Simulating AGN feedback in galaxy clusters with pre-existing turbulence

TL;DR

This study tests whether turbulence can balance radiative cooling in cool-core clusters by simulating a Perseus-like cluster with pre-existing turbulence and an impulsive AGN jet. Turbulent heating is quantified through velocity structure functions and the energy power spectrum, revealing that the turbulent dissipation rate in the core is smaller than the cooling rate, even with observationally constrained turbulence. The jet injects energy mainly as transient bulk motions and pressure perturbations, not as a sustained turbulent cascade; XRISM-observed central velocity dispersions can be reproduced by jet-driven motions but do not imply persistent turbulence. The results suggest turbulent heating alone cannot solve the cooling-flow problem, reinforcing the need for additional AGN-related processes (bubble mixing, weak shocks, sound waves) and other heating channels such as cosmic rays or conduction in a more complete model.

Abstract

Feedback from active galactic nuclei (AGN) is believed to play a significant role in suppressing cooling flows in cool-core (CC) clusters. Turbulence in the intracluster medium (ICM), which may be induced by AGN activity or pre-existing motions, has been proposed as a potential heating mechanism based on analysis of Chandra X-ray surface brightness fluctuations. However, subsequent simulation results have found the subdominant role of turbulence in heating the ICM. To investigate this discrepancy, we perform three-dimensional hydrodynamic simulations of a Perseus-like cluster including both AGN feedback and pre-existing turbulence, which is stirred to the observationally constrained level in the Perseus cluster. Our results indicate that, although the velocity field is dominated by the pre-existing turbulence, AGN heating through bubbles and shocks remains significant. More importantly, analysis of the velocity structure function and the energy power spectrum shows that the turbulent heating rate is smaller than the radiative cooling rate, especially in the cluster core. Our results offer insights relevant for recent XRISM observations and indicate that turbulent heating alone cannot offset radiative cooling in CC clusters.

Figures (11)

• Figure 1: Development of the pre-existing turbulence. Upper panel: Time evolution of the 3D root-mean-square (RMS) turbulent velocity, showing that the system with pre-existing turbulence reaches a steady state after $\sim 800~\mathrm{Myr}$. Lower panel: Time evolution of the kinetic energy spectrum, $E(k)$, shown from $t=200$ to $1400~\mathrm{Myr}$ with an interval of $200~\mathrm{Myr}$. The convergence of the spectra to a stable profile confirms that the turbulent cascade is fully established prior to the jet injection.
• Figure 2: Time evolution of the gas density slices at the x = 0 plane for JetOnly (top row) and Turb+Jet (bottom row). Columns correspond to snapshots taken at $t = 10$, 40, 70, 100, and 130 Myr after jet injection. Each slice is 200 kpc on a side. In Turb+Jet, the jets propagate faster and the cluster core is less centrally concentrated due to pre-existing turbulence, which also disrupts the bubbles more quickly compared to JetOnly.
• Figure 3: Columns from left to right correspond to TurbOnly, Turb+Jet, and JetOnly, shown at the same timestep, 50 Myr after the jet injection. The first row shows thin projections (4 kpc) of the velocity magnitude ($|{\boldsymbol v}|$) weighted by gas density. The second row shows the line-of-sight velocity dispersion ($\sigma_{\rm LOS}$) weighted by X-ray emissivity.Each panel spans 132 kpc on a side. Both $\lvert v \rvert$ and $\sigma_{\rm LOS}$ are dominated by pre-existing turbulence, indicating that the jet has a minor influence on the overall velocity field.
• Figure 4: Time evolution of $\sigma_{\rm LOS}$ averaged over the entire simulation domain for Turb+Jet and TurbOnly as a function of the time since jet injection. The overall $\sigma_{\rm LOS}$ is similar in both cases, indicating that the jet has a minor impact and that the velocity field is dominated by pre-existing turbulence.
• Figure 5: Entropy profiles as a function of radius for TurbOnly, Turb+Jet, and JetOnly, shown at the same timesteps after jet injection. Different lines correspond to different timesteps. The entropy increase is mainly due to the jet, with turbulence having little effect, as shown by the similar profiles of Turb+Jet and JetOnly, while TurbOnly remains nearly constant.