficiencies, while in the right panel, we further multiply these by accretion efficiencies (see § \ref{['sec:accretion_eff']}) to emphasize the net feedback efficiencies and to facilitate comparison between the different modes. For this figure, we used $r_\mathrm{tr,0}=10^4$ as found through calibration (§ \ref{['sec:calibration_results']}).