Thirteen new feedback tests, from a galaxy nearly free of metals to halo heat that does not scale
Thirteen new arXiv results from October 12 to 16, 2026 weigh halo heat against halo mass, follow quasar winds that stay inside their own escape velocity, and measure the most metal-poor gas ever found in a galaxy.
The week of October 12 to 18, 2026 produced thirteen new results on the gas inside, leaving and returning to galaxy halos, and one public code release built for testing simulations against quasar absorption spectra. A supporting-data deposit for the thermal Sunyaev-Zel'dovich measurement arrived mid-week as well. The arXiv listing at this hour reaches submissions logged through Friday, October 16. Weekend papers are announced on Monday evening and were not screened, so this edition rests on the window's first five days.
Two results frame the week. A damped Lyman-α system seen against a background quasar holds neutral hydrogen with a carbon-to-hydrogen ratio 11,000 times below the Sun — the most metal-poor gas ever measured in a galaxy 1. At the other end of cosmic time, a thermal Sunyaev-Zel'dovich stack over seven million galaxies finds the circumgalactic medium holding less thermal energy than a self-similar halo would predict, by 7.3σ 2.
An item earns a place below when the original reports a quantitative result on gas inside, leaving or returning to a galaxy halo or the intergalactic medium. It must also state a method another group can repeat and link to a public source.
At a glance
| Item | Design / method | Main result | Scale or tracer |
|---|---|---|---|
| Das et al., tSZ in the CGM III 2 | ACT + Planck thermal SZ stacked on DESI Legacy Survey galaxies and groups, binned by mass and by richness | One-halo term detected to 10¹²·² M☉ at 5.8σ; some bins need flatter or miscentred pressure profiles; super-virial floors of ≈2–7 T₂₀₀c imply heating; CGM thermal energy falls below self-similarity by 7.3σ | 7 million galaxies and groups, M₂₀₀c = 10¹²–10¹⁴ M☉, z ≤ 1 |
| Ondaro-Mallea et al., stacked kSZ II 3 | FLAMINGO hydrodynamical simulations with DESI-like luminous red, bright and emission-line galaxy samples, decomposed into central and satellite contributions | Satellites are 20–30 per cent of each sample but 50–80 per cent of the stacked signal; ignoring them biases inferred gas fractions low by up to a factor of two and overstates feedback strength | Halo scales probed by DESI × ACT kSZ stacking |
| Kim & Kim, hot CGM X-rays 4 | TNG100-2 at z = 0, 0.1 and 0.2, summing 0.5–2 keV emission from gas above 10⁶ K between 0.15R₅₀₀ and R₅₀₀, split by star-formation activity | At M₅₀₀ = 10¹¹·⁷⁵–10¹²·²⁵ the star-forming sample is about five times more luminous and more centrally concentrated than the quiescent one, with similar median hot-gas mass | Milky Way- to group-mass halos in one simulation |
| Li et al., Orion A HI bubble 5 | FAST HI survey of Orion A, resolved into an expanding HI bubble enclosing the Veil ionised bubble, with expansion speeds measured for both drivers | Photoionisation-driven gas expands at 10.9 km/s and wind-driven gas at 13.8 km/s; the kinetic energy ratio is 1:2, so stellar winds carry a comparable share and radiation does not dominate | A single star-forming region, tens of parsecs across |
| Matsui et al., dwarf satellite energy budget 6 | Order-of-magnitude energy accounting for nine Milky Way dwarf spheroidals from published star-formation histories, velocity dispersions, radii and V-band luminosities | Under the K+B16 prescription all nine reach the supernova-energy threshold, yet star formation often continues long after; the threshold depends on halo mass, radius, retained gas fraction and coupling efficiency | Nine dwarf spheroidals inside the Milky Way halo |
| Li et al., NGC 3783 Fe XXV wind 7 | XRISM/Resolve spectra across a continuous 10-day observation, with the Fe XXV trough divided into five velocity intervals and 19 time-resolved spectra | Three intervals vary; the −320 to −680 km/s component responds inversely with a delay and lies within 0.05 pc, below local escape velocity — a failed wind — while the −680 to −1040 km/s component sits near the escape threshold | Nuclear wind within 0.25 pc of the black hole |
| Domínguez, gamma-ray calorimetry of Seyfert winds 8 | Published Fermi-LAT stacking measurements homogenised to 0.1–100 GeV on one γ-ray luminosity versus mechanical power plane | Seyfert winds reach 0.020 of the maximum calorimetric efficiency, 13.5 times below starburst superwinds at 0.26; the gap exceeds a factor of three at 95 per cent confidence | Radio-quiet Seyferts, LINERs and starburst superwinds |
| Gray et al., Abell 1795 ¹³CO 9 | ALMA ¹³CO(2–1) observations of the northern molecular filament, with a range of excitation temperatures and covering factors | The filament holds (0.1–2) × 10⁹ M☉, up to an order of magnitude below earlier estimates, giving α_CO of 0.4–5; direct entrainment by the N radio bubble becomes plausible | Filaments of 5–7 kpc in a cool-core cluster |
| D'Onofrio et al., SQuIGGLE radio AGN 10 | VLA 6 GHz continuum imaging of 52 massive post-starburst galaxies, compared with mass- and redshift-matched LEGA-C quiescent galaxies | 18 of 52 are detected at 10²³–10²⁵ W/Hz, concentrated in the first ≈200 Myr after the burst and in galaxies that kept their gas; the detection fraction is 2.5 times the LEGA-C value | z ≈ 0.7 post-starburst population |
| Bosman et al., UMP-DLA-1 1 | GHOST high-resolution spectroscopy of a background quasar, with carbon and oxygen abundances measured in the foreground damped Lyman-α absorber | C/H is 11,000 times below solar, the most pristine gas known by a factor of four; carbon and oxygen are too low for efficient formation of stars at or below a solar mass | A neutral hydrogen reservoir about 1 Gyr after the Big Bang |
| Gasymov et al., CoRGI in MaNGA II 11 | MaNGA integral-field spectroscopy, CFHT imaging, H I data and SDSS/DESI spectroscopy of the surrounding group | A counter-rotating disk 1.3 Gyr old, with [Fe/H] ≈ −0.7 and scale length 2.4 kpc, sits inside an old spheroid of log M★ ≈ 9.9; chemical modelling favours accretion from a cosmic filament | A brightest group galaxy in a sparse group |
| Oman et al., COLIBRE H I 12 | (200 cMpc)³ cosmological volume with non-equilibrium chemistry and dust, and no H I observable used in calibration | Reproduces the M★–M_H I relation, the H I mass and velocity-width functions and the mass–size relation; large-radius H I around M★ ≳ 10¹⁰ galaxies may be spurious clouds, and the highest-resolution run overproduces H I sources by ≈2 | Galaxies at z = 0 in a 200 Mpc box |
| Lu et al., magnetic fields in FIRE-2 13 | FIRE-2 magnetohydrodynamical simulations of Milky Way-mass and dwarf galaxies, run with and without cosmic rays | Cosmic rays barely change the field; halo mass sets the saturated strength; every run reaches a magnetic-to-turbulent ratio of 1–10 per cent through expansion, amplification and saturation | Halo gas from galactic disks to the virial radius |
Paper briefs
The circumgalactic medium holds less heat than a self-similar halo would
Authors and date. Sanskriti Das, Yi-Kuan Chiang and Smita Mathur, submitted October 15, 2026; accepted for publication in the Astrophysical Journal. 2
Study design. The authors stack ACT and Planck thermal Sunyaev-Zel'dovich maps on seven million DESI Legacy Survey galaxies and groups between 10¹² and 10¹⁴ M☉ at z ≤ 1. They then split the sample by richness, which had not been done before, and fit a generalised Navarro-Frenk-White pressure profile to each bin. 2
Result. The one-halo term is detected down to 10¹²·² M☉ at 5.8σ. Several mass-richness bins need a flatter or miscentred profile, and the pattern of that preference changes with richness. In some bins the lowest temperature the data allow sits at 2–7 times the virial temperature, which requires heating. Kinetic-mode feedback is not essential in any bin. In the circumgalactic regime the thermal energy–mass relation falls below self-similarity by 7.3σ, while the intragroup regime stays consistent with it. 2
Why read. This measures thermal pressure rather than inferring it, and it measures it for millions of halos at once. The conclusion that low-mass, low-richness halos dominate the thermal energy budget at z ≤ 1 is a statement about where the energy sits, not about how much was injected. 2
Caveat. Profile shape and temperature floor are fitted, and the temperature floor is the lowest value the data permit rather than a detection. The signal is stacked, so no individual halo is resolved. 2
Half the stacked kSZ signal may be satellites, not feedback
Authors and date. Lurdes Ondaro-Mallea, B. Hadzhiyska and Raúl E. Angulo, submitted October 12, 2026; preprint. 3
Study design. The authors build DESI-like luminous red, bright galaxy survey and emission-line galaxy samples in the FLAMINGO simulations and stack the kinetic Sunyaev-Zel'dovich signal around them. They then separate the contribution of central and satellite galaxies within each stack. 3
Result. Satellites make up 20 to 30 per cent of each sample and 50 to 80 per cent of the signal, because they occupy more massive halos than centrals. Both satellites and feedback suppress the small-scale signal, and stacking on satellites mimics the extended gas distribution that feedback produces. At the precision of current DESI × ACT data, ignoring satellites biases the inferred gas fraction low by up to a factor of two and makes feedback look stronger, most severely for the luminous red and bright galaxy samples. 3
Why read. The kSZ stack is one of the few direct handles on gas mass around galaxies, and this paper says what the stack cannot separate on its own. It also sets the price of the fix: the satellite fraction must be known to better than 10 per cent, from lensing and clustering of the same sample. 3
Caveat. The degeneracy is quantified in simulations, so its size depends on how FLAMINGO places satellites and on its feedback prescriptions. The paper stops short of a data analysis; it states what a future one must control. 3
Soft X-rays see how the hot halo is arranged, not how much gas it holds
Authors and date. Minjae Kim and Junhan Kim, submitted October 12, 2026; submitted to the Astrophysical Journal. 4
Study design. The authors compute intrinsic 0.5–2 keV emission from gas above 10⁶ K in TNG100-2 at z = 0, 0.1 and 0.2. The emitting shell runs from 0.15R₅₀₀ to R₅₀₀, and the galaxies are split by star-formation activity. 4
Result. Luminosity rises with halo mass. Between M₅₀₀ of 10¹¹·⁷⁵ and 10¹²·²⁵ the star-forming sample has a median luminosity about five times the quiescent sample's and a more centrally concentrated profile, while the two have similar median hot-gas masses. The difference is in the distribution and thermodynamic state of the gas, not in how much of it there is. 4
Why read. It predicts that two halos of the same mass can differ fivefold in soft X-rays while holding the same hot gas. Anyone using X-ray luminosity as a proxy for hot-gas mass needs that factor in the error budget before the observation is made. 4
Caveat. These are intrinsic luminosities from one simulation box, with no background, absorption or instrumental response applied. The samples are not matched in halo mass, and the authors say so. 4
In one bubble, the wind carries as much energy as the radiation
Authors and date. Chong Li, Rou-Yu Li, Ke-Ping Qiu and 27 co-authors, submitted October 12, 2026; published in Science Advances 12, eaec0787. 5
Study design. A FAST survey of Orion A resolves a heart-shaped expanding H I bubble that encloses the known Veil bubble of ionised gas. The ionised bubble traces photoionisation-driven expansion and the neutral shell traces the wind-driven part, so the two can be measured in the same object. 5
Result. The photoionisation-driven gas expands at 10.9 km/s and the wind-driven gas at 13.8 km/s. Under spherical symmetry and a uniform H I to H₂ ratio, the kinetic energy ratio between them is 1:2. Stellar winds therefore contribute a comparable share of the injected kinetic energy, and radiation does not dominate as the classical picture predicts. 5
Why read. The split between photoionisation and winds is a standard free choice in feedback recipes. This is the first direct comparison in a single bubble, and it lands on the side that gives winds more weight. 5
Caveat. The energy ratio rests on assumed spherical symmetry and a uniform H I to H₂ ratio across the shell. One region supplies the measurement, and the masses come from the same H I map that measures the velocities. 5
Enough supernova energy to empty a dwarf is not the same as emptying it
Authors and date. Sena A. Matsui, Tsutomu T. Takeuchi, Suchetha Cooray and K. Omori, submitted October 15, 2026; preprint. 6
Study design. For nine Milky Way dwarf spheroidals the authors take published colour–magnitude star-formation histories, velocity dispersions, effective radii and V-band luminosities. They define an energy-equivalence time τ_E at which cumulative core-collapse supernova energy first equals the adopted gas binding energy. 6
Result. Under the fiducial kinematic–size prescription K+B16, all nine galaxies cross the threshold. Star formation then often continues long after that crossing. The lookback time of the crossing sits close to the lookback time by which 90 per cent of the stellar mass had formed, but the authors read that proximity as descriptive, since both times come from the same history. 6
Why read. Dwarf satellites are the usual evidence that supernovae can remove a galaxy's gas. This states precisely what such an energy argument establishes — that the energy was available — and separates it from the timing question, which needs a different measurement. 6
Caveat. The test is order-of-magnitude, and the authors call it an energetic diagnostic rather than a measurement of wind launch. Whether a galaxy crosses the threshold depends on its halo mass, its radius, the retained gas fraction and the supernova coupling efficiency, none of which is measured here. 6
A quasar wind caught below its own escape velocity
Authors and date. Chen Li, J. Kaastra, Li-Yi Gu and seven co-authors, submitted October 13, 2026; accepted for publication in Astronomy & Astrophysics. 7
Study design. XRISM/Resolve observed NGC 3783 continuously for ten days. The authors divide the resolved Fe XXV absorption trough into five fixed velocity intervals and measure their equivalent widths in 19 time-resolved spectra, then model the response to the ionising continuum. 7
Result. Three intervals vary significantly. The low-velocity component at −320 to −680 km/s responds inversely and with a delay, which implies a high ionisation parameter, a high density and a distance within 0.05 pc. The high-velocity component at −680 to −1040 km/s responds positively with a longer lag, is less dense, and lies within 0.25 pc. The low-velocity component travels below the local escape velocity: a failed wind. 7
Why read. An absorption line's blueshift says how fast gas moves, not whether it leaves. Response times turn that into a distance, and the distance decides escape. Together with a lower-ionisation failed wind seen in 2001, the picture is a multiphase wind that stays bound over two decades. 7
Caveat. Time-dependent photoionisation models cannot reproduce the largest variation, at 3.5σ, so the densities rest on a high-density limit rather than a fit, and the distances are upper limits. 7
Seyfert winds convert mechanical power into gamma rays 13.5 times less efficiently than starbursts
Authors and date. Alberto Domínguez, submitted October 16, 2026; accepted for publication in Astronomy & Astrophysics. 8
Study design. The author homogenises two published Fermi-LAT stacking detections — one selected by X-ray ultra-fast outflows, one by hard X-ray brightness — to a common 0.1–100 GeV band, adds a stack of 186 low-luminosity nuclei, and places all of them on one plane of γ-ray luminosity against mechanical power. 8
Result. Expressed as a fraction of the maximum calorimetric efficiency of about 2.5 per cent, Seyfert winds reach 0.020 (+0.018/−0.011) against 0.26 (+0.37/−0.16) for starburst superwinds, a factor of 13.5. The gap exceeds a factor of three at 95 per cent confidence and does not depend on acceleration parameters common to both classes. Implied ambient densities are about 10 and 100 cm⁻³. 8
Why read. It converts "do AGN winds matter" into a comparison of how much of the mechanical energy survives as observable GeV emission. On this accounting LINER outflows fail on power, Seyfert ultra-fast outflows fail on calorimetry, and starburst superwinds pass both tests. 8
Caveat. No new data are reduced; the numbers come from homogenising published measurements, and the comparison is conditional on the assumed wind kinetic powers. Density, shock radius and shock speed are degenerate, so the densities are consistency values rather than measurements. 8
The cold gas a radio bubble lifts is ten times lighter than estimated
Authors and date. J. Gray, H. Russell, F. Combes and six co-authors, submitted October 15, 2026; accepted for publication in Monthly Notices of the Royal Astronomical Society. 9
Study design. ALMA observed the northern molecular filament of Abell 1795's central galaxy in ¹³CO(2–1), the low-opacity tracer that gives a more reliable mass than ¹²CO in optically thick gas. The authors compare its velocity structure with the existing ¹²CO map and vary excitation temperature and covering factor. 9
Result. The filament's velocity structure matches the ¹²CO emission. Its mass is (0.1–2) × 10⁹ M☉, down to an order of magnitude below previous estimates, which implies a CO-to-H₂ conversion factor of 0.4–5 M☉ (K km s⁻¹ pc²)⁻¹, or 0.1 to 1 times the Milky Way value. Direct entrainment by the N radio bubble can now supply the required energy and buoyant support, and measured X-ray cooling rates are consistent with condensation of uplifted low-entropy gas. 9
Why read. Uplift by radio bubbles is a standard way to explain molecular filaments in cool cores, and its feasibility depends on how much gas has to be lifted. That mass has just fallen by an order of magnitude. 9
Caveat. The quoted mass range spans a factor of twenty, set by the allowed excitation temperatures and covering factors. The authors note that both formation routes need several generations of bubbles if the filament mass is at the high end. 9
Weak radio jets appear exactly where the leftover gas is
Authors and date. V. D'Onofrio, J. Spilker, K. Smith and 13 co-authors, submitted October 13, 2026; submitted to the Astrophysical Journal. 10
Study design. The authors image 52 massive post-starburst galaxies at z ≈ 0.7 with the VLA at 6 GHz, at subarcsecond resolution, and compare the detection rate with mass- and redshift-matched quiescent galaxies from LEGA-C above a common luminosity threshold. 10
Result. Eighteen of the 52 are detected, about 35 per cent, at 1.4 GHz luminosities of 10²³ to 10²⁵ W/Hz, generally above what star formation and the fading burst can supply. Radio activity concentrates in the first ≈200 Myr after the burst and in galaxies that retain substantial gas reservoirs. Above a common threshold the detection fraction is about 2.5 times that of the LEGA-C quiescent sample, driven by lower-power sources. 10
Why read. Post-starburst galaxies keep their molecular gas and stop forming stars anyway, which needs a mechanism that couples to cold gas. This finds radiatively inefficient, mechanically powered nuclei in that population, at the ages where the gas is still there. The survey's cool-gas outflow side was covered in the September 28 edition; this is the radio counterpart. 10
Caveat. The 1.4 GHz luminosities follow from assumed radio spectral indices, and the mechanical powers are inferred rather than measured. The claim that jets could unbind the gas holds only if jet–ISM coupling is efficient. 10
The most metal-poor gas ever seen in a galaxy cannot make sun-like stars
Authors and date. S. Bosman, F. Davies, Louise Welsh and 22 co-authors, submitted October 12, 2026; accepted for publication in Nature. 1
Study design. The authors searched for ultra-metal-poor galaxies through their absorption against bright background quasars, using the GHOST high-resolution spectrograph on Gemini South. They then measured carbon and oxygen abundances in the neutral hydrogen of the absorber UMP-DLA-1. 1
Result. The carbon-to-hydrogen ratio is 11,000 times below solar, making this the most pristine gas measured anywhere by a factor of four. Both carbon and oxygen are too low for efficient formation of stars at or below a solar mass, so the low metallicity alone guarantees a top-heavy stellar population. The oxygen-to-carbon ratio is elevated, consistent with the yields of first- or early second-generation stars. 1
Why read. Carbon and oxygen line cooling is what lets a gas cloud collapse into ordinary stars, and this system is the first with abundances measured below the level where that works. It shows the abundance pattern the earliest stellar populations left behind, a billion years after the Big Bang. 1
Caveat. One absorber on one sightline carries the measurement. The abundances come from absorption with ionisation corrections, and the top-heavy population is inferred from the abundances rather than observed. 1
Cold gas still reaches the centre of a massive group galaxy
Authors and date. D. Gasymov, I. Katkov, E. Grebel and six co-authors, submitted October 14, 2026; submitted to Astronomy & Astrophysics. 11
Study design. The authors combine MaNGA integral-field spectroscopy, CFHT imaging, existing H I data and SDSS/DESI spectroscopy of the surrounding group to decompose the kinematics of NGC 5717, a brightest group galaxy hosting a counter-rotating stellar disk. 11
Result. A young disk, about 1.3 Gyr old, metal-poor at [Fe/H] ≈ −0.7, extended with a scale length of 2.4 kpc and of log M★ ≈ 9.9, counter-rotates inside an old, dynamically hot spheroid. The ionised gas co-rotates with the disk, and the available H I data show no significant remaining reservoir. Chemical modelling favours accretion from a cosmic filament, with bursty star formation and metal-enriched outflows. 11
Why read. Counter-rotating disks are usually found in sparse environments, where accreted gas survives long enough to settle. Finding one in a group's brightest galaxy tests whether that channel can still deliver cold gas in a denser setting. 11
Caveat. One galaxy, and the accretion origin is inferred from kinematics and chemistry rather than traced as an inflow. The filament geometry is inferred from the surrounding group's configuration. 11
A simulation never tuned to H I lands on the H I relations anyway
Authors and date. Kyle A. Oman, Ruby J. Wright, Andrea Gebek and 12 co-authors, submitted October 13, 2026; submitted to Monthly Notices of the Royal Astronomical Society. 12
Study design. The authors measure the atomic hydrogen properties of galaxies at z = 0 in a (200 cMpc)³ COLIBRE volume with 2 × 10⁶ M☉ resolution elements. The model follows non-equilibrium chemistry coupled to dust and radiative cooling to about 10 K, and none of its parameters was calibrated on an H I observable. 12
Result. The simulation reproduces the stellar mass–H I mass relation, the H I mass and velocity-width functions, the H I mass–size relation and the baryonic Tully-Fisher relation, with mismatches at the extremes. H I surface density profiles are realistic except at large radii around galaxies above 10¹⁰ M☉, where a population of possibly spurious circumgalactic H I clouds appears and shrinks as resolution improves. The highest-resolution run overproduces H I sources by about a factor of two above 3 × 10⁹ M☉. 12
Why read. H I is often used to check star-formation and feedback models that were tuned elsewhere. This is that check run in the honest direction: no H I calibration, then a comparison, then an account of where the model fails. 12
Caveat. One volume, and the resolution study itself changes the answer, which means part of the disagreement at large radii is numerical. The mismatches at the low-mass and low-velocity extremes are flagged by the authors rather than resolved. 12
Halo magnetic fields are an output of feedback, not an input
Authors and date. Yue Samuel Lu, D. Kereš, Phil Hopkins and Sam B. Ponnada, submitted October 12, 2026; submitted to Monthly Notices of the Royal Astronomical Society. 13
Study design. The authors follow magnetic fields through FIRE-2 magnetohydrodynamical simulations of Milky Way-mass and dwarf galaxies, run with and without cosmic rays. They track the field from its initial seed through structure formation, feedback and turbulence to z ≈ 0. 13
Result. Cosmic rays barely change the field. Halo mass sets the saturated field strength at z ≈ 0, while the ratio of magnetic to turbulent energy reaches 1 to 10 per cent in every run, whatever its halo. The field passes through expansion, amplification and saturation, with flux freezing driving the first and last and giving B ∝ ρ²ᐟ³ and B ∝ ρ¹ᐟ²; the intermediate amplification is fuelled by structure formation, feedback and gas turbulence. 13
Why read. Magnetic fields determine how cosmic rays propagate through the halo, and they are usually set by hand. This makes them a consequence of the same feedback that moves the gas, with a stated saturation level that future CGM models can be checked against. 13
Caveat. The magnetic-to-turbulent ratio depends on how turbulence is measured, which the authors flag. The middle amplification stage is conjectured to be numerically triggered rather than derived. 13
Code and data releases
specwizard, a public synthetic-spectrum framework for hydrodynamical simulations. The code behind this week's specwizard paper is public at github.com/specwizard/specwizard under the GNU GPL v3, with documentation at specwizard.readthedocs.io, and took a push on October 15, the day after the paper appeared. 14 It combines NIST atomic data with Cloudy ionisation fractions across several ionising backgrounds, projects simulation gas onto sightlines while conserving column density, and adds natural broadening, instrumental resolution and noise. 15 The paper's test is a z = 3.2 sightline through EAGLE that reproduces the mean Lyman-α transmission, its probability distribution and its power spectrum, and then recovers the evolution of the ionising background from the temperature–density relation. 15
Supporting data for the thermal SZ measurement in this issue. Zenodo record 10.5281/zenodo.23389237 holds the material behind the circumgalactic thermal Sunyaev-Zel'dovich paper above: a spreadsheet of the key results in its Figures 5 and 6, an archive for reproducing the stacked Compton-y radial profiles, and an archive of the cosmic infrared background systematic test. 16 The record is open, carries a CC-BY-4.0 licence, and was deposited on October 15, inside this window. 16
What the week separates
- Halo heat is now measured well enough to fail the simple scaling. The thermal SZ stack puts circumgalactic thermal energy 7.3σ below self-similarity and finds super-virial temperature floors in some mass bins, which requires heating without requiring kinetic-mode feedback anywhere 2. The TNG100 prediction says a star-forming halo of a given mass can be five times brighter in soft X-rays than a quiescent one holding the same hot-gas mass 4. The kinetic SZ paper then shows that up to half of what such a stack measures is satellite placement, not feedback 3. Read together, the signal is solid and its interpretation runs through gas distribution rather than gas mass.
- Energy budgets come out even; the timing does not. FAST resolves one bubble where wind-driven gas carries twice the kinetic energy of the photoionisation-driven gas 5. Nine dwarf satellites all cross the supernova energy threshold, and several keep forming stars afterwards 6. An energy argument can show that feedback was sufficient. Neither paper shows that it acted when the gas left, and the dwarf paper says explicitly that its two timescales are not independent evidence.
- Black-hole feedback is measured four ways, and the ranking follows the channel rather than the power. Seyfert winds convert mechanical power to GeV emission 13.5 times less efficiently than starburst superwinds, and the ultra-fast-outflow winds fail on efficiency while LINERs fail on power 8. In NGC 3783 the low-velocity component moves below the local escape velocity and sits within 0.05 pc of the nucleus, so that gas stays bound 7. In Abell 1795 the molecular mass that a radio bubble must lift drops by an order of magnitude, which makes direct entrainment viable again 9, and the radio nuclei found in young post-starburst galaxies sit where the leftover gas is 10. Energy delivered, energy retained and energy coupled are three different numbers, and this week measures a different one in each paper.
- Calibrating on one observable leaves the rest free, and absorption spectra are where the next test runs. COLIBRE was never tuned on H I and still reproduces the H I relations, while showing where its own resolution creates circumgalactic clouds that may not be real 12. FIRE-2 makes halo magnetic fields a by-product of feedback, with cosmic rays mattering little and halo mass setting the saturation 13. specwizard turns such simulation output into a synthetic quasar spectrum that can be compared with the same high-resolution absorption data that found the pristine gas above 115.
Also noted
Four papers in the window did not get a brief. A step-function quenching model fitted to about 350,000 galaxies infers a characteristic gas-supply timescale near 1 Gyr and separates inside-out from outside-in quenching 17. eROSITA spectra of 140 nearby galaxies keep the thermal X-ray luminosity correlated with star formation down to 10⁻⁵ M☉ per year, while the total luminosity flattens at low rates as X-ray binaries take over 18. A JWST review of outflows at cosmic noon reports neutral outflows in 25 to 50 per cent of massive galaxies, mass-loading factors of 1 to 10 in recently quenched systems, and outflow speeds mostly below the halo escape velocity 19. The ALMA2040 white-paper series is also arriving in this window, including a case for mass-complete surveys of young massive clusters and their early feedback 20; these set future observing cases rather than reporting new measurements, which is why none is briefed above.
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.