Introducing RAFIKI: Refining AGN Feedback in Kinetic Implementations Journal: The Astrophysical Journal
Skylar Grayson, Evan Scannapieco, Romeel Davé, Arif Babul, Renier T. Hough · 2025
Read Introducing RAFIKI: Refining AGN Feedback in Kinetic Implementations Journal: The Astrophysical Journal by Skylar Grayson, Evan Scannapieco, Romeel Davé, Arif Babul, Renier T. Hough (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-19924.txt.
Introduction
Skylar Grayson and four co-authors introduce RAFIKI, a suite of galaxy simulations designed to pull apart how black hole feedback works. Cosmological simulations need active galactic nucleus feedback to stop massive galaxies forming too many stars, but the details are poorly pinned down. RAFIKI builds on the SIMBA-C simulation and varies the mass loading of black-hole-driven winds.
Unlike the base simulation, RAFIKI separates the efficiencies of the two kinetic feedback modes, so their effects on galaxies, black holes, and surrounding gas can be studied independently. Even with stronger loading, the slower quasar-type mode cannot quench massive galaxies, though it regulates black hole growth and star formation in mid-sized ones. The runs also expose parameter degeneracies that future observations could break.
Catalog record
| CREATOR | Skylar Grayson, Evan Scannapieco, Romeel Davé, Arif Babul, Renier T. Hough |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-19924
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Introducing RAFIKI: Refining AGN Feedback in Kinetic Implementations Journal: The Astrophysical Journal Skylar Grayson, Evan Scannapieco, Romeel Davé, Arif Babul, Renier T. Hough Abstract Modern cosmological simulations have now matured to the point of reproducing the evolution of realistic galaxy populations across cosmic time. These simulations rely on feedback from active galactic nuclei (AGN) to quench massive galaxies, yet the details of this process remain poorly understood. To address this issue, we introduce RAFIKI (Refining AGN Feedback In Kinetic Implementations), a novel suite of simulations built upon SIMBA-C that vary the mass loading of AGN-driven winds. Unlike the fiducial SIMBA-C simulation, RAFIKI separates the efficiencies of the two kinetic feedback modes, enabling a detailed study of their impact on galaxies, black holes, and the circumgalactic medium. We explore a range of galaxy and baryon properties in the RAFIKI runs, and find that even with enhanced mass loading, the lower-velocity quasar-type mode cannot quench massive galaxies. However, it plays a significant role in regulating black hole growth and star formation in intermediate-mass galaxies. We also uncover degeneracies in the parameter space that highlight the lack of current constraints on AGN feedback. RAFIKI provides a controlled framework to disentangle these degeneracies using current and upcoming observations. I Introduction Cosmological hydrodynamic simulations are an essential component of modern astrophysics, providing a tool for interpreting a wide range of multiwavelength observations. These models strive to realistically emulate the processes governing galaxy evolution, including gravity and structure formation, gas heating and cooling, chemical evolution, star formation, and feedback from stellar winds, supernovae, and active galactic nuclei (AGN) (see Crain & van de Voort 2023 , for a review) . AGN feedback, in particular, plays a key role, as it is essential to reproduce both the quenching of massive galaxies and their observed stellar mass distribution ( Thacker et al. 2006 ; Sijacki et al. 2007 ; Schaye et al. 2015 ; Pillepich et al. 2018 ; Davé et al. 2019 , e.g.) . However, substantial uncertainty remains regarding the physics of accretion and outflows from supermassive black holes. The complexity of magnetohydrodynamic plasma in accretion disks is not fully understood, nor are the mechanisms that power outflows ( Yuan & Narayan 2014 ; Pathak et al. 2025 , e.g.) . The energetics of these outflows are also often unconstrained, and there is still no consensus on how AGN feedback acts to quench galaxies ( Fabian 2012 ; De Lucia et al. 2026 ; King & Pounds 2015 ) . Beyond these uncertainties, the low resolution of cosmological simulations means AGN feedback must be treated as a sub-grid process, which cannot be modeled from first principles and must be approximated on resolved scales. Numerous free parameters are involved in this approximation, and every simulation adopts different prescriptions, sometimes leading to drastically different implementations. Despite this, most modern cosmological simulations are generally successful in matching the stellar properties of galaxies, because the stellar mass function, the stellar-halo mass relationship, and the stellar-black hole mass relationship are often used to calibrate the feedback parameters ( Crain & van de Voort 2023 ) . However, the differences in feedback models become far more consequential when considering diffuse baryons. Gas inflow and outflow rates vary between simulations, sometimes by a factor of 3 or more ( Crain & van de Voort 2023 ) , and this can impact the distribution of gas on large scales. Baryon fractions, particularly in lower-mass halos, are not well constrained observationally and differ drastically between simulations ( Davies et al. 2020 ) , and there is growing evidence that most simulations do not align with observations of gas reservoirs. Studies comparing stacked observations of the hot phase of the circumgalactic medium (CGM) using X-ray emission or the thermal Sunyaev Zel’dovich effect suggest that simulations such as SIMBA and EAGLE are failing to capture all the processes shaping the CGM ( Grayson et al. 2023 ; Grayson et al. 2025 , e.g.) . Thus there is a growing need to better understand and constrain the physical drivers of AGN feedback. Observationally, AGN have historically been separated into two modes, often labeled ‘radiative’ and ‘kinetic’ (see Fabian 2012 ; Heckman & Best 2014 for reviews). The first is thought to occur at high Eddington ratios (i.e. quasars), where the energy output is primarily through radiation, while the second is dominated by kinetic energy input, with accretion disks fueling jets of material that can form large radio lobes far beyond the disk of the host galaxy. These observational trends motivate how AGN feedback is implemented in simulations to various degrees. Some simulations, like EAGLE ( Schaye et al. 2015 ) and ROMULUS ( Tremmel et al. 2019 ) , deposit AGN feedback as thermal energy around the black hole, using a single mode that scales with the black hole mass and accretion rate. The recent OBSIDIAN simulation implements three modes of feedback, wherein very low accretion rates simultaneously power both a jet and a wind driven by the accretion flow ( Rennehan et al. 2024 ) . Others, like IllustrisTNG ( Pillepich et al. 2018 ) , Horizon-AGN ( Dubois et al. 2012 ) , and SIMBA ( Davé et al. 2019 ) attempt to capture a separated dual-mode behavior, providing two pathways for AGN feedback to occur depending on the accretion rate of the central supermassive black hole. In the case of IllustrisTNG, at high accretion rates the energy is deposited thermally, but at low accretion rates, outward velocity kicks are given to gas particles around the black hole ( Nelson et al. 2019 ) . SIMBA also follows this dual mode prescription, but it only uses kinetic feedback, with high velocities at low accretion rates (the ‘jet’ mode) briefly decoupled from the hydrodynamics to avoid numerical inaccuracies, prevent the feedback energy from being trapped around the black hole, and depositing the energy of the jet mode on scales comparable with observations of radio lobes. SIMBA also stands out in its choice of directionality, as the velocity kicks are not random, as in IllustrisTNG and OBSIDIAN, but rather the gas is ejected in a bipolar fashion orthogonal to the black hole accretion disk. When jets are at full speed in relatively gas-poor galaxies ( M_{\rm gas}/M_{*}<0.2 ), SIMBA also includes an ‘X-ray heating’ mode, which injects a small percentage of the jet energy as thermal energy around the black hole ( Davé et al. 2019 ) . A common trend in these and other simulations that use dual mode AGN feedback is the dominance of one mode over the other in shaping galaxy properties. In both SIMBA and IllustrisTNG, the high velocity kinetic mode associated with low accretion rates is most important for quenching star formation and shaping a realistic galaxy population at high mass scales ( Scharré et al. 2024 ; Tillman et al. 2025 ) . It is also the mode that impacts the large-scale environments, changing the properties of the circumgalactic medium and impacting the dynamics of the intracluster medium. By contrast, the high accretion rate mode plays a less significant role, although it does have an impact at lower mass scales. In SIMBA, this so-called ‘wind’ mode is believed to play a small