CGM/IGM Feedback Physics Weekly: August 10–16, 2026
Nine new arXiv preprints connect early baryon recycling, phase-dependent galactic winds, AGN timing, FRB electron clustering, hot-CGM absorption, and IGM-heating model uncertainty; no qualifying code or dataset release was verified.
The week's papers put the same feedback problem on three different clocks. A JWST spectrum catches a cool outflow beginning to recycle only 500 million years after the Big Bang; a local survey measures how warm winds change with galaxy mass; and FRB dispersion measures now reject a no-feedback baryon distribution at 9σ. The remaining studies ask how AGN feedback changes morphology, how much hot gas sits around M31, and whether the simulations used to interpret IGM heating agree closely enough for parameter inference.
Nine arXiv preprints submitted from August 10 through 16 qualify. No new GitHub release, repository, or public astronomy data release tied directly to CGM/IGM feedback physics was verified in the same window.
The scan
| Item | What is new | Method and scale | Read priority | Main caveat |
|---|---|---|---|---|
| Gz9p3 baryon cycle | First direct cool-outflow density at z=9.311; high mass loading | JWST spectroscopy; a merging galaxy 500 Myr after the Big Bang | High for early feedback and recycling | One galaxy; inferred loading and escape depend on geometry and modeling |
| Cosmic Noon outflow scaling | Outflow velocities up to 3× higher than at low redshift at fixed galaxy property; no comparable evolution in mass loading | 387 galaxies at z≈0–9; Keck/MOSFIRE line decomposition | High for empirical wind prescriptions | Heterogeneous redshift sample and warm-ionized phase only |
| CLASSY XV | Cool and warm ions trace different wind behavior; average maximum velocity reaches 620 km/s | 17 local star-forming outflows; UV absorption radiative-transfer modeling | High for phase-dependent loading factors | Small sample and model-dependent ionization structure |
| AGN and disk suppression | AGN becomes effective just after disk spin-up, suppressing late thin-disk growth | Two Milky Way-mass FIRE-2 halos, with and without several AGN models | High for morphology–quenching links | Two halos cannot establish population-level robustness |
| FRB DM × galaxy and kSZ | 6.5σ baryon signal; no-feedback gas-tracing scenario disfavored at about 9σ | 130 localized FRBs, DESI galaxies, simulations, and a halo model | High for feedback constraints beyond individual halos | Current FRB sample and halo-model assumptions limit precision |
| M31 hot CGM | Marginal inner-halo O VII/O VIII excess; inferred mass changes sharply with the CGM boundary | XMM-Newton absorption toward 15–18 background AGN, out to 190–730 kpc | High for a concrete hot-CGM census test | 2.2–2.3σ combined significance and strong foreground/boundary dependence |
| LINER outflow efficiency | Extended ionized LINER winds are inefficient γ-ray accelerators; compact jets remain favored for the most extreme cases | 17 years of Fermi-LAT plus spatially resolved MEGARA kinematics | Medium for energy partition and AGN-wind models | Mostly upper limits; it tests particle acceleration, not CGM escape |
| X-ray heating code comparison | Similar mean IGM temperatures but about 30% different 21-cm power spectra; posterior shifts can exceed 1σ | Licorice 3D radiative transfer versus Beorn 1D radiative transfer | High for IGM-feedback inference pipelines | The comparison isolates code approximations, not a new physical feedback measurement |
| Recycled gas feeding SMBHs | Recycled, metal-rich gas dominates SMBH accretion in the simulated massive-galaxy sample | 30 high-resolution cosmological zoom-ins with particle-history tracing | Medium for the supply side of AGN feedback | Results apply to z=0 massive simulated galaxies and do not measure feedback coupling |
The most direct observation-to-model chain runs from the early outflow to the local wind samples: density and mass loading set the material budget, while phase and velocity determine how much of that material can remain in or leave the halo. The FRB and M31 papers then test the accumulated gas on larger scales, but their observables are not interchangeable: one measures electron clustering statistically, the other absorption along sparse sightlines.
Per-item briefs
1. An emerging baryon cycle in a galaxy 500 million years after the Big Bang
Authors and date: Shengzhe Wang et al.; submitted August 10, 2026. Source: arXiv abstract page.
Study design. The authors use deep JWST spectroscopy of Gz9p3, a merging galaxy at redshift 9.311, to resolve the gas phases associated with a merger-driven tidal structure. Fine-structure absorption supplies an electron-density measurement for the cool outflowing gas, while the spectrum also shows a substantial neutral reservoir and a multiphase outflow.
Result. The cool outflow has an electron density of about 17 cm⁻³ and a mass-loading factor among the highest measured for galaxies of comparable stellar mass. The outflow is unlikely to escape the host halo, so metal-enriched material can remain available for later recycling through the CGM.
Why read it. This is unusually direct evidence for the transition from a burst that built stars rapidly to a feedback-regulated baryon cycle during reionization. It connects a measured cool-gas density to the question that simulations often answer only statistically: whether early stellar feedback ejects gas permanently or seeds a later fountain.
Caveat. The claim rests on one merging system. Mass loading, geometry, and the conclusion that the gas remains bound are model-dependent, so the paper establishes a physical case study rather than a population-average duty cycle.
2. Scaling relations of galactic outflows across cosmic time
Authors and date: Tiffany Liou et al.; submitted August 10, 2026. Source: arXiv abstract page.
Study design. The study builds a joint sample of 387 galaxies spanning approximately z=0–9, including 98 new Cosmic Noon galaxies from the Keck Baryonic Structure Survey and the Keck Lyman Continuum Spectroscopic Survey. High signal-to-noise Keck/MOSFIRE emission lines are decomposed into narrow and broad components to identify warm-ionized outflows and compare their velocities, mass outflow rates, and loading factors with stellar mass, star-formation rate, and star-formation surface density.
Result. At fixed stellar mass, star-formation rate, or star-formation surface density, Cosmic Noon outflows reach maximum velocities up to three times those at low redshift. Their mass outflow rates are also higher at fixed stellar mass, but the mass-loading factor shows no corresponding evolution at fixed stellar mass. The authors find evidence that most of the outflowing gas is recycled at roughly 0.03 Rvir, and infer an orientation- and geometry-corrected outflow occurrence rate of at least 30%.
Why read it. The result separates wind speed from mass loading. A feedback model that matches only the velocity evolution can still put the wrong amount of mass into the CGM. The small recycling radius also gives simulations a concrete scale at which to test fountain prescriptions.
Caveat. The warm-ionized phase is not the full multiphase wind, and the redshift-assembled sample combines different observing programs and selection functions. The reported recycling scale therefore constrains the observed phase and adopted outflow model, not necessarily the total baryon return path.
3. CLASSY XV: kinematics and spatial distributions of outflows in local highly star-forming galaxies
Authors and date: Mason S. Huberty et al.; submitted August 12, 2026. Source: arXiv abstract page.
Study design. The authors model UV absorption in 17 local CLASSY galactic outflows, following Si II, Si III, and Si IV as tracers of cool and warm phases. They use radiative-transfer modeling to derive mass, momentum, and energy loading factors and compare the inferred wind profiles with FIRE-2 and CGOLS hydrodynamic simulations.
Result. Si II-traced winds behave differently from the warmer Si III and Si IV components. Loading factors scale inversely with stellar mass, and the mass and momentum results agree with the FIRE-2 comparison. The modeled velocity profiles reach an average maximum of 620 km/s. Outflows associated with young star-forming regions are more likely to have cooler-gas-dominated columns, while mass outflow rates decrease with radius.
Why read it. This paper supplies an observational warning against treating one ion as a proxy for the whole wind. The phase contrast matters for both feedback calibration and the interpretation of CGM absorption: a change in tracer can look like a change in mass or energy transport when it is actually a change in the gas phase being sampled.
Caveat. Seventeen systems cannot fix a universal loading law, and the ionization and geometry assumptions enter the conversion from column density to mass and energy. The comparison with simulations is therefore a consistency check, not a unique model selection.
4. Old disks die hard: how AGN feedback suppresses disk formation in Milky Way-mass galaxies
Authors and date: Patricia Fofie et al.; submitted August 13, 2026. Source: arXiv abstract page.
Study design. The authors analyze two Milky Way-mass halos simulated with FIRE-2 physics. Each halo is evolved without AGN feedback and with up to three AGN implementations, allowing the study to compare morphology and star formation across otherwise related histories.
Result. The no-AGN runs retain prominent, thin, star-forming disks at z=0. AGN runs have less late-time star formation, a higher spheroid fraction, and a lower thin-disk mass fraction; two produce quenched lenticular galaxies with no cool gas at z=0. The timing is the central result: AGN feedback becomes effective just after disk formation begins, at the simulated ``spin-up'' stage. It suppresses late thin-disk growth while leaving the early-built spheroid broadly similar.
Why read it. The paper ties quenching to when feedback couples, not only to the integrated energy injected. It gives a morphological diagnostic for CGM and galaxy-formation models: two runs can end with similar spheroids but different thin disks if their feedback turns on at the wrong phase of assembly.
Caveat. Two halos and several feedback implementations cannot establish a population-level causal law. The result is also specific to FIRE-2 physics and the chosen AGN prescriptions, so the timing should be tested rather than transplanted unchanged into other models.
5. A new measurement of the FRB DM-galaxy cross correlation and a first joint analysis with the kinematic SZ effect
Authors and date: Samuel McCarty et al.; submitted August 11, 2026. Source: arXiv abstract page.
Study design. The authors cross-correlate dispersion measures from 130 localized fast radio bursts with the DESI Legacy Survey Bright Galaxy Sample, adding new host redshifts from DSA-110. They measure the signal in configuration and harmonic space, compare it with simulations and a halo model, and combine it with kinematic Sunyaev–Zeldovich measurements.
Result. The FRB DM–galaxy signal is detected at 6.5σ, the highest significance reported in the paper's comparison. The measurements disfavor a no-feedback model in which gas simply traces dark matter at about 9σ. The paper also presents a first joint attempt to use FRBs and kSZ to constrain the optical-depth degeneracy and the growth parameter fσ8.
Why read it. This is a feedback constraint on electron distribution beyond a single halo. It connects the small-scale redistribution of gas by feedback to a large-scale clustering statistic, giving simulations a target that includes both halo gas and inter-halo electrons.
Caveat. The current FRB sample remains small, and the inference depends on the halo model and simulation calibration. The result establishes that feedback is required by the statistic, but it does not by itself identify which feedback channel or radial coupling profile is responsible.
6. An X-ray absorption-line survey of the circumgalactic medium of M31 with XMM-Newton
Authors and date: Kaile Wang et al.; submitted August 10, 2026. Source: arXiv abstract page.
Study design. The study searches for O VII and O VIII absorption toward 15 and 18 background active galactic nuclei, respectively, with sightlines reaching impact parameters of about 190–730 kpc around M31. It compares inner and outer sightlines to isolate a hot-CGM contribution above the Milky Way foreground.
Result. The inner sightlines show a marginal excess consistent with hot gas associated with M31. The mean equivalent widths are about 4–9 mÅ for inner O VII sightlines near 310 kpc and about 17 mÅ for inner O VIII sightlines near 200 kpc, with a combined nominal significance of 2.2–2.3σ. If all of the excess is assigned to gas inside the virial radius, the inferred mass exceeds the nominal missing-baryon budget; extending the assumed boundary to 400–500 kpc lowers the estimate to a few ×10¹¹ solar masses.
Why read it. The result exposes how much a hot-CGM census depends on the spatial boundary chosen before the mass is inferred. For feedback studies, that is not a bookkeeping detail: the same absorption signal can imply very different retention and heating histories when the CGM is defined at different radii.
Caveat. The signal is marginal, foreground subtraction is difficult, and the mass is strongly model-dependent. The paper does not yet isolate AGN or stellar feedback from other sources of hot halo gas; deeper exposures and more inner-halo sightlines are needed.
7. Constraining the gamma-ray efficiency of LINER outflows with Fermi-LAT and MEGARA
Authors and date: Alberto Domínguez et al.; submitted August 12, 2026. Source: arXiv abstract page.
Study design. The authors combine spatially resolved optical integral-field kinematics from MEGARA at the Gran Telescopio Canarias with 17 years of Fermi-LAT observations. The optical data provide ionized-outflow kinetic powers; the gamma-ray data provide detections or upper limits over 0.05–500 GeV for a local LINER sample.
Result. No LINER in the sample is formally detected in gamma rays. The strongest constraint, for the radio-loud LINER NGC 1052, is a gamma-ray radiative efficiency below 41%; for the remaining systems, upper limits generally sit above the available outflow kinetic power. The authors conclude that extended ionized outflows are inefficient high-energy particle accelerators and that compact nuclear jets are favored for the most extreme gamma-ray efficiencies.
Why read it. The paper narrows the energy budget of one AGN-wind channel. If extended winds cannot efficiently turn kinetic power into high-energy emission, gamma-ray luminosity should not be used as a simple proxy for the energy deposited into the wider CGM.
Caveat. Most conclusions are upper-limit based, and the study tests particle acceleration rather than the fate of wind material on CGM scales. It is most useful as an energy-partition constraint, not as a direct measurement of halo heating.
8. How X-rays heat the IGM in different 21-cm simulation codes
Authors and date: Romain Meriot, Jonathan R. Pritchard, and Timothée Schaeffer; submitted August 14, 2026. Source: arXiv abstract page.
Study design. The authors compare Licorice, a three-dimensional radiative-transfer code, with Beorn, a one-dimensional code. They feed Beorn sources extracted from Licorice while holding source physics fixed, then compare IGM temperature fields, global 21-cm signals, power spectra, and posterior distributions from an emulator-based MCMC analysis.
Result. The two setups agree on luminosity fields, mean IGM temperatures, and the global 21-cm signal, but their spatial temperature and 21-cm distributions differ. The resulting 21-cm power spectra differ by about 30%, and the inferred one-dimensional-code posteriors show typical biases of at least 1σ for a noise level corresponding to 100 hours of SKA observations.
Why read it. The study identifies a failure mode in feedback inference: agreement in a mean thermal history does not guarantee agreement in the spatial statistic used to recover astrophysical parameters. This matters for X-ray heating from stellar remnants and AGN, where the inferred feedback history can inherit the radiative-transfer approximation.
Caveat. The paper compares modeling approximations rather than measuring a new feedback signal, and the stated posterior bias depends on the assumed instrument noise and emulator pipeline. It is a methods warning, not evidence for one heating history.
9. Recycled gas dominates the metal-rich fuel of supermassive black holes
Authors and date: Dongyun Kwak et al.; submitted August 12, 2026. Source: arXiv abstract page.
Study design. The authors trace individual gas particles through 30 high-resolution cosmological zoom-in simulations of massive galaxies at z=0. They classify gas accreted by the central SMBH as early, external, recycled, or smooth IGM-accreted material, and follow its chemical enrichment history.
Result. Recycled gas, including material enriched by stellar evolution, AGB winds, and supernova ejecta, dominates the SMBH accretion budget in the simulated sample and is already metal-rich at early times. Gas from other origins becomes enriched in the galactic environment before accretion, while the mean abundance ratios evolve only weakly with redshift and are broadly compatible with the high metallicities inferred for quasar broad-line regions.
Why read it. Feedback models often focus on how AGN energy leaves the nucleus. This paper supplies the complementary supply-side constraint: stellar recycling can dominate the metal-rich fuel available before an AGN episode, linking stellar feedback, CGM return flows, and SMBH growth.
Caveat. The result applies to 30 simulated massive galaxies at z=0 and does not measure the coupling efficiency or outflow fate of the resulting AGN activity. The category labels also depend on how the simulation defines recycled and external gas.
Cross-paper comparison
| Axis | What this week's papers constrain | Where the evidence is strongest | What remains open |
|---|---|---|---|
| Coupling location | AGN feedback can act just after disk spin-up; FRB clustering requires baryons to deviate from dark-matter tracing; M31 absorption tests gas hundreds of kpc out | The FIRE-2 controlled comparisons and the FRB detection provide the clearest causal and statistical tests | Neither identifies a unique radial coupling profile for AGN versus stellar feedback |
| Gas phase | Gz9p3 and CLASSY separate cool or warm outflowing gas; M31 probes hot ions; FRBs probe free electrons; the 21-cm comparison follows volume-filling IGM heating | Phase-resolved spectroscopy and the FRB electron statistic complement each other | Loading factors and mass budgets cannot be combined without a multiphase conversion model |
| Tracer lifetime | Absorption lines sample an instantaneous line of sight; FRB and kSZ statistics accumulate gas over cosmological structure; morphology records the integrated timing of feedback | The contrast is clearest between CLASSY's ion-specific winds and the FRB galaxy cross-correlation | A single tracer still cannot establish the lifetime of gas between launch, recycling, and escape |
| Sampled scale | Gz9p3 and the local wind surveys start near the galaxy; M31 reaches roughly 200–730 kpc; FRB statistics extend beyond individual halos; 21-cm power spectra sample the IGM field | The combination covers the missing link between launch and inter-halo gas | Cross-scale comparisons still require a consistent definition of the CGM boundary and feedback energy budget |
The practical reading order is therefore not the order of the table. Start with Gz9p3, CLASSY XV, and the Cosmic Noon scaling relations if the question is how much gas is launched and recycled. Add the AGN morphology paper for timing, then the FRB result for a large-scale baryon constraint. Use M31 and the code comparison to test whether the inferred hot phase and IGM heating are robust to boundary and radiative-transfer assumptions. The LINER and SMBH-fuel papers are the narrower follow-ups for energy partition and the supply side of AGN feedback.
Coverage note: No qualifying GitHub code release or major astronomy-repository dataset release was verified for August 10–16, 2026. The code and dataset categories are left empty rather than backfilled with older material.
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.