Effects of Varied Cosmic Ray Feedback from AGN on Massive Galaxy Properties

Active galactic nuclei (AGN) provide energetic feedback necessary to `turn off’ star formation in high-mass galaxies (M${\rm halo} \geq $ 10$^{12.5}$ M${\odot}$, $10.4 \leq \log(\frac{M_*}{M_\odot}) \leq 11$) as observed. Cosmic rays (CRs) have been proposed as a promising channel of AGN feedback, but the nature of CR feedback from AGN remains uncertain. We analyze a set of high-resolution simulations of massive galaxies from the Feedback in Realistic Environments (FIRE-3) project including multi-channel AGN feedback, explicitly evolving kinetic/mechanical, radiative, and spectrally-resolved CRs from the central black hole. Specifically, we explore different CR feedback and transport assumptions, calibrated to Milky Way local ISM constraints, and compare them to observed galaxy scaling relations. We find that all parameterizations explored self-regulate within agreement with observed galaxy scaling relations, demonstrating that CR injection efficiencies varied by $\sim$1.5 dex and locally-variable transport produce quenched galaxies with reasonable bulk properties; however, they feature orders-of-magnitude variant circumgalactic medium (CGM) gas properties. Our results indicate that multi-wavelength synthetic observations probing these varied halo properties from larger simulated samples in conjunction with observational comparisons may place novel constraints on how AGN physically quench star formation in massive galaxies.

💡 Research Summary

This paper investigates how cosmic rays (CRs) emitted by active galactic nuclei (AGN) influence the evolution of massive galaxies (halo mass ≥ 10¹²·⁵ M⊙, stellar mass ≈ 10¹⁰·⁴–10¹¹ M⊙). Using the latest FIRE‑3 suite, the authors run high‑resolution cosmological zoom‑in simulations of four “m13” halos (M_halo ≈ 10¹³ M⊙) that include a full CR‑magnetohydrodynamic (CR‑MHD) treatment, stellar feedback, and multi‑channel AGN feedback (mechanical jets, radiation, and CRs).

Two CR transport prescriptions are explored: (1) a constant diffusion model (CD) with a power‑law rigidity dependence (κ_eff ∝ R⁰·⁶), and (2) a variable‑diffusion model (VD) whose diffusion coefficient is derived from local ISM properties and calibrated to Milky Way Voyager and AMS‑02 data. For the VD case three AGN CR injection efficiencies are tested (ϵ_BH,cr = 10⁻⁴, 10⁻³, 3×10⁻³), giving a total of seven distinct runs.

The main findings are:

1. Galaxy‑scale regulation – All models reproduce the observed stellar‑mass–halo‑mass relation (Behroozi et al. 2019), the SMBH mass–velocity‑dispersion relation (Greene et al. 2020), and lie in the quenched region of the SFR–mass diagram at z = 0. Higher CR injection efficiencies lead to stronger quenching, but even the low‑efficiency VD runs achieve regulation comparable to the CD runs because the variable diffusion enhances CR pressure in the central ISM.

2. SMBH growth – The inclusion of CR feedback modestly affects black‑hole accretion histories. The highest‑efficiency VDHiCR run produces SMBHs ≈0.2 dex more massive than the median observational relation, yet the M_BH–σ correlation remains intact.

3. Circumgalactic medium (CGM) diversity – Despite similar bulk galaxy properties, the CGM differs dramatically across models. Radial profiles of gas temperature, density, metallicity, and CR energy density vary by 1–2 dex between low‑efficiency (VDLoCR) and high‑efficiency (VDHiCR) runs. In the latter, CR pressure inflates the halo, lowering temperatures below 10⁶ K and boosting high‑ion column densities (e.g., O VI, Ne VIII). The CD models, with higher diffusion, retain hotter, denser CGM.

4. Observational signatures – The authors predict distinct multi‑wavelength diagnostics: UV absorption lines (O VI, C IV, Mg II) show strong dependence on CR efficiency; soft X‑ray surface brightness (0.5–2 keV) is reduced by up to 0.3 dex in CR‑rich halos; the thermal Sunyaev‑Zel’dovich signal is modestly suppressed (~10 %) because part of the pressure support is non‑thermal CR pressure. These differences provide a pathway to discriminate between CR transport models using current (HST/COS, XMM‑Newton, Chandra) and upcoming (XRISM, Athena, eROSITA) facilities.

5. Implications and future work – The study demonstrates that AGN‑driven CRs are sufficient to quench massive galaxies while simultaneously reshaping the CGM. However, the CR transport physics remains highly uncertain; the two prescriptions used bracket plausible extremes but do not capture the full complexity of MHD turbulence and wave‑particle interactions in low‑density halos. Further progress will require (i) a broader simulation suite covering diverse environments (cluster cores, field galaxies), (ii) more realistic CR acceleration spectra tied to jet physics, and (iii) a tight coupling between synthetic observations and real data to constrain the “micro‑physics” (diffusion coefficients, streaming speeds, scattering rates).

In conclusion, the paper provides compelling evidence that CR feedback from AGN can simultaneously satisfy galaxy‑scale quenching constraints and generate observable CGM signatures that differ by orders of magnitude. Multi‑wavelength observations of the CGM thus represent a powerful probe of the otherwise hidden CR‑AGN feedback channel, opening a new avenue for testing galaxy formation models.