CGM/IGM Feedback Physics Weekly: August 3–9, 2026

Six research preprints submitted to arXiv from August 3 through August 9 examine different links in the CGM/IGM feedback chain: halo gas depletion, the X-ray luminosity–temperature relation, cold-gas kinematics, intracluster gas motions, and a joint density–pressure probe. The shared problem is measurement: a lower X-ray signal, a missing gas reservoir, and a pressure deficit do not identify the same physical process.

The code and dataset categories remain empty this week. Targeted searches did not produce a qualifying GitHub release or public astronomy data release tied directly to CGM/IGM feedback physics in the same window.

Authors and date: Haojie Xia, Feng Yuan, Bocheng Zhu, Haoen Zhang, Tingfang Su, Aoyun He, and Suoqing Ji; submitted August 4, 2026. 1

Study design: The paper runs high-resolution 3D hydrodynamical simulations with the MACER3D framework for a dwarf elliptical, a massive elliptical, and a cluster-central galaxy. Controlled runs isolate AGN winds and supernova feedback, leaving out cosmological inflow and environmental effects. 1

Result: The dominant regulator depends on halo mass. In the cluster-central model, neither AGN winds nor supernovae alone suppress the gas density and X-ray luminosity enough; their combined action pushes the model below the observed relation, while adding AGN jet feedback resolves that mismatch. In massive ellipticals, AGN feedback brings the model broadly into the observed –$T$ relation. In dwarfs, AGN winds transport supernova-enriched gas to intermediate radii, raise the metallicity there, and strengthen cooling, creating a fountain-like recycling loop. 1

Why read: It is a useful warning against treating “AGN feedback” as one interchangeable knob. The same broad label covers winds, jets, removal, and recycling, and the observable response changes from dwarfs to cluster centrals.

Caveat: The three idealized galaxy contexts are controlled experiments, not a cosmological population. The exclusion of inflow and environmental effects makes the feedback comparison cleaner, but it also limits how directly the –$T$ result can be mapped onto real groups and clusters. 1

Authors and date: Natan de Isídio, Paola Popesso, Sandro Tacchella, Anna Pasquali, Ilaria Marini, Daudi Mazengo, Victoria Toptun, and Sean McGee; submitted August 4, 2026. 2

Study design: The authors reconstruct baryonic, dark-matter, structural, and chemical histories for about 7,300 IllustrisTNG galaxies, including 2,800 satellites, using time since infall as the clock. The sample is analyzed through MaNGA-like mock galaxies rather than direct observations. 2

Result: Satellites lose about 90% of their hot CGM within 4.2 ± 0.6 Gyr after infall, with the loss increasing with residence time and showing little dependence on stellar mass. Hot-gas mass correlates strongly with star-formation rate. Quenched satellites were accreted about 6.5 ± 0.3 Gyr ago, compared with 4.3 ± 0.3 Gyr for star-forming satellites, and they continue forming stars for at least roughly 3 Gyr after infall before declining rapidly. 2

Why read: The paper gives the delayed-then-rapid picture a concrete reservoir. The relevant clock is not simply “infall, then quenching”; it is the gradual erosion of the hot CGM that can keep supplying the satellite for several gigayears.

Caveat: The result is a model-history inference inside IllustrisTNG, not a direct measurement of hot gas around an individual satellite. Ram-pressure stripping, tidal effects, and the simulation's treatment of the surrounding medium still shape the inferred timeline. 2

Authors and date: Connor Jennings, Earl P. Bellinger, Imad Pasha, Pieter van Dokkum, and Pratik J. Gandhi; submitted August 4, 2026. 3

Study design: The authors use 182 Milky Way/Andromeda analogues from the TNG50 cosmological simulation. They generate H-alpha emission maps with Cloudy models and line-of-sight-averaged 2D velocity maps, then train a UNet to infer plane-of-sky velocities from the emission morphology. Noise is forward-modeled to match future observing facilities. 3

Result: The network generally recovers the overall flow direction. At depths expected for facilities such as MOTHRA, the typical RMS error in the plane-of-sky velocity direction is 0.3–0.5 virial velocities. If the emission maps reach those depths, they could add two phase-space dimensions that traditional line-of-sight measurements cannot constrain. 3

Why read: This is a measurement-design paper rather than a new feedback diagnosis. Its payoff is practical: a morphology map could help select sightlines and targets where follow-up spectroscopy has the best chance of revealing CGM circulation.

Caveat: The network learns from TNG50 and from forward-modeled H-alpha emission. The quoted error is therefore a simulation-to-observation forecast, not a demonstrated recovery on real CGM maps; mismatches in emissivity, cloud structure, or feedback physics could change the result. 3

Authors and date: Erwin T. Lau, Naomi Ota, and Daisuke Nagai; submitted August 5, 2026. 4

Study design: Using TNG-Cluster, the paper compares intrinsic three-dimensional gas motions with mock XRISM/Resolve observations. It models how cool-core state, formation history, projection, and azimuthal variation alter the inferred non-thermal pressure fraction. 4

Result: Projection and azimuthal structure lower the observed velocity-dispersion and non-thermal-pressure signals, especially at larger radius, partly hiding the intrinsic outward rise. They still cannot explain the extremely low values measured for Abell 2029: that system falls below roughly the 0th–6th percentiles of the simulated cool-core distribution at every measured radius under both turbulence-only and turbulence-plus-bulk definitions. 4

Why read: The paper separates an observational geometry problem from a model problem. That distinction matters when feedback studies use cluster pressure support to infer gas ejection or hydrostatic mass bias.

Caveat: The remaining Abell 2029 tension is not assigned to a single missing mechanism. The authors leave open rare dynamical conditions and missing physics in current ICM models, so the result is a constraint on the model space rather than a feedback detection. 4

Authors and date: Isabel Medlock and Daisuke Nagai; submitted August 6, 2026. 5

Study design: The paper uses the Baryon Pasting framework with separate parameters for feedback efficiency and non-thermal pressure amplitude. The tSZ effect supplies an electron-pressure tracer, while FRB dispersion measures supply an electron-density tracer. The authors combine the cross-power spectrum with existing detections and forecast samples from DSA-2000, SO, and CMB-HD. 5

Result: In the noise-free limit, the conditional parameter correlation drops from 0.96 to 0.08, a 12-fold reduction, and the joint figure of merit improves by a factor of 19. The information is concentrated at cluster-interior scales with multipole . For 50,000 DSA-2000 FRBs, the forecast reaches 3.8% precision on the non-thermal pressure amplitude with SO and 2.3% with CMB-HD. 5

Why read: This is the week's clearest answer to a recurring calibration problem: integrated thermal pressure alone cannot tell you whether a cluster's gas was expelled, rearranged, or supported non-thermally. Pairing pressure and density gives the parameters more room to separate.

Caveat: The precision forecasts depend on the Baryon Pasting parameterization, assumed FRB populations, survey noise, and the availability of large FRB samples. The paper reports consistency with recent detections, but the new constraints remain a model-and-survey forecast rather than a direct measurement of feedback efficiency. 5

Authors and date: A. Kruglov, N. Lyskova, I. Khabibullin, V. Biffi, and K. Dolag; submitted August 6, 2026. 6

Study design: The authors make a one-to-one comparison between stacked soft X-ray surface-brightness profiles of several dozen low-redshift massive clusters and population-averaged predictions from Magneticum. The comparison uses the 0.3–2.3 keV band and focuses on the gain from stacking faint outer regions. 6

Result: The simulation agrees well with the recent eROSITA stacked profile outside the central region. A clear central mismatch suggests that the implemented AGN feedback redistributes gas too effectively in cluster cores. Beyond several virial radii, the simulated signal becomes so noisy and faint that it sits orders of magnitude below a fluctuating, radially flat background. 6

Why read: It gives feedback calibration a spatial scale. A model can match the population-averaged X-ray profile through much of the halo while still moving too much gas in the core, and the outskirts may be limited by background rather than sample size.

Caveat: Agreement with a stacked profile does not identify which feedback process produced it. The central discrepancy is described as likely AGN-related, and the outer-halo conclusion is limited by detectability; neither point is a direct measurement of an AGN energy budget. 6

Start with Circumgalactic medium depletion drives satellite quenching if your immediate question is how a galaxy's environment removes its fuel: it supplies a timescale and a reservoir, while keeping the result inside a simulation. Read The mass-dependent interplay... next if you need a controlled test of which feedback channel changes the hot-gas observable at different halo masses.

For cluster work, pair Non-Thermal Pressure... with Beyond Feedback.... The first says how projection and gas motions bias an observable; the second asks how pressure and density can be combined so the bias is less degenerate with feedback efficiency. Average soft X-ray surface brightness... adds a radial calibration check, especially for core redistribution and the faint outskirts.

The cold-CGM CNN paper is the most forward-looking of the six. It does not establish a feedback effect, but it offers a way to measure flow geometry that the other papers mostly infer from integrated or line-of-sight quantities.

The common thread is a demand for matched scales and matched tracers. Hot-CGM depletion after satellite infall, X-ray emission from cluster gas, cold-cloud morphology, non-thermal ICM pressure, and FRB electron columns are not interchangeable feedback indicators. A model that fits one can still fail another because the coupling location, gas phase, and tracer lifetime differ.