04-07 December 2023 -- Flatiron Institute
- Accretion theory does well in XRBs but not in the AGN regime.
- Environment -- star clusters, objects in disk
- Mass supply -- ???
$\rightarrow$ see talks later today - Accretion -- radiation pressure is important
- Opacities -- e- scattering dominates in XRBs but atomic opacities are much more important in AGN
- Open questions
- How does feeding and environment impact accretion disks?
- How are B-field brought in and what determines their geometry?
- Radiation might drive winds at
$v_{\rm out} \sim v_{\rm in}$ . What are the implications for AGN feedback?
Multi-scale cosmological simulations of quasar fueling and feedback
- (Angles-Alcazar+ 2021) Hyper-refinement simulations.
- Strong variability 0.001-10 Msun/yr within 1 Myr. At late times, there's a low AGN duty cycle set by ISM physics.
- Grav. torques from stars dominate AM transport at all scales over gas self-torquing and pressure gradients.
- A large fraction of inflow rate goes into SF on all scales.
- Accretion correlates with nuclear SFR but time variability decouples ANG-global SFR
- SMBH-galaxy scaling driven by common gas supply?
- Gas AM: time-dependent misalignment of gas
- (Hopkins+ 2023) FIRE-3 + STARFORGE.
- Within the accretion disk (<0.01 pc), there's a
$m=1$ disk mode induced by larger-scale asymmetry. - Highly supersonic turbulence (t_cool < t_dyn)
- Strongly magnetized -- flux-freezing from ISM (
$\beta \ll 1$ ).
- Within the accretion disk (<0.01 pc), there's a
Bridging the Gap: Modeling Supermassive Black Holes Feeding and Feedback at the Meso-Scale
- Performed HD simulations with Athena++ of M87-like galaxy with jet feedback.
- Probed turbulent kinematics of the two-phase gas down to 250M.
- Next steps: include MHD and expand to different galaxy types and compare with observations.
Simulating Multi-Waveband Quasar Light Curves Using Radiation Magnetohydrodynamic Simulations
- (Secunda+ 2023) Cooling can cause an entropy inversion, which can drive instabilities.
- In their RHD simulations, they split their radiation into UV and X-ray groups.
- X-ray scattering opacity is dominant over absorption.
- Correlation (
$r=0.54$ ) between intrinsic and reprocessed UV, but little correlation between UV and X-ray.
Toward Horizon-scale Accretion onto Supermassive Black Holes in Elliptical Galaxies
- (Guo+ 2023; also see Xu 2023) Toy model: Turbulent Bondi accretion. Never forms a disk. They inspect density profiles and inflow rates.
- Two stages: chaotic stage and then disk stage. They swap between the two during their simulations. AM changes frequencies.
- Radial profile is proportional to 1/r instead of
$r^{-3/2}$ in the inner regions. The CDAF (intermediate radii) is proportional to$r^{-1/2}$ .
Black Hole Accretion Within The Bondi Radius
- Accretion model: infalling gas will circularize according to their AM. A fraction of the gas mass is accreted within this variable accretion radius (mass-weighted average).
- Subgrid model: Have a gas reservoir that accretes onto the disk on a free-fall timestep. The AGN then accretes on a viscous timescale. They assume
$\alpha = 0.1$ . - Includes primordial gas cooling, dust self-shielding (with TreeCol; Clark+ 2016?)
- Future work: cosmological simulations, radiative transfer, wind feedback
The origin of gas that feeds supermassive blackholes in cosmological simulations
- The external gas supply is the most important for SF providing about 40% of the reservoir, much lower than the recycled gas within the galaxy at 20%.
- They predict super-solar metallicities at
$z=2-3$ because of the recycled gas. - There's a large spread in the lag between the merger and AGN phase with 3 +/- 1 Gyr.
Observability of Low-Luminosity AGN in the Early Universe with JWST
- (Jeon+ 2023) SMBH / AGN model in AREPO first galaxy simulations (4 cMpc box).
- They find that Stellar feedback suppresses SMBH accretion.
- They seeded the SMBHs with light seeds (300 Msun) and wanders around the galactic potential.
- They generally find accretion rates varying around 0.01
$\dot{M}_{\rm Edd}$ .
The growth and impact of the first gargantuan black holes - insights from cosmological simulations
- (Bennett+ 2023) Using the FABLE model for galaxy formation (Henden+ 2018), focusing on a very massive halo with
$M = 7 \times 10^{12} M_\odot$ at$z=6$ . - Modified the SMBH model to encourage SMBH growth -- lower rad. efficiencies, earlier seeding, higher accretion rate limit (
$2\dot{M}_{\rm Edd}$ ). - Generally reproduces the
$z \sim 6$ QSOs, GN-z11, and the Larson+ (2023)$z \sim 8$ AGN. - The SMBH accretion rate varies between 0.2-2 of Eddington.
- Without AGN, the column densities are Compton-thick. But with feedback, the column density is reduced by a factor of ~100 down to
$\sim 10^{22} cm^{-2}$ . - X-ray emission from hot halo gas is hard to disentangle from AGN emission. With Athena, it is nearly impossible, but another planned mission (AXIS) may be able to distinguish them. However Lynx (if funded) can definitely do this.
The role of super-critical accretion in the early growth of high-z AGNs observed with JWST
- Using CAT (Trinca+ 2022), a SAM for high-z AGN/galaxies. Uses a "detailed seeding prescription"
- Compares a Bondi, Eddington-limited accretion model and a super-Eddington model. The latter allows light seeds to contribute the AGN population.
- (Schneider+ 2023) The different evolutionary history affects the SED and observational prospects. However, need simulations to understand the exact differences.
- (Sassano+ 2023) Can grow rapidly within the first 1 Myr -- seed BH of
$10^3 M_\odot$ grows$10^4 M_\odot$ -- but quenches shortly afterwards.
A new theory for hot accretion flows within the Bondi radius
- Aiming to explain the density slope within the Bondi radius, especially the
$\alpha = -1$ regime. - (Xu 2023) Even when they inject 5% perturbations in these most simple model, self-similar large-scale turbulence develops.
- Converges to a self-state solution with
$\alpha = -0.8$ that has an infall velocity of 1/4 Keplerian with the rest in tangential velocities. - The main contribution to the turbulence is the gas buoyancy.
- Why
$\alpha = -0.8$ ? There is strong convection (mixing length$\xi \sim r$ ). The slope comes from momentum balnace + approximate up-down symmetry --$\alpha = -1$ + small number. - This argument of mean vs. typical acceleration still holds when more physics is added.
- When converting the density slope to an accretion rate (and luminosity), this matches observations well.
Rapid talk: Exploring the long-term effects of short-lived high accretion rate black hole feedback on galaxies
- Build growth histories within UniverseMachine, using observations as a baseline -- MBH-Mstar relation, SFR-Mstar relations at
$z=0$ . - Easy to grow less massive SMBH in star forming galaxies, but harder to grow more massive SMBHs, as the gas supply has been quenched.
Rapid talk: Shaping Massive Galaxies at Cosmic Noon: Probing AGN Feedback in the FIRE Simulations
- (Mercedes+ 2023ab) Find some positive AGN feedback on small-scales, generating higher SFEs and SFR densities within outflowing material.
- Outflows compress nearby clumps to collapse and form stars.
Rapid talk: Black hole mass correlation with a galaxies stellar mass, bulge mass, and velocity dispersion
- Attempts to conduct a symbolic regression model on TNG simulations. Works well on 100 and 300 Mpc boxes but not 50 Mpc.
- Uses a random forest model. SVM models give similar results.
- Uses Vmax, Zstar, SFR, Mhalo (these 4 features are the most important), Mstar, Rstar, Mgas, VmasRad, Zgas, Rgas, ...
AGN Feedback in the Low-z Ly$\alpha$ Forest
- CAMELS-Simba implies that the low-z forest is a candidate for future constraints on galactic feedback models.
AGN Feedback Review
- (Costa+ 2020) BOLA: Creates a spherical boundary in AREPO with HEALPix surface pixels in which they inject feedback
- Various forms of feedback: thermal, jets, radiation, RP on dust, physical models that include the different phases (exactly like the SN remnants)
- (Talbot+ 2021; Wagner+ 2016; Tanner & Weaver 2022) Simulations of jet feedback. If they
- (Bahe+ 2022; Chen+ 2022) BH growth can be artifically suppressed in simulations by wandering due to unresolved dynamical friction.
- (Massonneau+ 2023) Super-Eddington accretion but it doesn't result in substantial growth because of the strong feedback.
On the Connection between AGN radiative feedback and Massive Black Hole Spin
- AGN feedback impact is determined by its power, not angular pattern that is more relevant for highly accreting MBHs in thick/spherical galaxies.
- High spin MBHs can easily suppress SF and hinder MBH growth.
- Binary migration: feedback can clear a cavity around the BBH, leading to bursty evolution. Migration is suppressed by feedback because the lack of gas.
Efficient Suppression of Cooling Flows by Cosmic Rays in Jet Cocoon Shock Fronts
- Including CR injection in jets. They explore different injection sites: at the BH or at the shock front.
- When they inject CRs at the shock front, the feedback is much more sustained with longer duty cycles (~Gyr), which can lead to galaxy quenching.
- Jet precession changes the shock front locations and thus CR distribution. They find that the "sweet spot" for quenching is ~30 Myr.
Recent progress in simulating supermassive black hole accretion, spin evolution, jets and winds
- (Bourne+ 2023) SMBH binaries in circumbinary disks with small-scale AREPO simulations, exploring different parameters. They find some orbital hardening and shrinkage.
- They also explore the alignment of the BH spins and disk.
- Preferential accretion onto the primary BH.
Bridging Scales between Supermassive Black Holes and Their Host Galaxies via Multi-annuli Approach
- (Cho+ 2023) Running simulations with different annuli, ranging from the horizon to 10 kpc. The annuli influence each other by transfering flux through their boundaries.
- Tested their new technique in Bondi case, both with
$B=0$ and strongly magnetized. - Most of their feedback is driven by convection when reconnection occurs.
Event Horizon-scale Insights into Accretion, Feedback, and Spin Evolution
- (Narayan+ 2022) Powering a jet will spin down the BH
- (Ricarte+ 2023) Ran 38 GRRMHD simulations, varying masses, spins, and Eddington ratios. They fit spin evolution models for their suite that can be used in subgrid models.
- Also calculated the equilibrium spin as a function of Eddington ratio. Most massive SMBHs may have spun down.
- Spin down may help reduce the radiative efficiencies of super-Eddington accretors at high-z. The accretion duty cycle matters.
High Resolution Exascale Simulations of Magnetized AGN Feedback in Galaxy Clusters and Groups
- Using AthenaPK to simulate AGN feedback in idealized simulations. Within ~3 kpc, plasma
$\beta \le 1$ . - Through an energy transfer analysis (see below), they find that kinetic energy transfers to magnetic energy through magnetic field tension.
Problematic flexibility: How our modeling choices for "live" BH accretion and feedback affect growth and quenching in FIRE galaxies
- Posed question: What are the physical properties of the models that reproduce the observed galaxy properties (Mstar-Mhalo, M-sigma, SFR-Mstar)?
- (Wellons+ 2023) They ran a simulation suite (~500 runs) that explored a parameter space with different accretion efficiencies, mechanical feedback, radiative feedback, CR feedback, and different implementations (pushing, spawning, collimating).
- BH growth depends on accretion and feedback efficiencies and this relationship is halo-dependent. They term this "responsiveness" through the product of these two efficiency parameters.
AGN Jet Feedback Simulation with GADGET3-Osaka
- AGORA simulations are being extended to include an AGN model in the isolated galaxy, exploring seeding, accretion, feedback, and then the full model.
- Their group is trying to reproduce and improve on the results of Husko & Lacey (2023), where they had a self-similar jet model and ballistic jet model.
The impact of radiatively efficient AGN-driven winds on galaxy and black hole growth
- Zoom-in simulations of selected TNG100 galaxies -- a quenched galaxy and a star forming galaxy at
$z=3.5$ . Quenching coincides with the triggering of the AGN. - Find little growth within 40 Myr of modeling.
- They need to better model AGN feedback in the radiatively efficient regime to study high-z galaxies.
BH Seeding Review
- (Beckmann+ 2021) NewHorizon seeding motivated by observations. One seed per halo, excluding formation if there's another BH within 50 kpc.
- (Bellovary+ 2011; Taylor+ 2014; Tremmel+ 2017; Chon+ 2020) Seeding based on local properties, e.g. metallicity, LW flux, H2 fraction.
- Initial seed masses chosen either by (i) observations, i.e. scaling relations, (ii) subgrid models depending on local properties, or (iii) selecting either light or heavy seeds.
- (Habouzit+ in prep) Comparing JWST discoveries with simulations
Modeling Super-Massive Black Holes in cosmological contexts from their origin as BH-seeds down to the local Universe
- (Spinoso+ 2023) Using L-Galaxies to understand how the seeds grow. They have four seeding channels: light, intermediate, heavy through mergers, heavy through DCBHs.
- (Izquierdo+ 2020, 2022) MBH evolution, including kicks. 2022 paper models binaries (pairing, hardening, and merger) with physically motivated delay between galaxy and MBH mergers.
- (Lops+ 2023) Uses their model to make PTA and LISA predictions.
An insight on Black Hole Growth from semi-analytic model L-Galaxies-BH: Advancements and Prospects
- Using tidal disruption events (TDEs) to seed / grow BHs in L-Galaxies
- (Hoyer+ 2021) Observed nuclear star cluster (NSC) occupation fraction at
$z=0$ - There are ~1,000x more TDEs in NSCs than bulges.
Black hole seeding in cosmological hydrodynamic simulations: A novel flexible approach
- Two seeding models: (i) gas based model in resolved halos, (ii) stochastic model in unresolved halos
- (Bhowmick+ 2021, 2022) Gas-based: dense and metal-poor gas and some minimum halo mass. Converged.
- (Bhowmick+ 2023) Stochastic model: galaxy mass criterion, galaxy enviornment criterion. Seed the smallest resolved higher mass descendants.
- (Bhowmick+ in prep) BRAHMA simulations -- uses this seeding model in a cosmological simulations -- 9, 18, and 36 Mpc boxes
UNCOVERing Red AGN Candidates with JWST and ALMA
- (Greene+ 2023) Finding SMBHs with masses around
$10^8 M_\odot$ at$z \sim 6$ in galaxies with$M_{UV} \simeq -20$ . - UV-selected QSO LF really flatten at magnitudes dimmer than -22. UNCOVER is finding AGN hosts with
$M_{UV} \simeq -18$ around$10^{-5}$ Mpc$^{-3}$ mag$^{-1}$. - SED is still uncertain. UV can be explained by scattered light or a SF component.
- (Williams+ 2023) The Mid-IR spectra flatten above 5 microns, which indicate that it's a reddened stellar population with age ~ 500 Myr, not an AGN.
- (Kokorev+ 2023) Finding another overly massive SMBH --
$M_{BH} = 10^8 M_\odot$ and$M_\star = 10^9 M_\odot$ . - From Q&A: BH mass estimates are uncertain because the true luminosity is unknown. Reddening is unknown. Line widths give you some estimate, but it's not the whole story.
Detecting active IMBHs at z~0 with novel approaches to photoionization simulations
- (Richardson+ 2022) Using CLOUDY to explore different geometries (spherical vs plane parallel) on the emission line properties.
- (Polimera+ 2022) [O I] 6300 line is a metallicity insensitive diagnostic for AGN selection in metal-poor dwarfs. Other lines: He II 4686.
- Mid-IR lines can also give some insight -- [Ne III] 15.56 microns / [Ne II] 12.x microns versus [O IV] 25.9 microns, and another pair that I didn't have time to write down.
- (Lebouteiller & Ramambason 2022) MULTIGRIS model. Given photo-ionization models and a suite of emission lines, it will return posterior PDFs of the physical quantities with Bayesian inference.
Testing models of supermassive black hole seeding and growth with observations of dwarf galaxies
- (Chadayammuri+ 2023) Explored light and heavy seeds when compared to broken power law AGN LF, power-law AGN LF, and main sequence AGN in X-ray luminosity versus stellar mass space.
- Observations are flux-limited. Observed AGN fractions depends on survey flux limit, which can look similar for different seeding models at the flux limit.
- They infer a SMBH occupation fraction from the observed AGN fraction -- log-linear increase between
$10^7$ and$3 \times 10^9 M_\odot$ . Strongly favors heavy seeds because light seeds would over-predict the AGN fraction.
The effect of AGN feedback in simulated dwarf galaxies
- (Arjona-Galvez+ 2023) Uses the AURIGA simulations to study the difference between 12 dwarf galaxy simulations with and without AGN model.
- Found that the SFR ratio between the simulations with and without AGN feedback is inversely correlated with the BH mass.
- The DM density is decreased by 65% within the central 1 kpc when AGN feedback is included.
Where to Seed Massive Black Holes
- Joe McCaffrey is looking into the seed BH masses within the Renaissance Simulations.
- Need to determine how much error is being introduced by skipping the BH history between actual formation and seeding within the simulation.
- From Q&A: Should compare with SAMs of BH seeding and growth.
Continual formation of intermediate mass black holes and its effect in SMBH formation and galaxy evolution
- (Stevens & Adam 2016) Dark Sage SAM that includes SMBH formation and growth.
- (Porras+ in prep) Generally cold-mode and instabilities contribute the most to BH growth at all redshifts.
Are we surprised to find SMBHs with JWST at z > 9?
- (Sassano+ 2022) Showed the variation of the BH mass function, depending on the seeding mechanism, i.e. LW flux or supercompetitive accretion model (Chon & Omukai 2020)
- (Schneider, Trinca+ 2023) Investigated the properties of the JWST-detectable
$z>5$ BHs, looking at both Eddington-limited and super-Eddington growth models.
The NANOGrav Experiment: Results and Future Directions
- Evidence for GW background in a common amplitude. Next up local nHz sources within a decade.
Recoiling and off-center supermassive black holes in galactic nuclei
- Core stalling and bouyancy occur in systems with a "shallow" or inflection in f(E)
- Driven by near-resonant effects between stars and BHs
- More careful modeling of dynamical friction in cosmo sims, getting insight from idealized sims.
3D radiation hydrodynamics simulations of wandering seed black holes
- 3D RHD simulations with anisotropic radiation and dust sublimation, using SFUMATO (M1 method). Improvements on Toyouchi+ (2020) and Ogata+ (2021).
- Inject photons within some solid angle. Radiation force is ineffective because of dust sublimation.
- Suite of idealized simulations with
$\log_{10} M_{BH}/M_\odot = 3-6$ and varying wind speeds and ambient densities. - Outflows can return back to the BH from the opposing side of the wind.
- Super-Eddington when
$M_{BH} n_B > 10^{9} M_\odot cm^{-3}$ . - The D-type front is asymmetric in the flow direction and can accelerate the BH.
- Future work: Within cosmological simulations (e.g. Ishiyama+ 2021) will boost the Bondi rate (i.e.
$\alpha$ boost factor) derived from these simulations.
Massive Black Hole Dynamics in and beyond Cosmological Simulations: Bridging the Gap between Galaxy Mergers and Gravitational Wave Emission
- The coarse DM discretization, relative to low-mass SMBHs, could cause some errors, even when a DF subgrid model is used.
- (Chen+ in prep) Resimulated 15 mergers within ASTRID by restarting before the merger and split the particles. Can trace the orbital decay to the hardening regime, aided by higher densities at smaller radii.
- The orbits circularize to an eccentricity of 0.8.
- To sink further in dynamic and complex environments, BHs need a nuclear star cluster above
$5 \times 10^6 M_\odot$ .
Black hole seed formation and dynamics in high-redshift galaxies
- Need to observe MBHs in low-mass galaxies.
- Pros: BH mass is closer to the seed mass, very abundant in the Universe
- Cons: harder to observe, may wander from the galaxy center
- Need to model DF correctly (e.g. Tremmel+ 2015), treat individual stars, primordial chemistry, and X-ray feedback from low-mass BHs.
- (Lupi+ in prep) Gizmo, KROME, M1 RT, Pop III -> II SF, stellar winds, BH accretion / radiation / winds / mergers
- Have run it to
$z=6$ . Some BHs are within the central 1 pc.$M_\star = 3 \times 10^3 M_\odot$ ,$M_{BH} = 2 \times 10^4 M_\odot$ - There are dozens of BHs with
$\sim 10^3 M_\odot$ . - (Lupi+ submitted and in prep b) Another simulation including super-Eddington accretion. The BH accretion / feedback have three phases (details?). The BHs grow the most in the MAD / HMAD phase. Only started to grow substantially after
$M_\star = 10^{10} M_\odot$ .
Massive Black Hole Binaries: Modeling Populations for Multi-Messenger Astronomy
- (Siwek+ 2023ab) Very large suite of circumbinary disk simulations, varying mass ratio, eccentricity, and separation.
- There are no MBHBs with low mass ratios. There's an upper limit to eccentricity of 0.6, which then hardens to
$e=0$ .
Modeling the coalescence of MBHs over the cosmic history
- (Bird+ 2022; Ni+ 2022) ASTRID simulation.
- The dual AGN fraction is around 0.01, and triple fraction is around
$10^{-4}$ - Tidal stripping produces an extensive population of wandering BHs
