Thirteen new feedback tests, from a wind that only exists below two parsecs to metals the halo keeps
Thirteen new arXiv results from October 26 to 30, 2026 test whether supernova winds carry neutral gas only at parsec resolution, whether the Milky Way's halo keeps the metals its galaxies made, and whether AGN feedback works through geometry rather than strength.
The week of October 26 to November 1, 2026 brought thirteen entries on the gas galaxies drive out, on the gas that waits in their halos, and on the heat that keeps it there. Fourteen papers fill them, because one entry covers a matched pair from a single simulation. No open-source release and no qualifying dataset deposit appeared in the window.
The arXiv listing at this hour carries submissions logged through Friday, October 30, so this edition rests on the window's first five days. The October 31 and November 1 tail is announced on Monday evening and was not screened.
An item earns a place below when the original reports a quantitative result on gas inside, leaving, returning to, or between galaxies and their halos. It must also describe a method another group could repeat and be reachable from a public link.
Two results set the week's range. A simulation of supernova-driven winds carries most of its mass as neutral clouds only when its cells reach about one parsec; coarsened to four parsecs, the neutral outflow disappears 1. Twenty-five fast radio burst sightlines put the Milky Way halo's metallicity at 0.12 to 0.39 times solar. That sits above what infalling intergalactic gas can supply 2.
At a glance
| Item | Design / method | Main result | Scale or tracer |
|---|---|---|---|
| He et al., neutral mass in supernova winds 1 | Quokka simulations of supernova-driven winds at 1, 2 and 4 pc resolution | Winds are neutral-dominated at 1 pc; the neutral outflow weakens at 2 pc and vanishes at 4 pc | pc-scale cells; HI cloud sizes and column densities |
| Li et al., thermal stellar feedback 3 | "Anchor-and-refine" scheme that pins each star particle's host cell and refines it when the cooling radius is unresolved | Hot superbubble interiors survive; outflow mass loading plateaus near 0.2 above 1 kpc; smoother star formation history than hybrid feedback | Supernova cooling radii; 1 kpc to halo scales |
| Piacitelli et al., CGM emission around dwarfs 4 | Marvelous Massive Dwarfs simulations, predicted Hα, [N II] and [O III] surface brightness | Outflows supply about 40% of the surface brightness and inflows about 50%; emission is eccentric, along the minor axis, and tracks recent star formation | Dwarfs below 10⁹ M☉; MUSE, KCWI and MOTHRA bands |
| Sanchez et al., simulated X-ray halos, part I 5 | Romulus25, more than 750 galaxies, forward-modelled 0.5–2 keV emission compared with eROSITA stacks | Stacked surface brightness comes out too bright, worst in the inner halo; a zero-metallicity emission model lowers L_X by nearly an order of magnitude | 10 < log M★ < 11.5; 0.5–2 keV |
| Sanchez et al., simulated X-ray halos, part II 6 | Same suite, tracing emissivity drivers in L★ halos with the PIONEER zoom-ins | Hot gas mass is set by the combined energy of stellar and AGN feedback; at fixed hot gas mass, metal-rich halos are brighter and metal-poor ones dimmer | Gas above 10⁵·⁵ K; soft X-rays |
| Crocker & Krumholz, cosmic rays in halos 7 | Semi-analytic transport and loss model tested against Fermi-LAT limits, the gamma-ray background, eROSITA stacking and pressure stability | Two scenarios survive; in the slower one cosmic rays supply the heat balancing X-ray cooling and about half the injected energy is still in the halo at z = 0 | L★ halos; GeV gamma rays and 0.5–2 keV |
| van Donkelaar et al., dense clumps at high redshift 8 | Two cosmological zoom-ins of one massive halo, one with mildly super-Eddington early black hole growth | Clump masses and sizes barely change; the enhanced run adds metal-rich outward-moving clumps during blow-out, some beyond the virial radius | One halo above z ≈ 6 |
| Su et al., anisotropic AGN feedback 9 | Galaxy-scale model carrying the angular structure of multi-zone GRMHD simulations into isolated M87*-like and Sgr A*-like galaxies | Angle-dependent feedback suppresses late stellar growth more than the isotropic reference at similar accretion rates; anisotropy matters more than polar mass loading | Three decades in black hole mass |
| Kumar et al., cool halo density and pressure 10 | HST/COS and STIS spectra of 23 galaxies and groups behind QSOs, with photoionization modelling | Cool-gas pressure matches hot-gas pressure from X-ray stacks; the cool phase's baryon fraction falls from 6–19% at 10¹² M☉ to 0.6–1.5% at 10¹³ M☉ | z = 0.4–0.7; 10¹²–10¹³ M☉ |
| Harvey et al., halo gas after a starburst 11 | DESI DR2 Mg II stacking over about 143 million galaxy–quasar pairs, with post-starburst ages from star formation histories | Post-starburst galaxies show the strongest absorption, 1.58 ± 0.32 Å within 71 kpc; older systems lose inner absorption and gain it at 300–1000 kpc | 0.4 < z < 0.76; 10.6 < log M★ < 12 |
| Roy et al., halo metallicity from FRB dispersion 2 | 25 FRB sightlines combined with metallicity-dependent oxygen absorption column estimates, bootstrap errors | Median halo metallicity 0.12, 0.26 or 0.39 Z☉ for diffuse baryon fractions of 0, 0.6 or 0.8 | Milky Way CGM; dispersion measures |
| Chen et al., a line-locking web in one quasar 12 | VLT/X-shooter medium-resolution spectroscopy of a quasar at z = 2.15 | Twelve absorption components and fourteen line-locking pairs across three ions; the chance probability is below 10⁻⁵ | About 2700 km/s; C IV, N V, Si IV, Lyα |
| Bisht & Sarkar, what lights the eROSITA bubbles 13 | Hydrodynamic simulations of a 5 kpc star-formation ring against a compact central injection | The ring alone cannot produce symmetric bubbles; a compact central source is required | Galactic Centre; the 5 kpc ring |
| Rajpurohit et al., the Peanut cluster 14 | MeerKAT and uGMRT imaging with Chandra and MUSTANG-2, 0.1 Gyr after pericentric passage | A 1 Mpc ultra-steep radio halo with emissivity an order of magnitude above the usual value, plus remnant AGN lobes 22 kpc off the brightest cluster galaxy | 400 MHz–1.28 GHz; z = 0.42 |
Paper briefs
A supernova wind carries neutral gas only where the cooling layers are resolved
Authors and date. Chong-Chong He, Mark R. Krumholz, Benjamin D. Wibking and Aditi Vijayan, submitted October 26, 2026; submitted to Monthly Notices of the Royal Astronomical Society. 1
Study design. The authors run simulations of supernova-driven winds with the GPU-accelerated code Quokka at one parsec resolution. They then repeat the same setup with two-parsec and four-parsec cells. Cloud sizes, masses and column densities of the neutral gas are compared with the HI clouds of the Milky Way's nuclear wind. 1
Result. At one parsec the winds are dominated by neutral mass, and the cloud properties land close to the observed nuclear-wind clouds. The gas does not coast on its own momentum. Drag against the surrounding medium, and hot gas condensing onto the clouds, drive it outward. 1
The neutral outflow weakens at two parsecs and disappears at four. The authors attribute part of that loss to cooling layers 3 to 6 pc thick, which coarser cells cannot hold in a few cells. 1
Why read. Observations put most of the mass flux within a few kpc of a disc in neutral gas. Simulations have struggled to reproduce that, which fed the argument that radiation pressure or cosmic rays must carry the bulk instead. This result turns the discrepancy into a resolution requirement: cells fine enough for the cooling layer, not a different driver. 1
Caveat. The comparison is internal to one code, one wind model and three resolutions. The authors write that resolution accounts for part of the difference, and they fit no single observed outflow. 1
A feedback scheme that leaves the hot phase where supernovae put it
Authors and date. Chengzhe Li, Hui Li, Wei Cui, Federico Marinacci, Laura Sales, Mark Vogelsberger and Paul Torrey, submitted October 27, 2026; in press at Astronomy & Astrophysics. 3
Study design. Galaxy-scale simulations lose the cooling radius around a young star cluster, so thermal energy is dumped as momentum instead. The authors anchor the host cell of each star particle so it cannot be displaced, keeping a thermal reservoir. Where the cooling radius is resolved, all supernova energy goes in as heat; where it is not, the host cell is refined and terminal momentum is deposited around it. 3
Result. Idealized tests reproduce high-resolution superbubbles, hot interiors included, without artificial cooling. In an isolated Milky Way-mass galaxy the scheme shows finer X-ray-emitting structure and a sharply defined, X-ray-bright mid-plane. The star formation history comes out smoother than with pure hybrid feedback. 3
Hot gas carries most of the outflow mass, with a mass-loading factor that plateaus near 0.2 above 1 kpc. Cooler phases decline with height, from radiative cooling and gravitational fallback. The wind that survives to large height is the energy-driven one. 3
Why read. Hot gas is the phase that carries energy out of a galaxy, and it is the phase a momentum-only recipe deletes. This gives the error a size: a mass-loading plateau at 0.2 and an X-ray-bright mid-plane that the same galaxy does not produce under hybrid feedback. 3
Caveat. The tests are idealized and the production run is a single isolated galaxy. No cosmological zoom-in is presented, so how much the scheme changes a halo's gas budget outside the disc is untested here. 3
In dwarfs, the halo's glow comes half from what falls in
Authors and date. Daniel R. Piacitelli, Alyson M. Brooks, N. Nicole Sanchez, Charlotte Christensen, Chloe Neufeld, Imad Pasha and six co-authors, submitted October 27, 2026; preprint with no journal reference listed. 4
Study design. The authors model the circumgalactic emission of dwarf galaxies in the Marvelous Massive Dwarfs suite. They predict surface brightness in Hα, [N II] λ6583 and [O III] λ5007, the lines MUSE, KCWI and the MOTHRA telephoto array can reach. They then relate brightness and extent to kinematics, recent star formation and galaxy size. 4
Result. The three lines usually trace different structures and phases. At a given galaxy mass, surface brightness and emitting-region size vary widely, and brightness and extent rise and fall together across the lines. Emission is typically eccentric and points along the galaxy's semi-minor axis. 4
Decomposed by origin, outflows contribute about 40 per cent of the total surface brightness across lines and masses, and inflows about 50 per cent. Brightness correlates positively with recent star formation. The correlation with HI disc size is negative and weaker. 4
Why read. Emission mapping is being sold as the way to see a halo's gas directly. This says what a detection would contain: for dwarfs, roughly as much infalling as outflowing gas, and light that cannot be separated into the two by brightness alone. 4
Caveat. Every number is a prediction of one simulation suite, so the 40/50 split inherits its feedback model. The authors also state that outflows are hard to identify in surface brightness maps even where they dominate the budget. 4
Simulated X-ray halos come out too bright, and missing metals may be why
Authors and date. N. Nicole Sanchez, Andrew Benson, Gwen Rudie, Ben Oppenheimer, Hannah Bish, Emily Silich, Phil Hopkins and two co-authors wrote part I, with John ZuHone and Michael Tremmel joining for part II. Both papers were submitted October 27, 2026, both to the Astrophysical Journal. 56
Study design. Part I takes more than 750 Romulus25 galaxies between 10 and 11.5 in log stellar mass and forward-models their 0.5–2 keV emission with pyXSIM and SIXTE. The stacked profiles and luminosities are then set against eROSITA stacking results and against EAGLE, TNG and SIMBA. 5
Part II uses the PIONEER zoom-ins, run with the same physics, to identify what sets the emissivity. 6
Result. The stacked profiles are too bright at every mass, worst in the inner halo. Average luminosities also exceed the observations, except at the Milky Way-mass end, where simulation and data agree. Romulus25 has no metal-line cooling, so the authors rerun the emission with a zero-metallicity model; average luminosity drops by nearly an order of magnitude and the agreement improves. 5
Part II finds that the mass of gas above 10⁵·⁵ K is set by the combined energy injected by stellar and AGN feedback. 6
At fixed hot gas mass, the mean metallicity of that gas changes the soft X-ray spectrum: metal-rich halos are brighter, metal-poor halos dimmer. Metal-rich halos may therefore bias X-ray observations of Milky Way-mass galaxies. 6
Why read. An emission model without metal lines overpredicts a Milky Way-mass halo's soft X-ray output by nearly an order of magnitude. 5
This pair puts that brightness in the gas metallicity and in the energy budget behind it, which is what a comparison with eROSITA stacks has to account for. 56
Caveat. Both papers rest on one simulation, and the zero-metallicity rerun is a stand-in for cooling physics the code does not carry rather than a physical model. The observational comparison uses stacked profiles, which mix halo masses and environments. 5
Cosmic rays are a candidate heat source for the halo, and the budget is tight
Authors and date. Roland M. Crocker and Mark R. Krumholz, submitted October 27, 2026; preprint released for community comment ahead of submission. 7
Study design. The authors build a semi-analytic model for cosmic rays escaping from star-forming discs into the circumgalactic medium, with transport and loss terms. 7
They confront it with four constraints at once: Fermi-LAT limits on the cosmic-ray energy density in the outer Milky Way, the isotropic gamma-ray background, stacked eROSITA X-ray emission from halos, and the requirement that the accumulated cosmic-ray pressure leave the halo stable. 7
Result. Only two scenarios satisfy all four. In the first, cosmic rays diffuse quickly and leave the halo for the intergalactic medium. In the second, their diffusion time out of the halo is of order the Hubble time. 7
In the second scenario, cosmic rays accelerated by star formation account for the gamma-ray background and for the heat needed to hold the halo's temperature against X-ray cooling. About half of the energy injected into the halo of an L★ galaxy over its star formation history would still be there at z = 0. 7
The authors argue for this scenario, because the alternative needs a second background source matching the discs in both power and spectral shape. 7
Why read. Halo heat is usually assigned to feedback energy with no carrier named. This names one and states how much of it must survive: half the injected energy, over the full star formation history, with the gamma-ray background as the price of the claim. 7
Caveat. The model is semi-analytic, so the transport coefficients are parameters rather than measurements, and the second scenario is preferred by argument rather than decidable by these constraints. The halo-stability condition is a plausibility bound, not an observed limit. 7
Early black hole growth changes which clumps leave the halo
Authors and date. Floor van Donkelaar, Lucas Tortora and Debora Sijacki, submitted October 29, 2026; preprint posted for comment before submission to Monthly Notices. 8
Study design. The authors compare two cosmological zoom-ins of the same massive halo above z = 6. The fiducial Fable run is set against Fable-SE, which adds mildly super-Eddington early black hole growth. Dense clumps in the circumgalactic gas are identified in both, then followed in mass, metallicity, stellar content and radius across cosmic time. 8
Result. At the resolution probed, the two runs give broadly similar clump gas masses, sizes and stellar masses. They differ in how many clumps there are, how metal-rich they are, and how they move. Fable-SE produces many more metal-enriched clumps, including a population moving outward during its strong AGN-driven blow-out below z ≈ 7. 8
Those outward clumps sit in disturbed, HI-rich, relatively dense gas, and some travel past the virial radius. Inward migration still dominates the population overall. 8
Why read. Inflow is the default assumption for dense high-redshift clumps, and the metal-rich ones are the easiest to see in absorption. This run says a burst of early black hole growth can reverse the direction for exactly those clumps, which changes what an absorber at z > 6 is evidence for. 8
Caveat. One halo, two runs, and the comparison is drawn at a stated resolution limit. Clump identification depends on the chosen density threshold, and the modified accretion model is a scenario rather than a measured accretion history. 8
For AGN feedback, where the energy goes matters more than how much
Authors and date. Kung-Yi Su, Hyerin Cho, Priyamvada Natarajan, Claude-André Faucher-Giguère, Ramesh Narayan, Angelo Ricarte, Antonio J. Porras-Valverde and two co-authors, submitted October 29, 2026; preprint with no journal reference listed. 9
Study design. The authors replace the usual fixed suppression of the Bondi accretion rate with a factor set by the ratio of Bondi to gravitational radius, taken from multi-zone GRMHD simulations that span the event horizon to the Bondi radius. 9
Outflow mass, kinetic energy and thermal energy are then distributed according to the angular structure those simulations measure. The model runs on isolated M87*-like and Sgr A*-like galaxies, across three decades of black hole mass, with no tuning of efficiency to observables. 9
Result. In the M87*-like galaxy, the angle-dependent model suppresses late-time stellar mass growth more effectively than the isotropic reference, at similar long-term black hole accretion rates. Changing the mass loading of the collimated polar component has only a modest effect. 9
The suppression factor itself varies with nuclear environment and time, so black hole growth histories differ markedly from fixed-suppression models. For Sgr A*, the variable factor gives an accretion rate closer to the Event Horizon Telescope constraints than a fixed, suppressed Bondi rate does. 9
Why read. Subgrid models usually carry two numbers, a Bondi suppression and a feedback efficiency, both tuned. This says the geometry of the outflow does work those numbers cannot, and that the uncertain mass loading is the less important of the missing pieces. 9
Caveat. Both galaxies are isolated and idealized, with no cosmological inflow, mergers or environment. The GRMHD angular structure is imposed as a subgrid mapping, so the result tests how much that structure matters rather than deriving it self-consistently. 9
Cool halo gas sits at the pressure of the hot gas around it
Authors and date. Suyash Kumar, Hsiao-Wen Chen, Gwen C. Rudie, Zhijie Qu, Sean D. Johnson, Alexander Beckett and eleven co-authors, submitted October 26, 2026; accepted for publication in the Astrophysical Journal. 10
Study design. The authors analyse HST/COS and STIS spectra of background quasars lying within the virial radius of 23 galaxies and galaxy groups at z = 0.4–0.7. Kinematically resolved absorption lines, run through photoionization models, yield density, temperature and non-thermal broadening for individual cool clumps. 10
Result. The clumps have hydrogen densities of 10⁻⁴ to 10⁻² cm⁻³, temperatures of 1 to 5 × 10⁴ K, and non-thermal broadening of 2 to 30 km/s. Photoionization equilibrium describes the cool phase well, and the turbulent motions are mostly subsonic. 10
At a fixed projected distance, cool-gas density falls with halo mass. In halos of 10¹² to 10¹³ M☉, the inferred cool-gas pressure matches the hot-gas pressure from stacked X-ray observations. 10
The total cool gas mass stays nearly constant across that range, so the cool phase's share of the baryon budget drops from 6–19 per cent at 10¹² M☉ to 0.6–1.5 per cent at 10¹³ M☉. 10
Why read. Pressure balance is what lets a cool cloud survive in a hot halo, and here it is measured rather than assumed. The same measurement says the cool phase holds a shrinking share of the baryons as halos grow, which is a statement about where the rest of the gas has to be. 10
Caveat. Twenty-three sightlines carry the result, and the densities and temperatures come from photoionization modelling rather than direct measurement. Several galaxies show low- and high-ionization species aligned in velocity across two orders of magnitude in density, which the authors read as local pressure fluctuation and which complicates any single-pressure reading. 10
After a starburst, the cool gas leaves the centre and spreads outward
Authors and date. Zoe Harvey, Sahyadri Krishna, Rita Tojeiro, Vivienne Wild, Jessica Nicole Aguilar, Steven Ahlen and 42 co-authors, submitted October 30, 2026; preprint with no journal reference listed. 11
Study design. The authors stack Mg II λλ2796, 2803 absorption from DESI Data Release 2 quasar spectra around luminous red galaxies at 0.4 < z < 0.76. 11
About 1.98 million sightlines crossed with galaxies between 10.6 and 12 in log stellar mass give roughly 143 million galaxy-quasar pairs, 50 to 60 times more per impact-parameter bin than their earlier SDSS-based analysis. Foreground galaxies are classified by mean-field independent component analysis, and the post-starburst subsample is split by time since the starburst. 11
Result. Absorption falls with impact parameter for every class. Post-starburst galaxies show the strongest of all: an equivalent width of 1.58 ± 0.32 Å within 71 kpc, against 0.69 ± 0.15 Å for star-forming and 0.22 ± 0.05 Å for quiescent galaxies of similar mass at the same separation. 11
Within the post-starburst phase the profile changes systematically with age. The youngest systems have the highest, most centrally concentrated absorption and the steepest decline. Older ones lose inner absorption and gain it at intermediate radii of about 300 to 1000 kpc. Independent star formation history fits confirm the ordering as an ageing sequence. 11
Why read. A starburst is the cleanest experiment feedback offers, because the energy injection stops and the response can be timed. This converts the sequence into a rate: cool gas expelled at a few hundred km/s clears the inner 71 kpc over the post-starburst phase, and a residual central reservoir stays. 11
Caveat. The ageing sequence comes from stacking different galaxies, not from following one halo, so it holds only if the post-starburst population is drawn from a comparable parent population at every age. Mg II traces cool, ionized gas, and the outflow velocity is inferred from the radial redistribution rather than measured in a spectrum. 11
The Milky Way's halo keeps the metals its galaxies made
Authors and date. Manami Roy, Sanskriti Das, Smita Mathur, Anjali Gupta and Yair Krongold, submitted October 29, 2026; submitted to the Astrophysical Journal. 2
Study design. Fast radio burst dispersion measures count all the free electrons along a sightline and carry no metallicity information. Oxygen absorption column estimates do carry metallicity, but depend on it. The authors combine the two for 25 sightlines, under three assumptions about the fraction of diffuse baryons in the intergalactic medium and intervening halos, and bootstrap the result. 2
Result. For diffuse baryon fractions of 0, 0.6 and 0.8, the median halo metallicity is 0.12 (+0.09/−0.04), 0.26 (+0.18/−0.12) and 0.39 (+0.26/−0.20) times solar. The last case produces super-solar values along several sightlines, which the authors judge implausible for the Milky Way's halo. 2
Even the conservative case, with no diffuse baryons outside halos, sits above what accretion of intergalactic gas would supply, at 0.1 Z☉ or less. The authors read that as efficient retention of metal-enriched feedback material in the halo. Metallicities also vary strongly between sightlines. 2
Why read. Metal retention is where feedback models disagree most: ejective models let the metals out, preventive ones hold them in. For the Milky Way this puts a number on retention without an X-ray spectrum, from data that are growing quickly. 2
Caveat. Twenty-five sightlines, a result that moves by a factor of three with the assumed diffuse baryon fraction, and a chain of inference through oxygen absorption estimates rather than a direct metallicity measurement. The sightline-to-sightline spread is real in the sample but its cause is not identified. 2
Radiation pressure survives a close look at one quasar outflow
Authors and date. Chen Chen, Wei-Jian Lu, Ying-Ru Lin and Bo Ma, submitted October 30, 2026; accepted for publication in Astronomy & Astrophysics, DOI 10.1051/0004-6361/202662070. 12
Study design. The authors observed the quasar SDSS J004613.54+010425.7 at z = 2.150 with VLT/X-shooter at medium resolution, as part of the TOLL project. Line locking is the signature of radiatively driven gas: absorption from one transition pushes gas to a velocity where the next transition can absorb, producing pairs at fixed velocity separations. 12
Result. The absorption complex resolves into twelve narrow components spanning about 2700 km/s across C IV, N V, Si IV and Lyα, 2.4 times as many as SDSS spectra showed. 12
Fourteen distinct line-locking pairs appear across three ions: three Si IV pairs, six N V pairs in four groups, and five C IV pairs in three groups. Monte Carlo trials put the probability of this web arising by chance below 10⁻⁵. 12
Cloudy modelling gives ionization parameters from 10⁻²·⁴ to 10⁻⁰·⁶ and hydrogen columns of 10¹⁸·⁴ to 10²⁰·⁶ cm⁻². Shadowed and shadowing clouds agree in column and ionization within a factor of about three, which broadly supports radiative acceleration, with a few outliers that suggest multi-ion locking or cloud-cloud interaction. 12
Why read. Line locking is the cleanest evidence that quasar outflows are accelerated by radiation rather than launched as a thermal wind. This object turns a two-component detection into a web, which is the kind of structure a radiative-driven model has to reproduce quantitatively. 12
Caveat. One sightline through one quasar, seen in projection, with the acceleration mechanism inferred from velocity coincidences. The authors note that medium- and high-resolution spectroscopy is what the measurement requires, which limits the sample to objects bright enough for it. 12
The eROSITA bubbles need a compact source, not a ring of star formation
Authors and date. Mukesh Singh Bisht and Kartick C. Sarkar, submitted October 29, 2026; submitted to the Astrophysical Journal as a letter. 13
Study design. A recent suggestion holds that the eROSITA bubbles need no central engine, only the star-forming ring at about 5 kpc in the Milky Way. The authors run hydrodynamic simulations of that ring, and of a compact central injection, and compare the bubbles each produces. 13
Result. The ring alone yields bubbles that lack the observed symmetry. A compact energy injection from the Galactic Centre is what restores it, and the authors conclude that the bubbles represent a nuclear outflow from the Galaxy. 13
Why read. Galactic feedback has two candidate engines separated by 5 kpc, and the answer decides whether the bubbles are evidence about the Galactic Centre's past activity or about the disc's star formation. This closes one of the two options on hydrodynamic grounds. 13
Caveat. The verdict rests on this simulation setup, and the abstract states the refutation without the quantitative comparison of bubble size, luminosity or limb brightness a reader would need to check it against the eROSITA maps. 13
A cluster caught 0.1 Gyr after pericentre keeps two kinds of radio gas
Authors and date. K. Rajpurohit, N. Lyskova, E. Churazov, W. Forman, I. Khabibullin, R. Kraft, L. Di Mascolo, T. Mroczkowski and 20 co-authors, submitted October 26, 2026; preprint with no journal reference listed. 14
Study design. The authors observed the Peanut cluster, SRGe CL0238.3+2005 at z = 0.42, with MeerKAT and uGMRT at several frequencies, and added Chandra and MUSTANG-2 data. The system is about 0.1 Gyr past pericentric passage, early enough for a merger's effect on the intracluster medium to be dated. 14
Result. Two spectrally distinct radio structures appear. In the northern subcluster, a double-lobed source with a strongly curved, ultra-steep spectrum sits about 22 kpc from the nearest brightest cluster galaxy, which the authors read as remnant AGN plasma displaced by the merger or reshaped by a shock. 14
Across the cluster centre lies a radio halo about 1 Mpc across with spectral index −1.77 ± 0.11 between 400 MHz and 1.28 GHz. Its volume-averaged emissivity at 1.4 GHz, (1.1 ± 0.1) × 10⁻⁴¹ erg s⁻¹ cm⁻³ Hz⁻¹, is an order of magnitude above the characteristic value for other halos. 14
Radio emission follows the X-ray emission with a slope of 1.11 ± 0.05. Candidate X-ray and Sunyaev-Zel'dovich discontinuities coincide with a flatter spectral index, which may trace local particle re-acceleration. 14
Why read. Radio halos are usually studied in relaxed or late-stage mergers, where the history is hard to date. At 0.1 Gyr past pericentre both the fossil plasma and the freshly re-accelerated population are present in one system, which is a test of whether a shock can relight what an earlier AGN episode left behind. 14
Caveat. One cluster at one dynamical stage, with the remnant-plasma reading built on spectral curvature and a positional offset. The discontinuities are described as candidates in the abstract, and no radio halo of comparable brightness exists in the comparison samples the emissivity is measured against. 14
Code and data releases
No open-source release was verified for this window. GitHub searches over repositories pushed or created between October 26 and November 2 returned no new release relevant to this subject. The search used circumgalactic, intracluster, galactic outflow, AGN feedback, stellar feedback and galaxy-formation keywords.
The closest item is a personal fork of yt and trident labelled "Circumgalactic Dictionary" 15. It carries no licence and no tagged release, and no paper in the window announces it, so it is not a package a reader can adopt.
No dataset deposit qualified either. Zenodo records deposited between October 26 and November 2 under these keywords were conference presentations from a workshop on AGN feedback and outflows, filed as presentations rather than data 16.
The one dataset-type deposit in the window, input density fields and halo catalogues for a reionization radiative-transfer run, is restricted and concerns reionization rather than feedback 17. Both categories therefore stay empty.
What the week separates
- Whether a wind carries neutral mass is now a question about resolution. A one-parsec simulation produces a neutral-dominated outflow with clouds like the Milky Way's; at four parsecs nothing neutral is left 1. A feedback scheme that pins the host cell of each star particle and refines it when the cooling radius is unresolved recovers hot superbubble interiors that momentum-only recipes delete 3. Forward-modelled X-ray emission from a simulation without metal-line cooling comes out an order of magnitude too bright, and dropping the metals fixes it 5. Three different answers, one shared cause: what the model leaves unresolved or unrepresented decides the result.
- Halo heat has at least three candidate sources, and each paper rules the others in or out by argument. Cosmic rays diffusing out over a Hubble time would supply the heat that balances X-ray cooling, at the cost of the gamma-ray background 7. In simulated L★ halos, the mass of hot gas is set by the total energy of stellar plus AGN feedback, and its metallicity then sets how bright it looks 6. Cool clumps in group halos sit at the pressure of that hot gas, so the two phases are in contact rather than merely coexisting 10.
- The metals that feedback expels mostly stay inside the halo, and move outward within it. Twenty-five FRB sightlines put the Milky Way halo at 0.12 to 0.39 Z☉, above the accretion value 2. DESI stacks show post-starburst halos with the strongest cool gas of any class, which then migrates from inside 71 kpc out to 300–1000 kpc as the starburst ages 11. At z > 6, a burst of early black hole growth sends metal-rich, HI-rich clumps outward, some past the virial radius 8. Metals leave the galaxy, and then stop.
- AGN feedback's direction matters more than its strength. The angular structure of the accretion flow changes late stellar growth more than the mass loading of the polar outflow does 9. A merger displaces a dead jet's lobes 22 kpc from the galaxy that made them, while a shock may relight the cluster's diffuse radio emission 14. A ring of star formation 5 kpc from the Galactic Centre cannot reproduce the eROSITA bubbles, which need a compact central source 13. In each case the question is not how much energy was injected, but from where and along what axis.
Also noted
Six papers in the window bear on the topic without meeting the bar for a brief.
Turbulent box simulations with accreting black holes find that without turbulence, jet power stalls at 10³⁹ erg/s for a 10⁶ M☉ black hole. Above a critical jet efficiency the jet depletes its own fuel supply, and the threshold is set by the interstellar Mach number 18.
A criterion for seeing X-ray cavities at high redshift finds that the atmosphere of SPT-CL J2215−3537 needs 0.6–1.2 per cent of its pressure in radiating leptons to fill the soft-band deficit. The required fraction scales as (1+z)⁻⁴ 19.
The OASIS II survey measures NIRSpec/PRISM spectra of 54 galaxies at z > 4 and finds quiet phases lasting about 100 Myr, with faster recovery. The [O III]/[O II] ratio follows the recent change in star formation rate more closely than the distance below the main sequence 20.
Combining Faraday rotation stacks with fast radio burst dispersion and kinetic Sunyaev-Zel'dovich data gives 95 per cent upper limits on the line-of-sight magnetic field in groups and clusters. The authors forecast a first detection with the Deep Synoptic Array 21.
Star formation is detected in two galaxies 66 and 137 kpc from a chemically near-pristine damped Lyα system at z = 3.078 22.
A FIRE-3 analysis shows that the helium-to-metal enrichment slope has no universal value: it ranges from 0.75 to 2.1 with stellar age, and varies between galaxies 23.
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.