{"id":"f5d075ca-ffe1-4120-b736-25454562deab","arxiv_id":"2501.18681","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Stacked X-ray images of 93 disk galaxies reveal soft CGM emission enhanced along the galactic minor axis, with the anisotropy appearing only in the high-star-formation subsample.","lead":"By stacking millions of seconds of Chandra X-ray observations of 93 edge-on disk galaxies, the authors detected faint hot gas around the galaxies and found that the gas is brighter along the direction perpendicular to the galactic disks. The finding could mean that outflows driven by star formation are common in such galaxies, though black-hole-driven bubbles remain a possible alternative.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SFR-anisotropy correlation rests on a single F-test with p=1.16% among six subsample tests; after Bonferroni correction the result is not significant, and no direct high- vs low-SFR amplitude comparison is made.","rationale":"The reader's weakest assumption identifies the same load-bearing issue: the SFR-anisotropy correlation is supported by one marginal F-test among six subsample comparisons, with no multiple-comparison correction and no direct amplitude comparison between high- and low-SFR subsamples. This matters because the paper's physical conclusion favors star-formation-driven outflows over SMBH-driven bubbles specifically on the basis of that SFR correlation. If the correlation is not statistically robust, the origin of the observed minor-axis enhancement becomes unconstrained, even though the stacked CGM detection and the full-sample anisotropy may remain valid. The paper has genuine strengths: the 6.5-sigma soft-band detection, the hard-band non-detection, the estimate of unresolved X-ray binary contamination, and the careful TNG50 mock-image comparison. The concern here is not about data handling or fabrication but about the statistical inference connecting a subsample-selected F-test to a physical correlation. The proposed permutation test directly addresses the multiple-testing problem by constructing the null distribution of the maximum F-statistic over all six subsample tests, and the bootstrap amplitude comparison would check whether the apparent high-SFR effect is merely a signal-to-noise artifact. The reader's conditional verdict already requires addressing this issue, so no change of verdict is needed.","tokens_in":17678,"tokens_out":7293,"duration_ms":77540,"concrete_test":"Run a permutation test: shuffle SFR labels among the 93 galaxies 10^4 times, rebuild top/bottom-third subsamples, and compute the maximum F-statistic (or minimum p-value) across the six constant-vs-sinusoid F-tests on each shuffled sample. If the observed high-SFR F-statistic does not exceed the 95th percentile of the permutation distribution, the 'only high-SFR' claim is not supported. As a complement, report bootstrap 68% confidence intervals for the sinusoid amplitudes of the high- and low-SFR subsamples to test whether the apparent difference is a detection-threshold effect.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2 reports that, after splitting the sample into top and bottom thirds in SFR, stellar mass, and SMBH mass (six subsample tests), only the high-SFR subsample satisfies the F-test for adding a sinusoidal component, with F=1.16%. But the paper applies a 5% acceptance threshold without multiple-comparison correction. With six tests, Bonferroni requires p<0.0083, so p=0.0116 would no longer be significant. The abstract's claim that 'only high star formation rate galaxies exhibit significant anisotropies' is therefore not established by the reported statistics. Moreover, the paper never compares the anisotropy amplitude of the high-SFR subsample directly with that of the low-SFR subsample; a significant F-test in one subsample and a non-significant one in another does not demonstrate a correlation with SFR, especially if the high-SFR subsample has better stacked statistics. Consequently, while the 6.5-sigma CGM detection and the minor-axis enhancement in the full stack may stand, the attribution of the anisotropy to star-formation-driven outflows rather than SMBH activity is the weakest link in the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper stacks archival Chandra observations of 93 edge-on disk galaxies (133 pointings, ~2.2 Ms) to search for diffuse X-ray emission from the circumgalactic medium (CGM) and for azimuthal anisotropies. The authors report a 6.5 sigma detection of soft 0.3-2 keV emission within a 14 kpc aperture, with luminosity (4.2 +/- 0.7) x 10^39 erg/s, no significant 3-8 keV signal, and an unresolved X-ray binary contamination estimate a factor of ~20 below the detected flux. The azimuthal profile of the stacked emission shows an enhancement along the galactic minor axis. Dividing the sample into top and bottom thirds by SFR, stellar mass, and SMBH mass, the only subsample for which a sinusoidal component significantly improves the fit is the high-SFR subsample (F-test p = 1.16%). The authors compare with mock Chandra observations of TNG50 galaxies and find that simulated SMBH-driven bubbles would appear as a major-axis enhancement at 14 kpc and correlate with SMBH mass, opposite to the observations. They conclude that the observed anisotropies are probably associated with star-formation-driven outflows, or with AGN bubbles on smaller scales than those in TNG50.","tokens_in":17914,"tokens_out":9207,"duration_ms":85492,"significance":"If the SFR-anisotropy correlation holds, the paper provides the first stacked detection of anisotropic CGM emission in external disk galaxies and a direct observational constraint on feedback processes. The core CGM detection appears robust: the signal is seen only in the soft band, the contamination estimate is well below the measured luminosity, and the comparison with TNG50 mock images is a useful and nontrivial test. The main weakness is that the paper's central interpretive claim, that anisotropy is linked to SFR rather than SMBH activity, rests on a set of uncorrected F-tests and lacks a direct subsample amplitude comparison. The paper is therefore significant but requires additional statistical work before the headline claim is established.","major_comments":[{"comment":"The claim that only the high-SFR subsample shows significant anisotropy is derived from six F-tests (high/low thirds for SFR, stellar mass, and SMBH mass), each evaluated at a 5% threshold. For six independent tests the Bonferroni-corrected threshold is p < 0.0083, so the reported p = 0.0116 for the high-SFR profile is no longer significant after correction. Please report the p-values for all six fits and either apply a multiple-comparison correction or provide a clear justification for treating the tests as independent. If the corrected significance does not survive, the abstract and Section 6 statements that 'only high star formation rate galaxies exhibit significant anisotropies' must be softened.","section":"3.2"},{"comment":"The paper does not compare the anisotropy amplitude of the high-SFR subsample directly with that of the low-SFR subsample. A significant F-test in one subsample and a non-significant result in another does not demonstrate that the profiles differ, particularly if the stacked signal-to-noise ratios differ. Please fit the constant-plus-sinusoid model jointly to both subsamples and test whether the sinusoidal amplitudes are equal (for example by bootstrap or permutation), reporting the amplitudes and uncertainties for each subsample.","section":"3.2"},{"comment":"Section 4 is used to argue that the observed minor-axis enhancement is inconsistent with TNG50-like SMBH bubbles, but it does not provide a positive prediction for star-formation-driven bubbles at ~10 kpc scales. The flat high-SFR profile in TNG50 is explained by model-specific effects (bright CGM dilution and the mix of thermal-mode SMBHs), so the simulation comparison alone does not establish that SNe-driven outflows produce the observed feature. Please either add a direct test of an SF-driven bubble model (for example, by constructing a simple outflow geometry and simulating it with the same pipeline) or present the SF interpretation explicitly as speculative rather than as the favored conclusion.","section":"4 (and 5)"}],"minor_comments":[{"comment":"The text states 133 individual observations in Section 2.1 but 120 Chandra observations in Section 6; please reconcile these numbers.","section":"2.1 and 6"},{"comment":"The notation 'F = 1.16%' is ambiguous: 1.16% is presumably the p-value of the F-test, not the F statistic. Please label it clearly and report the p-values for all six subsample fits, preferably in a small table.","section":"3.2"},{"comment":"The statement that the CGM 'extends up to 14 kpc' is not directly demonstrated; the analysis uses a fixed 14 kpc aperture and does not show a radial profile. Please rephrase as 'within 14 kpc' or include a radial profile.","section":"Abstract and 3.1"},{"comment":"The SMBH mass distributions of the observed and simulated samples differ by an order of magnitude; while this is acknowledged, the selection of simulated galaxies to match SFR but not SMBH mass makes the comparison of the SMBH-mass split difficult to interpret. Please show the TNG50 predictions for the SMBH-mass split with a matched mass range, or explicitly state how the mass mismatch affects the comparison.","section":"4"},{"comment":"The supernova energy estimate E_SN = 7.6 x 10^41 erg/s is a useful order-of-magnitude check, but the text should state explicitly that only a small fraction of this energy is converted into X-ray emission and that the comparison to L_X ~ 4 x 10^39 erg/s is therefore not a quantitative energy budget.","section":"5"},{"comment":"There are a few typographical errors: 'Star F ormation' in the title line, 'the the bulges' in conclusion item 5, and the wording of the Table 1 note about 'an apex SB' should be corrected.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The observational core of this paper is solid and the CGM detection appears robust. My main concern is the statistical support for the SFR-anisotropy correlation: the reported p-value does not survive a Bonferroni correction, and no direct amplitude comparison is made. If the authors can supply a corrected significance or a direct subsample comparison, the paper could be suitable for publication; if not, the conclusion should be substantially softened. The TNG50 comparison is interesting but is not by itself a positive test of the SF-driven scenario. I do not see grounds for rejection, because the central detection is sound and the interpretive overreach is fixable with additional analysis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things. The core detection is real and new: stacking 93 edge-on disk galaxies in Chandra gives a 6.5σ soft-band detection of extended CGM out to 14 kpc, L ≈ 4×10^39 erg/s, and the azimuthal profile shows a clear minor-axis enhancement that previous stacks, which average over azimuth, could not have seen. That part of the paper is careful and convincing. The hard-band non-detection and the factor-of-20 estimate for unresolved X-ray binary contamination support the thermal interpretation.\n\nThe weak link is the SFR correlation. The paper reports F = 1.16% for the high-SFR subsample with a 5% acceptance threshold, but that is one of six F-tests (top/bottom thirds of SFR, stellar mass, and SMBH mass). Bonferroni would require p < 0.83%, so the single 'significant' result does not survive. And the authors never directly compare the anisotropy amplitude of the high-SFR subsample with the low-SFR one. A significant F-test in one bin and a null in another is not a correlation. The abstract's 'only high star formation rate galaxies exhibit significant anisotropies' overstates what the statistics establish.\n\nThe TNG50 comparison is genuinely useful. The mock observations match the observed CGM luminosity, and the result that TNG50 bubbles would show up as a major-axis enhancement at 14 kpc, the opposite of what is seen, is a nice test. The authors note the simulated SMBH masses are ten times higher than the observed sample but then somewhat downplay that; it is a real caveat for the interpretation, though it works in the direction of saying the observed anisotropy is probably not TNG-like SMBH bubbles.\n\nThe paper deserves a serious referee. A good referee should ask for either a direct high-versus-low SFR amplitude comparison or a softened SFR claim, and a discussion of the multiple-testing issue. The detection and the morphology are solid enough to stand even if the SFR attribution turns out to be tentative.\n\nTake it to reading group; it will generate a good discussion about stacking statistics and what counts as a correlation.","headline":"Solid first detection of anisotropic CGM in stacked edge-on disks, but the SFR correlation relies on an uncorrected F-test and should be treated as tentative.","tokens_in":18447,"tokens_out":2737,"would_cite":true,"duration_ms":25601,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Stacked X-ray data from 93 edge-on disk galaxies reveal soft diffuse halo emission with a significant excess along the minor axis, seen only in the most star-forming third of the sample.","keywords":["circumgalactic medium","galaxy nuclei","X-ray astronomy","high energy astrophysics","galactic winds","X-ray bubbles","star formation feedback","stacked observations"],"falsifier":"Re-run the azimuthal analysis on the same stacked data with high- and low-star-formation subsamples compared directly—fitting the difference in sinusoidal amplitude rather than testing each profile against a constant—and apply a multiple-comparison correction for the six parameter splits. If the high-star-formation amplitude is not significantly larger than the low-star-formation amplitude, the claimed star-formation correlation fails even if a minor-axis excess in the full stack is genuine.","tokens_in":17436,"feed_emoji":"🌌","tokens_out":14077,"duration_ms":125162,"temperature":0.7,"pith_summary":"This paper tries to establish that the hot gas halo around ordinary disk galaxies is not a smooth, spherically symmetric envelope. By stacking public X-ray observations of 93 edge-on disk galaxies, the authors detect soft (0.3–2 keV) diffuse circumgalactic emission extending to 14 kpc at roughly 6.5σ significance, with an average luminosity of (4.2±0.7)×$10^{39}$ erg/s, and they show that the surface brightness is enhanced along the galactic minor axis—the direction perpendicular to the disk. When the sample is split by stellar mass, central black hole mass, and star formation rate, only the highest star formation rate third shows a statistically significant minor-axis enhancement. The paper argues that this pattern favors star formation- or supernova-driven outflows over the large black-hole-inflated bubbles seen in modern cosmological simulations, or else requires such bubbles to be much smaller than simulated. If correct, this is the first population-level evidence that disk galaxies commonly expel hot gas perpendicular to their disks, and that the amount of expulsion tracks how actively the galaxy is forming stars.","feed_headline":"Stacked X-rays reveal hot-gas outflows along galaxy minor axes","feed_subtitle":"The asymmetry appears only in galaxies with high star formation, pointing to supernova-driven winds over black hole bubbles.","key_machinery":"The argument is carried by two procedures. The first is stacking: since no individual galaxy is detected, photons from 93 edge-on galaxies are co-added in a fixed 14 kpc circular aperture after masking the galactic disk and excluding point sources, and the surface brightness is measured in five azimuthal bins symmetric about the minor axis; the presence of an anisotropy is judged by whether a sinusoidal component significantly improves the fit of the azimuthal profile, assessed with an F-test. The second is a synthetic-observation pipeline in which simulated galaxies are projected as X-ray images with realistic background and exposure and analyzed with the same extraction and fitting steps, so that the predicted azimuthal morphology of large black-hole bubbles (a major-axis excess in a 14 kpc aperture) can be compared directly with the observed minor-axis excess.","core_discovery":"The central discovery is that the circumgalactic medium of stacked disk galaxies is anisotropic. In the 0.3–2 keV band, the co-added photons produce 389±60 net counts over the 14 kpc aperture—a 6.5σ detection—corresponding to an average luminosity of (4.2±0.7)×$10^{39}$ erg/s; the 3–8 keV band shows no significant signal, indicating thermal rather than binary or power-law X-ray emission. An azimuthal profile folded about the disk plane, with θ=90° along the minor axis, is best fit by a constant plus a sinusoidal excess at θ=90°, and in subsample splits the F-test favors the sinusoidal component only for the high-SFR top third (p=1.16%). Comparison with mock X-ray images of galaxies from a large cosmological simulation shows that the simulated black-hole-driven bubbles extend to about 50 kpc and, when viewed through the same 14 kpc aperture, produce an excess near the major axis rather than the minor axis—the opposite of the observed morphology. The paper concludes that the observed anisotropy is therefore either due to star formation/starburst-driven outflows or to black-hole bubbles confined to roughly 10 kpc scales.","pith_inferences":["The six-way subsample split means chance alone could produce one nominally significant F-test; the conservative reading is that the full-stack minor-axis excess is the robust new result, while the star-formation attribution needs a direct two-sample comparison before it is treated as established.","A testable prediction follows from the simulation comparison: if black-hole bubbles are the cause, the minor-axis excess should appear only when the azimuthal profile is extracted from apertures large enough to enclose the bubbles; measuring the radial dependence of the anisotropy (e.g., 10, 20, 40 kpc apertures) would distinguish small bubbles from large ones.","The same stacking method applied to face-on or intermediate-inclination galaxies, or to samples split by environment, would test whether the apparent minor-axis excess is truly a perpendicular outflow geometry rather than an artifact of masking the disk.","If supernova-driven winds are responsible, the minor-axis emission should be hotter (harder spectrum) than the surrounding CGM, as the paper's hardness ratios hint, and the anisotropy should correlate with SFR surface density rather than total SFR; both are testable with deeper stacking."],"forward_implications":["The hot circumgalactic medium of Milky Way-mass disk galaxies is not spherical at 10–14 kpc scales, so X-ray-based mass and baryon-budget estimates that assume spherical symmetry would need to account for this bipolar structure.","Because the minor-axis enhancement appears only in the high-star-formation third, ongoing star formation—through supernovae and starburst winds—is the more likely driver of such outflows in the present-day universe than central black hole activity.","If black-hole feedback does produce Milky Way-like bubbles in these galaxies, the bubbles must be roughly 10 kpc in size rather than the ~50 kpc scales produced by the simulation, constraining feedback energetics and coupling.","Individual galaxies in the high-SFR subsample should show minor-axis X-ray outflows at ~14 kpc in sufficiently deep exposures, making them priority targets for follow-up X-ray spectroscopy."],"supporting_citations":[{"why":"Supplies the parent galaxy catalog from which the 93 edge-on disk galaxies and their stellar masses, distances, and star formation rates are drawn.","marker":"Kovlakas et al. 2021"},{"why":"Provides the cosmological-simulation prediction that Milky Way-like X-ray bubbles are common and correlate with central black hole mass; the observed anisotropy pattern is compared against it.","marker":"Pillepich et al. 2021"},{"why":"Earlier analysis of the same simulation's CGM anisotropies that motivates the expectation that bubbles produce detectable azimuthal structure on 10–50 kpc scales.","marker":"Truong et al. 2021a"},{"why":"Defines the magnetohydrodynamic simulation run from which the simulated galaxies in the mock sample are drawn.","marker":"Nelson et al. 2019"},{"why":"Supplies the code that turns simulated gas density, temperature, and abundances into X-ray photons for the mock observations.","marker":"ZuHone et al. 2014"},{"why":"Supplies the detector simulator that converts the mock photons into event files analyzed with the same pipeline as the real data.","marker":"ZuHone et al. 2023"},{"why":"Reports the minor-axis X-ray outflow in a local disk galaxy attributed to supernovae, used as a nearby analog for the detected anisotropy.","marker":"Bogdán & Gilfanov 2008"},{"why":"Reports a similar supernova-driven minor-axis outflow in another disk galaxy, supporting the star-formation interpretation.","marker":"Li et al. 2011"},{"why":"Provides the starburst classification criterion applied to galaxies in the sample, linking the high-SFR subsample to starburst-driven outflows.","marker":"Strickland et al. 2004"}],"fun_headline_variants":["High star formation, not black holes, drives CGM anisotropies","Stacked X-rays tie galaxy halo anisotropies to starbursts","CGM anisotropies in disk galaxies: star formation wins over AGN","Disk galaxy gas halos lopsided only in high star formation","X-ray maps of galaxy halos point to starburst-driven outflows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion that star formation drives the asymmetry depends on treating a 1.16% chance that the high-star-formation subsample's azimuthal profile is flat as significant, without accounting for the fact that six subsample tests were performed; if a multiple-comparison correction is required, the star-formation correlation is not formally established.","fun_headline_variants_meta":{"raw":{"variants":["High star formation, not black holes, drives CGM anisotropies","Stacked X-rays tie galaxy halo anisotropies to starbursts","CGM anisotropies in disk galaxies: star formation wins over AGN","Disk galaxy gas halos lopsided only in high star formation","X-ray maps of galaxy halos point to starburst-driven outflows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001193,"raw_usage":{"total_tokens":4997,"prompt_tokens":1095,"completion_tokens":3902,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":711,"completion_tokens_details":{"reasoning_tokens":3804}},"tokens_in":711,"tokens_out":3902,"duration_ms":28405,"temperature":1.0,"reasoning_tokens":3804,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T22:52:24.156240+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the azimuthal analysis on the same stacked data with high- and low-star-formation subsamples compared directly—fitting the difference in sinusoidal amplitude rather than testing each profile against a constant—and apply a multiple-comparison correction for the six parameter splits. If the high-star-formation amplitude is not significantly larger than the low-star-formation amplitude, the claimed star-formation correlation fails even if a minor-axis excess in the full stack is genuine.","supporting_citations":[{"cited_title":"A., Vikhlinin, A., Tremblay, G","cited_arxiv_id":null,"evidence_quote":"Supplies the detector simulator that converts the mock photons into event files analyzed with the same pipeline as the real data."}],"review_version":1}