Survival and synthetic observables of neutral atomic hydrogen in galactic wind simulations

Andrei Antipov; Daniel Villarruel; Wladimir E. Banda-Barrag\'an

REVIEW 4 major objections 5 minor 16 references

Survival and synthetic observables of neutral atomic hydrogen in galactic wind simulations

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Galactic-wind HI is reborn from recycled gas, not surviving primordial clouds.

desk verdict A well-written IAU proceedings summary that points to real companion papers, but the headline recondensation claim is not backed by diagnostics shown here. read the letter →

arxiv 2412.08491 v1 pith:MDVVOBS6 submitted 2024-12-11 astro-ph.GA

Wladimir E. Banda-Barragán , Andrei Antipov , Daniel Villarruel This is my paper · ORCID

classification astro-ph.GA

keywords galacticwindsneutralhydrogenHIspectrallinesmagnetohydrodynamicsrecondensationmagneticdrapingsyntheticspectramultiphaseinterstellarmedium

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Radiative magnetohydrodynamic simulations of cloud systems embedded in hot, fast outflows show that the dense neutral atomic hydrogen (HI) seen in galactic winds need not be primordial cloud gas that survived ablation. Instead, the paper argues, HI is produced self-consistently: shock-shredded cloud material mixes with the hot wind, and the thermally unstable mixed gas cools and recondenses into fast-moving, filamentary HI cloudlets. In single-cloud wind models, magnetic fields control both morphology and synthetic observables: a magnetic field transverse to the flow drapes around the cloud, compresses it into a sheet, and produces broader HI absorption lines than an aligned field. If these results carry over to the Milky Way's nuclear wind, HI line spectra of outflows encode the orientation of the magnetic field and the recycling history of the gas.

What carries the argument

The load-bearing machinery is a pair of idealised radiative MHD setups. The shock-multicloud model sends a Mach 10 shock through a background medium into a layer of cold clouds with a log-normal density distribution, letting a hot post-shock wind shred and mix the clouds; the wind-cloud model resolves a single cloud in a magnetised wind at higher resolution to isolate microphysical effects. These setups are used to construct synthetic HI column densities and absorption spectra. The mechanisms doing the work are recondensation (thermal instability of mixed gas), hydrodynamic shielding (reduced drag in cloud chains), and magnetic draping (a transverse field wrapping the cloud and altering its shape), and the comparison of aligned versus transverse field runs is what produces the line-width signature.

What would settle it

Measure HI absorption line widths toward a sample of outflow clouds in the Milky Way's nuclear wind for which the magnetic field orientation is independently known from radio polarization or Faraday rotation; if clouds with transverse fields do not show systematically broader lines than aligned-field clouds, the central spectral claim is falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that HI in galactic winds is not a passive survivor but an actively recycled product. In shock-multicloud simulations, the original dense gas is destroyed and mixed, and the mixed gas, being thermally unstable, cools back into dense HI that moves with the outflow. The same models identify two further survival channels: hydrodynamic shielding, in which closely spaced clouds reduce drag on one another in supersonic flows, and, in wind-cloud models, magnetic draping, where a transverse magnetic field envelops a cloud and stabilises its interface. The new observable result is that the same transverse-field draping broadens synthetic HI absorption lines compared with an aligned-field configuration, implying that line width carries magnetic-field-orientation information.

Load-bearing premise

The argument stands or falls on whether these idealised setups—a single cloud or a layer of clouds, a homogeneous background flow, a fixed cooling floor near $10^{2}$ K, and an imposed field orientation—capture the real, multi-scale, multiphase nuclear wind closely enough for the survival mechanisms and line-width difference to transfer to observations.

Editorial extensions

If this is right

Fast-moving HI detected in the Milky Way's nuclear wind can be explained as recondensed gas rather than primordial cloud material.
HI cloudlets in multiphase outflows will be spatially coincident with cold 'molecular' cores and surrounded by warmer filamentary shells.
Observed HI absorption spectra of outflowing clouds should be broader where the ambient magnetic field is transverse to the flow than where it is aligned.
Cold gas mass in a wind can grow by recondensation of hot wind material, so HI mass is not a simple tracer of entrained ISM mass.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Going beyond the paper: if the line-width orientation effect holds in real winds, HI absorption-line surveys could be used as a magnetic-geometry diagnostic for clouds where Faraday rotation measurements are unavailable.
Inference: the same recondensation picture suggests that HI column density along a wind may not correlate monotonically with initial cloud mass, because shredded gas is recycled; a test would be measuring HI mass growth rates in higher-resolution simulations with varied cooling floors.
Inference: the transverse-field broadening could be confused with turbulent broadening in observations; separating the two would require comparing line widths with polarization or Faraday-rotation maps of the same clouds.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

This is an IAU proceedings summary of work already presented in the group's earlier and companion papers, not a self-contained research paper. It reads as a progress report: the novel bit—that transverse magnetic fields make HI absorption lines broader than aligned fields—is delegated to Villarruel et al. (2024, in press), and the recondensation picture in shock-multicloud runs is delegated to Antipov et al. (2025, submitted). By itself, the text contains no new equations, no new measurements, and no quantitative plots with values.

That said, it does what a proceedings piece should do: it is clearly written, well organised, and honest about the idealized nature of the simulations. The figures (synthetic HI column density maps and absorption spectra for aligned vs transverse fields) are illustrative, and the references to the underlying papers are explicit. The broader-science argument—that HI survival in outflows can proceed via recondensation of mixed gas rather than only survival of cold clumps—is physically plausible and consistent with the published companion papers (Casavecchia et al. 2024; Villares et al. 2024).

The soft spots are proportionate to the format, but worth naming. The abstract's headline claim rests on recondensation, and Section 3.1 asserts that 'pristine dense gas does not survive' without showing any tracer, passive scalar, or mass budget that distinguishes newly condensed gas from compressed, initially cold cores. That distinction matters: if the late-time dense HI is mostly original cloud material that was never fully mixed, the 'recondensation from the hot wind' claim is overstated. The stress-test note raises exactly this, and it lands. I would want to see the diagnostic in Antipov et al. before accepting the strong version of the claim. Similarly, the quoted mass content (~0.7) and volume filling factors (>0.9) come without error bars or time evolution, and the line-width broadening in Fig. 2 is described qualitatively—no velocity widths are quoted in the text.

The citation pattern is heavily self-referential, but that's normal for a summary of a group's own program; the published companions are real and checkable. The one genuine problem is that the central quantitative support is in a submitted, not yet public, paper.

My take: as a conference proceedings contribution, this deserves a light review and publication; the claims are plausible and the work is serious. As a standalone research paper for a journal, it would not be self-contained enough to referee. If you review it for the proceedings, ask the authors to add one sentence noting that the recondensation claim is traced to a specific companion paper, and to flag that the submitted status of Antipov et al. means the details are not yet public. It's a solid summary, but the punchline is in the companions.

Referee Report

4 major / 5 minor

Summary. This IAU proceedings paper summarises two sets of idealised radiative-MHD simulations of galactic winds: shock-multicloud models of a shock interacting with a turbulent cold layer, and wind-cloud models of a single magnetised cloud in a hot wind. The paper claims that (i) HI is not merely surviving initial cloud gas but forms by recondensation of mixed, thermally unstable gas; (ii) hydrodynamic shielding and magnetic draping help preserve dense gas; and (iii) magnetic field orientation leaves an observable imprint, with transverse fields producing broader HI absorption lines than aligned fields. It presents two figures of column density and synthetic HI spectra and delegates quantitative details to companion papers (Antipov et al. 2025, submitted; Villares et al. 2024; Casavecchia et al. 2024; Villarruel et al. 2024).

Significance. If the recondensation claim is correct, it would update the usual survival picture of cold gas in outflows and would imply that fast HI in the Galactic nuclear wind is newly formed from a mixed phase, with consequences for mass-loading and gas-recycling estimates. The line-broadening prediction offers a concrete observable diagnostic for magnetic field geometry in outflows, which is of direct interest to HI observers. The manuscript is clearly written and builds on a credible set of published companion works, and the models are described with enough context to place the claims. However, the quantitative support for the headline claims is not contained in this submission: the text asserts conclusions without presenting the supporting measurements (line widths, mass budgets, or tracers). For a proceedings contribution, deferral of details to companion papers is a common and acceptable practice, but the claims are worded more strongly than the evidence shown here.

major comments (4)

[§3.1(1) and §4] The central claim that HI arises from recondensation of mixed gas, while 'pristine dense gas does not survive', is not demonstrated in this manuscript. No passive tracer, Lagrangian particle, or mass budget tagged by initial gas origin is presented, so the late-time dense gas in the shock-multicloud simulation could equally be the surviving, compressed cores of the initial layer. The text itself refers the reader to Antipov et al. (2025, submitted) for the quantitative analysis, but the abstract and conclusions rely on this distinction. Please either add a diagnostic that separates original and recondensed gas (e.g., a tracer field or a time-resolved phase budget) in this summary, or weaken the wording to 'dense gas persists at late times and is replenished by cooling' until such support is presented.
[§3.2 and Figure 2] The statement that 'transverse magnetic fields produce broader spectral lines of HI than aligned magnetic fields' (also in the abstract and conclusions) is made without reporting any measured line width, FWHM, or uncertainty. Figure 2 shows normalised absorption spectra that appear to differ in width, but no numerical values are quoted. Since this is the main synthetic-observable prediction, the authors should either include the fitted line widths (with uncertainties) in the figure or text, or explicitly label the comparison as qualitative.
[§3.1] The reported numbers 'hot gas has high volume filling factor > 0.9' and 'warm ~1e4 K and cold ~1e2 K gas has high mass content ~0.7' are given without error bars, time evolution, or definitions of the temperature ranges. As they are used as evidence for the recondensation picture, their provenance and uncertainty should be provided (or a reference to the figure/table in the companion paper).
[§3.1(1) vs §3.1(2)] The text says in §3.1(1) that 'pristine dense gas does not survive' but in §3.1(2) that hydrodynamic shielding 'promotes the survival of dense gas'. These statements need to be reconciled: if shielding allows original clouds to survive, then the term 'pristine' needs a precise definition (e.g., unmixed gas), otherwise the reader cannot tell whether the simulation's late-time HI is survival or reformation. This is connected to the first major comment.

minor comments (5)

[Throughout] The paper uses both 'H I' and 'HI' (e.g., Section 3.1 header vs. Section 3.2 and the abstract); please standardise to one notation.
[Figure 2 caption] The caption says the column densities are 'at a distance of 50 kpc', but the axes are labelled in pc; please clarify that the 50 kpc refers to the assumed observer distance in the synthetic spectral calculation, not the projection plane.
[§2.1] The shock Mach number M=10 is stated without specifying whether it is magnetosonic or sonic; a brief definition would help readers assess the parameter choice.
[References] The entry 'Antipov A., Banda-Barragán W. E, Birnboim Y., Federrath C., Gnat O, Brüggen M. (submitted)' lacks an arXiv identifier or year; consider updating before publication.
[§2.2 and §3.1(3)] Typo: 'super-Alfvenic' should be 'super-Alfvénic'; also 'Br¨uggen' in the references appears to be a LaTeX rendering issue.

Circularity Check

0 steps flagged · score 1.0 of 10

No construction-level circularity: the HI recondensation and magnetic line-width claims are emergent simulation outcomes, not fitted or definitional; same-group companion citations are normal pointers, though the recondensation diagnostic is deferred to a submitted companion paper.

full rationale

The paper's derivation chain runs from idealized PLUTO radiative-MHD initial conditions (shock plus log-normal dense layer, or single cloud plus wind) to synthetic HI column densities and absorption spectra. The central claims, that HI can be produced by recondensation of mixed gas and that transverse magnetic fields produce broader HI lines, are computed outputs of these simulations, not parameters fitted to the conclusions. No equation in the paper defines the target result into the inputs: the cooling floor near 10^2 K does not by itself force the specific statement that some cold gas 'comes from the hot wind', and the line-width comparison is an emergent spectral synthesis result displayed in Figure 2. The main evidentiary weakness is in Section 3.1(1), where the quantitative support (mass content ~0.7, volume filling factors, and the origin of the cold gas) is delegated to Antipov et al. (2025, submitted), a same-group companion paper, and no tracer or Lagrangian mass budget is shown in this text to distinguish recondensed gas from compressed surviving initial cores. That is a support and verifiability gap, not a circular reduction: the claim does not become equivalent to the initial conditions by construction, and the companion-paper citations are used as pointers to detailed analyses rather than as an authority that forbids alternative interpretations. The paper is therefore not significantly circular.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical particles, forces, or conserved quantities; its entities are existing numerical models and magnetic field configurations.

free parameters (4)

shock Mach number = M = 10
Chosen initial condition for shock-multicloud models (Sec. 2.1); not constrained by observations in this paper.
cooling floor temperature = ~10^2 K
Radiative cooling floor in the simulations (Sec. 2.1); temperature where dense gas accumulates and affects HI mass fractions.
cloud density contrast and cloud size in wind-cloud models
Required initial conditions for cloud survival and spectra; values are not stated in this preprint and are supplied in Casavecchia et al. (2024) and companion papers.
magnetic field strength and orientation
Controls magnetic draping and line broadening; orientation is varied between aligned and transverse, but strength is not quantified in this paper.

assumptions (3)

domain assumption Radiative cooling and heating functions used in PLUTO are appropriate for the temperatures and densities encountered.
The recondensation mechanism depends on the mixed gas being thermally unstable; the paper does not validate the cooling curve against observations within this text (Sec. 2.1, Sec. 3.1).
domain assumption The idealized cloud geometries represent the key dynamics of dense gas in galactic winds.
The paper extrapolates from single-cloud and cloud-layer setups to the multiphase nuclear wind; it calls the models 'idealised' in Sec. 3.1(2).
domain assumption H I tracers derived from the simulations are faithful proxies for observed HI emission and absorption.
Synthetic observables depend on an unpublished Python generator and chosen HI column and spectrum calculation methods (Sec. 3.2, Fig. 2).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Survival and synthetic observables of neutral atomic hydrogen in galactic wind simulations." pith.science (2026). https://pith.science/paper/MDVVOBS6

@misc{pith2026241208491,
  author       = {Pith},
  title        = {Pith review of: Survival and synthetic observables of neutral atomic hydrogen in galactic wind simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MDVVOBS6}},
  note         = {Machine review of arXiv:2412.08491}
}

read the original abstract

Connecting numerical simulations to observations is essential to understanding the physics of galactic winds. Our Galaxy hosts a large-scale, multi-phase nuclear wind, whose dense gas has been detected using HI and molecular line observations. In this paper, we summarise our recent numerical work devoted to producing synthetic HI observables and measuring the properties of HI gas in galactic wind simulations. We discuss the evolution of radiative cloud systems embedded in star formation-driven galactic winds. Our shock-multicloud models show that multicloud gas streams are able to produce significant fractions of HI gas via recondensation. Our wind-cloud models show that magnetic fields have significant effects on the morphology and spectral signatures of HI gas. Cooling-driven recondensation, hydrodynamic shielding, and magnetic draping promote the survival of dense gas and the development of filamentary outflows. The orientation of magnetic fields also has an effect on synthetic observables, particularly on HI spectral lines. Transverse magnetic fields produce broader spectral lines of HI than aligned magnetic fields. Our models and analysis suggest that the fast-moving HI gas observed in the nuclear wind of our Galaxy may arise from multi-phase flows via recondensation.

Figures

Figures reproduced from arXiv: 2412.08491 by the authors.

**Figure 1.** Left: Number density rendering from a shock-multicloud simulation showing the filamentary structure of dense gas in a multi-phase outflow (see Banda-Barragan et al. ´ 2021; Antipov et al. 2025). Right: snapshot of a wind-cloud model showing the cloud number density and the 3D topology of a transverse magnetic field that drapes around the cloud (see Casavecchia et al. 2024). 2.1. Shock-multicloud models Our shock-mul… view at source ↗

**Figure 2.** Left: Synthetic column densities for H I for both orientations of the magnetic field (a: aligned field, b: transverse field) at a distance of 50 kpc from the starburst at 2.2 Myr. Right: H I absorption spectral lines for both models (a: aligned field, b: transverse field) obtained from our new python-based synthetic spectra generator (see our accompanying paper, Villarruel et al. (2024) for further details). has con… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 2 canonical work pages

[1]

S., Werk J

Tumlinson J., Peeples M. S., Werk J. K., 2017, ARA&A, 55,

work page 2017
[4]

R., Finkbeiner D

doi:10.1007/s00159-024-00152-1 Su M., Slatyer T. R., Finkbeiner D. P., 2010, ApJ, 724,

doi:10.1007/s00159-024-00152-1 2010
[11]

R., Churazov E., Heywood I., Fender R

doi:10.3847/1538-3881/ad4887 Ponti G., Morris M. R., Churazov E., Heywood I., Fender R. P., 2021, A&A, 646, A66. doi:10.1051/0004- 6361/202039636 Sarkar K. C., 2024, A&ARv, 32,

doi:10.3847/1538-3881/ad4887 2021
[31]

doi:10.3847/0004-637X/822/1/31 Cottle J., Scannapieco E., Br ¨uggen M., Banda-Barrag ´an W., Federrath C., 2020, ApJ, 892,

doi:10.3847/0004-637x/822/1/31 2020
[59]

S., Banda-Barrag´an W

doi:10.3847/1538-4357/ab76d1 Villares A. S., Banda-Barrag´an W. E., Rojas C., 2024, MNRAS, 535,

doi:10.3847/1538-4357/ab76d1 2024
[66]

M., Green J

doi:10.1038/nature11734 McClure-Griffiths N. M., Green J. A., Hill A. S., Lockman F. J., Dickey J. M., Gaensler B. M., Green A. J., 2013, ApJL, 770, L4. doi:10.1088/2041-8205/770/1/L4 Di Teodoro E. M., McClure-Griffiths N. M., Lockman F. J., Armillotta L., 2020, Natur, 584,

doi:10.1038/nature11734 2013
[228]

A., Heckman T

doi:10.1086/513316 Thompson T. A., Heckman T. M., 2024, arXiv, arXiv:2406.08561. doi:10.48550/arXiv.2406.08561 Banda-Barrag´an W. E., Br¨uggen M., Federrath C., Wagner A. Y ., Scannapieco E., Cottle J., 2020, MNRAS, 499,

arXiv doi:10.48550/arxiv.2406.08561 2024
[330]

E, Birnboim Y ., Federrath C., Gnat O, Br¨uggen M

doi:10.1088/0004- 637X/703/1/330 Antipov A., Banda-Barrag´an W. E, Birnboim Y ., Federrath C., Gnat O, Br¨uggen M. (submitted). Br¨uggen M., Scannapieco E., 2016, ApJ, 822,

doi:10.1088/0004- 2016

Show all 16 references

[347]

doi:10.1038/s41586-019-1009-6 Mignone A., Bodo G., Massaglia S., Matsakos T., Tesileanu O., Zanni C., Ferrari A., 2007, ApJS, 170,

2007 doi
[364]

S., Riquelme D., Kim W.-J., Menten K

doi:10.1038/s41586-020-2595-z Veena V . S., Riquelme D., Kim W.-J., Menten K. M., Schilke P., Sormani M. C., Banda-Barrag´an W. E., et al., 2023, A&A, 674, L15. doi:10.1051/0004-6361/202346702 Ponti G., Hofmann F., Churazov E., Morris M. R., Haberl F., Nandra K., Terrier R., e...

2023 doi
[389]

R., McQuinn M., Werk J

doi:10.1146/annurev-astro-091916- 055240 Faerman Y ., Piacitelli D. R., McQuinn M., Werk J. K., 2024, arXiv, arXiv:2406.03553. doi:10.48550/arXiv.2406.03553 Kim J.-A., Chung H., Vargas C. J., Hamden E., 2024, AJ, 168,

2024 doi arXiv
[1044]

M., Staveley-Smith L., Haverkorn M., Purcell C., Gaensler B

doi:10.1088/0004-637X/724/2/1044 Carretti E., Crocker R. M., Staveley-Smith L., Haverkorn M., Purcell C., Gaensler B. M., Bernardi G., et al., 2013, Natur, 493,

2013 doi
[1163]

E., Casavecchia B

doi:10.1093/mnras/stae2392 Villarruel D., Banda-Barrag ´an W. E., Casavecchia B. (in press, IAU Proceedings Series: Astronomy in Focus. doi:10.48550/arXiv.2411.08704). Gregori G., Miniati F., Ryu D., Jones T. W., 2000, ApJ, 543,

doi arXiv
[2173]

E., Br ¨uggen M., Heesen V ., Scannapieco E., Cottle J., Federrath C., Wagner A

doi:10.1093/mnras/staa2904 Banda-Barrag´an W. E., Br ¨uggen M., Heesen V ., Scannapieco E., Cottle J., Federrath C., Wagner A. Y ., 2021, MNRAS, 506,

2021 doi
[3454]

E., Br¨uggen M., Brighenti F., Scannapieco E., 2024, A&A, 689, A127

doi:10.1093/mnras/stx2541 Casavecchia B., Banda-Barrag´an W. E., Br¨uggen M., Brighenti F., Scannapieco E., 2024, A&A, 689, A127. doi:10.1051/0004-6361/202449461 Cooper J. L., Bicknell G. V ., Sutherland R. S., Bland-Hawthorn J., 2009, ApJ, 703,

2024 doi
[5658]

E., Federrath C., Crocker R

doi:10.1093/mnras/stab1884 Banda-Barrag´an W. E., Federrath C., Crocker R. M., Bicknell G. V ., 2018, MNRAS, 473,

2018 doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

S., Werk J

Tumlinson J., Peeples M. S., Werk J. K., 2017, ARA&A, 55,

work page 2017

[2] [4]

R., Finkbeiner D

doi:10.1007/s00159-024-00152-1 Su M., Slatyer T. R., Finkbeiner D. P., 2010, ApJ, 724,

doi:10.1007/s00159-024-00152-1 2010

[3] [11]

R., Churazov E., Heywood I., Fender R

doi:10.3847/1538-3881/ad4887 Ponti G., Morris M. R., Churazov E., Heywood I., Fender R. P., 2021, A&A, 646, A66. doi:10.1051/0004- 6361/202039636 Sarkar K. C., 2024, A&ARv, 32,

doi:10.3847/1538-3881/ad4887 2021

[4] [31]

doi:10.3847/0004-637X/822/1/31 Cottle J., Scannapieco E., Br ¨uggen M., Banda-Barrag ´an W., Federrath C., 2020, ApJ, 892,

doi:10.3847/0004-637x/822/1/31 2020

[5] [59]

S., Banda-Barrag´an W

doi:10.3847/1538-4357/ab76d1 Villares A. S., Banda-Barrag´an W. E., Rojas C., 2024, MNRAS, 535,

doi:10.3847/1538-4357/ab76d1 2024

[6] [66]

M., Green J

doi:10.1038/nature11734 McClure-Griffiths N. M., Green J. A., Hill A. S., Lockman F. J., Dickey J. M., Gaensler B. M., Green A. J., 2013, ApJL, 770, L4. doi:10.1088/2041-8205/770/1/L4 Di Teodoro E. M., McClure-Griffiths N. M., Lockman F. J., Armillotta L., 2020, Natur, 584,

doi:10.1038/nature11734 2013

[7] [228]

A., Heckman T

doi:10.1086/513316 Thompson T. A., Heckman T. M., 2024, arXiv, arXiv:2406.08561. doi:10.48550/arXiv.2406.08561 Banda-Barrag´an W. E., Br¨uggen M., Federrath C., Wagner A. Y ., Scannapieco E., Cottle J., 2020, MNRAS, 499,

arXiv doi:10.48550/arxiv.2406.08561 2024

[8] [330]

E, Birnboim Y ., Federrath C., Gnat O, Br¨uggen M

doi:10.1088/0004- 637X/703/1/330 Antipov A., Banda-Barrag´an W. E, Birnboim Y ., Federrath C., Gnat O, Br¨uggen M. (submitted). Br¨uggen M., Scannapieco E., 2016, ApJ, 822,

doi:10.1088/0004- 2016