# Hydrodynamical simulations of feedback from AGN

Patrick Yates1, Stas Shabala1, Martin Krause1,2

1University of Tasmania
2University of Hertfordshire

Honourable mentions: Jonathon Rogers, Jesse Swan, Ross Turner, Katie Vandorou

• Hi, I'm Patrick Yates from the University of Tasmania, where I recently started a PhD with Stas Shabala.

## AGN Feedback

### Overview

• Radio mode feedback required to explain evolution of SFR density with redshift
• Cooling catastrophe
• Truncation of star-formation in order to reproduce bi-modal galaxy population

Shabala & Alexander 2009, Apj, 699, 525

• AGN feedback plays an important role in galaxy evolution
• In this talk I will be focussing on radio-mode feedback, aka jet feedback
• Plot shows model of the evolution of star formation rate density with redshift, black is without feedback, red is with. Feedback needed for model to agree with observations
• Feedback balances with cooling flows and prevents the cooling catastrophe
• Feedback can be responsible for truncating star formation, which explains the bi-modality of the galaxy population. Can explain "red and dead" galaxies, where star formation has ceased

## AGN Feedback

### Large-scale feedback

• Jet inflates bubbles outside the galactic disk
• Bubbles do work on hot gas halo, through uplifting gas and shock heating

Fabian et al. 2000, MNRAS, 318 (4), 65

• I will briefly mention two areas where AGN feedback occurs. The first is large-scale feedback.
• This is feedback that essentially affects everything outside the galactic disc
• Feedback over a large range of scales, extending out to hundreds or thousands of kiloparsecs
• As the jet propagates through the environment it shock-heats the gas and inflates bubbles outside the galactic disc
• These bubbles do work by uplifting gas
• On the left is a plot showing the bubbles in the Perseus cluster. Radio contours overlayed on top of X-ray data
• On the right is a movie of density plot from a hydrodynamic sim of a jet, showing the jet shock-heating the environment and inflating a bubble

## AGN Feedback

### Galaxy-scale feedback

• Jets drive supersonic turbulence, drive outflows, can trigger star formation through compression due to shocks and heating
• This is feedback in the HI region

Mukherjee et al. 2016, MNRAS, 461 (1), 967

• The other interesting area where radio-mode feedback occurs is in the interstellar medium
• In this area the jets do a bunch of interesting things, including heating the ISM through driving supersonic turbulence and outflows
• It is possible that jets in the ISM can also trigger star formation through compression due to shocks and heating
• Plots density slices of hydrodynamic simulations of a jet in the ISM
• Different panels correspond to different jet powers and initial ISM turbulence values - see the paper for more information!
• Can see how the jet drives outflows and turbulence in all four cases

# How does the large-scale environment affect feedback?

• For the rest of this talk I will focus on large-scale feedback, and specifically how the large-scale environment plays a role in this

## We can use numerical simulations to find out!

• Can gain insight into how the physics of radio jets are impacted by:
• Environment
• Duty cycle and outburst count
• Can compare with:
• Analytical models
• Observations
• Simulations carried out using the PLUTO hydrodynamic code

## Quantifying Environment

Gas density for two halo masses

• Use the NFW dark matter density profile
• Gas density profile found by assuming hydrostatic equilibrium (Makino et al. 1998)
• Assume system is isothermal

$$\rho_\mathrm{DM}(r) = \frac{\delta_\mathrm{c} \rho_{\mathrm{c}0} }{(r/r_\mathrm{s}) \left( 1 + \frac{r}{r_\mathrm{s}} \right)^2 }$$ $$\rho_\mathrm{g}(r) = \rho_{\mathrm{g}0} \exp \left( - \frac{27b}{2} \right) \left( 1 + \frac{r}{r_\mathrm{s}} \right)^{27b / (2r / r_\mathrm{s})}$$

## Injecting the jet

• Gas is in hydrostatic equilibrium
• Jet injected as a mass inflow boundary condition at $r \sim 1$ kpc
• Non-relativistic jet

Gas density profile showing log of the gas density

## Different Environments

### Density and Morphology

• Power: $10^{37}$ Watt
• Mach Number: 25
• Active Time: 40 Myr
• Total Sim Time: 200 Myr

Cluster

Poor group

## Different Environments

### Energy Components

• Environment affects how the injected energy is distributed
• Greater kinetic energy in the $10^{12.5} M_\odot$ mass halo

Cluster

Poor group

## Different Environments

### Feedback Efficiency

• Quantified as fraction of injected energy that couples to ambient gas
• ~80% for $10^{12.5} M_\odot$ mass halo
• ~50% for $10^{14.5} M_\odot$ mass halo

## Different Environments

### Observations

Cluster

Poor group

• Can "paint on" radio emission as synchrotron emission using pressure
• Can construct size-luminosity diagrams

## Different Outburst Count

### Density and Morphology

• Left: 1 outburst for 40 Myr
• Right: 4 outbursts for 10 Myr
• Both: Same energy injected

4 Outbursts

1 Outburst

## Different Outburst Count

### Energy Components

• Outburst count does not significantly affect how injected energy is distributed

4 Outbursts

1 Outburst

## Different Outburst Count

### Feedback Efficiency

• Feedback efficiency is not greatly affected by outburst count in rich environment

Cluster

## Different Outburst Count

### Feedback Efficiency

• Feedback efficiency is greatly affected by outburst count in poor environment

Poor group

## Different Outburst Count

### Preconditioning

• Morphological differences show up clearly in plot of outburst tracers
• Explains some of the variation in feedback efficiency

Poor group

Cluster

• These plots show the jet material corresponding to outbursts 1 through 4 at the end of each outburst, for the poor group and cluster environments
• Clearly see how the preconditioning of the environment affects subsequent outbursts
• This likely explains some of the variation in feedback efficiency with outburst count

# How many compact sources have invisible lobes?

## Invisible Lobes

• Radio AGN which appear compact with FIRST / NVSS may have large "invisible lobes"
• Invisible lobes are present in simulated radio emission plots
• Multi-frequency surveys such as GLASS can provide further insight into invisible lobes

1 Outburst

4 Outbursts

• Simulations I have carried out support the existence of invisible lobes
• This is where large-scale structures in the cluster are below detection limit and so are not visible in radio observations
• Example of this is shown in plots on the right - the first is a single outburst jet, the second is a four outburst jet, both 200 Myrs old in a poor group
• Can see that surface brightness is well below 1 mJy for both plots in most regions of the jet, however there are these large-scale structures that are not visible in the radio emission
• More insight into these invisible lobes will be provided with multi-frequency surveys such as GLASS, as well as semi-analytic dynamical models
• Glass: 4cm wide, deep (?)

## Dynamical Models

• Jet simulations are computationally intensive
• New semi-analytic model developed by Ross Turner can help with this
• Semi-analytic dynamical model for expansion and radio emission + a numerical model for backflow
• Get electron aging, spectral break frequencies and more!

Turner, Shabala+ in prep.

• Full hydrodynamic simulations of jets are computationally intensive
• Can combine semi-analytic dynamical models with numerical simulations
• Use the numerical simulation for the backflow
• Combine with semi-analytic model for expansion and radio emission
• Can accurately predict electron aging, spectral break frequencies
• Can also quantify fraction of radio galaxies for a given frequency and sensitivity that have large-scale structures that go undedetected - invisible lobes

## Summary

• The environment into which the jet is injected affects the jet morphology, dynamics and injected energy distribution
• Preconditioning of the environment in earlier outbursts affects the jet dynamics and energy of subsequent outbursts
• The feedback efficiency of the jet is greater in poor environments compared to rich environments ($\sim 80 \%$ vs $\sim 50 \%$)
• Size vs luminosity tracks are very different between environments
• Invisible lobes are produced for simulations in poor environments