{"id":"22a64f41-9338-4a44-a23b-445fcf4d7c6d","arxiv_id":"2506.14361","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"ALMA detects about 10^9.7 solar masses of molecular gas in a merging post-starburst pair at z~1.4, while four similar galaxies are undetected, hinting that mergers may preserve gas while quenching star formation.","lead":"Using ALMA, the authors measured cold molecular gas in five distant galaxies that recently stopped forming stars, and found a gas reservoir in only one object: a close pair of galaxies that appear to be merging. The finding hints that major mergers may stop star formation while keeping gas around, a scenario that future surveys can test.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prospector t_quench ages are the load-bearing link for the claimed absence of an age-gas trend; a ~400 Myr model offset would erase the result.","rationale":"The reader's weakest_assumption correctly identifies the Prospector t_quench scale as the key uncertainty, and the paper's own comparison of t50 and t_quench validates the concern. The abstract's stronger wording ('we find no correlation') goes beyond what n=5 and model-dependent ages can support. The ALMA detection itself is credible and the merger interpretation is properly hedged, so no rejection is warranted; the paper should remain conditional, with revisions that soften the no-correlation claim, report systematic age uncertainties, and propagate alpha_CO/r21 uncertainties. The proposed cross-fitting test is feasible with existing data and would settle whether the age offset is physical or methodological.","tokens_in":14936,"tokens_out":18419,"duration_ms":215724,"concrete_test":"Re-run Prospector on the SQuIGGLE galaxies using the exact HeavyMetal fitting configuration (same rest-frame wavelength windows, spectral S/N, photometric bands, dust prior, and non-parametric PSB SFH prior), and on the HeavyMetal galaxies using the SQuIGGLE configuration. If the median t_quench offset between samples drops below ~100 Myr, the claimed redshift evolution is a fitting artifact and the abstract should be revised. If the offset persists, the age-based interpretation is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the z~1.4 HeavyMetal post-starbursts show no correlation between molecular gas mass and time since quenching, and the explanation that they are older than gas-rich low-z post-starbursts, rests entirely on the Prospector t_quench ages. The paper itself shows that UVJ colors imply similar t50 ages for HeavyMetal and SQuIGGLE, while Prospector returns systematically older t_quench by ~400 Myr (Section 4). This discrepancy is attributed to a different old-star fraction, but it is equally consistent with a model-driven offset: the two samples have different rest-frame wavelength coverage, signal-to-noise, and dust priors, and the non-parametric post-starburst SFH prior can trade dust against age. If the HeavyMetal t_quench values are overestimated by ~400 Myr, the detected system would have t_quench ~200 Myr, placing it in the gas-rich regime of the low-z trend; the 'no correlation' and 'older than typical' statements in the abstract would both disappear. With one detection and four upper limits, the data cannot distinguish these scenarios, so the age scale is the most load-bearing assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents ALMA CO(2-1) observations of five spectroscopically confirmed post-starburst galaxies at z~1.4 from the HeavyMetal survey. One system, HM1-214340, is detected at ~5σ with M_H2 ≈ 10^9.74 M_sun; this system is a close pair of two massive post-starburst galaxies (HM1-214340 and UVISTA-214345) separated by ~15 kpc and ~300 km/s, likely in the early stages of a major merger. The other four galaxies are undetected, with 3σ upper limits of log M_H2 < 9.6-10.0. The authors argue that, unlike at z<1, there is no correlation between molecular gas mass and time since quenching, possibly because the HeavyMetal galaxies have older post-burst ages (t_quench > 500 Myr) than typical gas-rich low-redshift post-starbursts. They suggest that major mergers may quench galaxies while allowing them to retain molecular gas, perhaps through turbulence-driven decreases in star formation efficiency.","tokens_in":15208,"tokens_out":6909,"duration_ms":68244,"significance":"The measurement is a valuable addition to the small sample of molecular gas observations in quiescent galaxies at z>1. The ~5σ CO(2-1) detection in a pair of massive post-starbursts is robust, and the upper limits are conservatively derived (800 km/s extraction, 3σ). The paper is careful in describing the data reduction and the model dependence of the ages, and it explicitly discusses the discrepancy between UVJ-based t50 and Prospector-based t_quench. The identification of a gas-rich, early-stage major merger among the quiescent population is interesting and will motivate follow-up work. However, the statistical and systematic limitations of the sample mean that the headline 'no correlation' and the 'older than typical' explanation are not firmly established.","major_comments":[{"comment":"The statement in the abstract that 'we find no correlation between molecular gas mass and time since quenching' is not supported by the data. The sample consists of one detection and four upper limits, all at t_quench > 500 Myr, so the data cannot distinguish a correlation from its absence. The four non-detections are consistent with the low-redshift trend (gas-poor at t_quench > 150 Myr), and the only evidence against the trend is the single detection at t_quench ~ 600 Myr. Please rephrase the claim to reflect that the data are consistent with a lack of correlation but do not provide a meaningful correlation test, or that the detected system appears as an outlier rather than a demonstration of no correlation.","section":"Section 4 / Figure 4 (right) and abstract"},{"comment":"The conclusion that the HeavyMetal galaxies are older than typical gas-rich low-redshift post-starbursts, and hence that the age-gas trend is absent at z~1.4, rests entirely on the Prospector-derived t_quench ages. The paper itself demonstrates that UVJ-based t50 ages are consistent between SQuIGGLE and HeavyMetal while the Prospector t_quench values differ by ~400 Myr, with the authors attributing the offset to differences in the old-star fraction. If the t_quench values are systematically overestimated by ~400 Myr, the detected system would have t_quench ~ 200 Myr, placing it in the gas-rich regime of the low-z trend and eliminating the reported discrepancy. Because the age scale is the load-bearing element for both the 'no correlation' and 'older than typical' statements, please explicitly test the robustness of the conclusion to the choice of age indicator (e.g., by showing the same relation with t50) or state clearly that this result is contingent on the Prospector model. As written, the abstract overstates the certainty of a result that depends on a model-dependent age scale.","section":"Section 2.2 and Section 4 (age scale)"}],"minor_comments":[{"comment":"The caption refers to 'UVISTA-214340' but the text and labels use 'HM1-214340' for the primary target; please correct the caption to 'HM1-214340'.","section":"Figure 2 caption"},{"comment":"The caption lists 'HM2-57702' as one of the undetected targets; this should be 'HM2-25702'.","section":"Figure 3 caption"},{"comment":"The quantity A/(A+K)~0.6 is introduced without definition; please define the terms or rephrase the sentence for clarity.","section":"Section 4"},{"comment":"Carnall et al. 2024b is listed with the same journal, volume, and page as Carnall et al. 2024a; please correct the bibliographic details.","section":"References"},{"comment":"Expressions such as 'MH2 ∼10 9.7 M⊙' have spacing and superscript issues; please use consistent LaTeX notation (e.g., $M_{\\rm H_2} \\sim 10^{9.7}\\,M_\\odot$).","section":"Abstract and text"}],"recommendation":"major_revision","confidential_remarks":"The measurements are solid and the paper is honest about many of its caveats. My main concern is that the abstract and title make stronger claims than the data support, particularly the 'no correlation' statement and the 'major mergers as a pathway' interpretation. With one detection and four upper limits, and with the age scale being model-dependent, the conclusions should be appropriately softened or supported by an additional robustness analysis. The paper is a good candidate after such revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuine observational step forward. New ALMA CO(2-1) data on five spectroscopically confirmed post-starburst galaxies at z~1.4, with one clean ~5 sigma detection in a close pair of massive post-starbursts and four conservative upper limits. That is the first molecular gas measurement in this specific population at the peak quenching epoch, and the serendipitous confirmation of the merging companion (UVISTA-214345) is a nice bit of luck they exploited properly. The gas mass ~10^9.7 Msun and gas fraction ~4% are credible.\n\nThe paper is honest. They explicitly discuss the t50 vs t_quench discrepancy and what it means, they use an 800 km/s extraction for upper limits, they do a two-component fit to localize the CO, and they flag that the merger interpretation is tentative. The discussion of simulations and alternative quenching paths is level-headed.\n\nThe soft spot is the 'no correlation between gas mass and time since quenching' claim in the abstract. With one detection and four upper limits, all at t_quench > 500 Myr, you cannot test a trend. The age scale itself is the load-bearing assumption: Prospector returns systematically older t_quench for HeavyMetal than SQuIGGLE by ~400 Myr while UVJ colors imply similar t50. The stress-test note is right that a ~400 Myr model offset would move the detected system from t_quench ~600 Myr to ~200 Myr, placing it in the gas-rich regime of the low-z trend and wiping out the claimed difference. The paper acknowledges the age discrepancy but still draws the 'no correlation' conclusion from those ages. Also, alpha_CO and r21 are fixed to MW values without propagation; that could shift gas masses by factors, though it would not change the detection itself.\n\nFor the field, this is a useful data point: it increases the number of gas measurements in z>1 quiescent galaxies by about 1.5 and adds a genuinely interesting merging system. The interpretation is speculative but clearly labeled. I would send it to a serious referee. The referee should ask the authors to soften the abstract, explicitly frame the age-gas comparison as limited by the model-dependent ages and small sample, and propagate alpha_CO/r21 uncertainties. None of that sinks the paper; it just makes it match what the data actually support.","headline":"Solid new ALMA detection in a z~1.4 post-starburst pair, but the 'no correlation' claim overreaches the data and rests on model-dependent ages.","tokens_in":15825,"tokens_out":2206,"would_cite":true,"duration_ms":23539,"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":"At z~1.4, a merging pair of post-starburst galaxies retains molecular gas, suggesting mergers can quench galaxies without removing their cold fuel.","keywords":["post-starburst galaxies","molecular gas","galaxy mergers","galaxy quenching","high-redshift galaxies","ALMA CO(2-1)","star formation efficiency","HeavyMetal survey"],"falsifier":"CO(2-1) observations of a larger sample of $z\\sim1.4$ post-starburst galaxies with spectroscopically measured $t_{\\rm quench}<150$ Myr would test the claim: if such young systems are uniformly gas-poor, the age-gas trend actually persists at high redshift; if they are gas-rich, the trend holds and the HeavyMetal non-detections reflect older ages.","tokens_in":14766,"feed_emoji":"🌌","tokens_out":7216,"duration_ms":66116,"temperature":0.7,"pith_summary":"This paper tries to establish that molecular gas can survive in quenched galaxies at the peak epoch of quenching, and that major mergers may be the reason. The authors present ALMA CO(2-1) observations of five spectroscopically confirmed post-starburst galaxies at $z\\sim1.4$ from the HeavyMetal survey; four are undetected, but one system, a close pair of massive post-starburst galaxies, is detected with $M_{\\rm H_2}\\sim10^{9.7}\\,M_\\odot$. The pair shows no tidal features and is likely an early-stage major merger. The paper argues that this system is a case where merger-driven turbulence suppresses star formation while preserving cold gas, and reports that, unlike at $z<1$, there is no correlation between molecular gas mass and time since quenching. If correct, this points to mergers rather than post-burst age as the key factor controlling gas retention in high-redshift quiescent galaxies.","feed_headline":"Merging dead galaxies keep their molecular gas at z~1.4","feed_subtitle":"A massive close pair is quenched yet retains ~10^9.7 solar masses of gas, hinting mergers both quench and preserve fuel.","key_machinery":"The central object is the molecular gas reservoir traced by ALMA CO(2-1) emission, converted to $M_{\\rm H_2}$ assuming thermalized emission and a Milky Way-like $\\alpha_{\\rm CO}=4.0$. The argument is carried by comparing this gas mass against the post-burst age $t_{\\rm quench}$ derived from full spectral energy distribution fitting with Prospector, using a non-parametric post-starburst star-formation history. The paper also uses UVJ colors as a light-weighted age indicator ($t_{50}$) and finds a discrepancy between the two age scales for the HeavyMetal and SQuIGGLE samples. The merger interpretation rests on HST imaging and the serendipitous Keck spectrum of UVISTA-214345, which establish a close pair with no tidal features and a small projected separation, and on the modeled star-formation history showing a recent burst that ended $\\sim500$--$600$ Myr ago.","core_discovery":"On its own terms, the paper's central discovery is a single detection: the CO(2-1) line is detected at $\\sim5\\sigma$ in the HM1-214340/UVISTA-214345 system, a spectroscopically confirmed pair of massive ($M_*\\sim10^{11.1-11.2}\\,M_\\odot$) post-starburst galaxies at $z=1.418$ and $1.421$ separated by $\\sim15$ kpc and $\\sim300$ km/s, with no clear tidal features. The inferred molecular gas mass is $\\sim10^{9.74}\\,M_\\odot$, a gas fraction of about 4%, and the best two-component fit places at least $\\sim80$% of the CO flux on UVISTA-214345. The paper interprets this as an early-stage major merger in which both members have already quenched, contrary to the usual picture that star formation is enhanced during the close-pair phase. It then uses the full five-galaxy sample to show that the low-redshift correlation between gas mass and time since quenching (Bezanson et al. 2022; French et al. 2018) is not reproduced at $z\\sim1.4$: the detected pair has $t_{\\rm quench}\\sim600$ Myr, similar to the non-detected systems. This leads to the claim that at high redshift, mergers may be the primary driver of whether a quenched galaxy retains molecular gas, possibly by injecting turbulence that lowers star formation efficiency.","pith_inferences":["The merger-turbulence explanation predicts that the CO line width of the detected system should be broader, or its star formation efficiency lower, than in isolated post-starbursts of the same mass; this could be tested by comparing high-resolution CO kinematics with SFR limits in a larger sample.","The $t_{50}/t_{\\rm quench}$ discrepancy suggests that UVJ-color age estimates may systematically understate the true post-burst ages of dusty high-redshift galaxies; re-fitting existing $z>1$ quiescent samples with full spectroscopic modeling could change the inferred gas-age relation.","If mergers quench while preserving gas, some of the apparently quiescent massive galaxies at cosmic noon may have recently passed through a gas-rich, low-star-formation phase; deep ALMA observations of close pairs of quiescent galaxies could reveal more such systems."],"forward_implications":["At $z\\sim1.4$, quenched galaxies can retain molecular gas fractions comparable to some star-forming galaxies, so quiescence does not require complete removal of cold gas.","The absence of a gas-age correlation at $z\\sim1.4$ means the simple 'young and gas-rich, old and gas-poor' picture established at $z<1$ does not extend unchanged to the peak quenching epoch.","If mergers are the key retention channel, samples of post-starburst galaxies with merger features should show higher CO detection rates than isolated post-starbursts.","Young UVJ-selected samples may be systematically older in $t_{\\rm quench}$ than their colors imply, which affects the interpretation of any UVJ-selected high-redshift quiescent sample.","The detected pair's depletion time of $\\sim1$ Gyr is similar to star-forming galaxies, so the gas could reignite star formation as the merger proceeds, meaning this system may not end as a typical gas-poor remnant."],"supporting_citations":[{"why":"Supplies the SQuIGGLE sample at z~0.7 and the correlation between gas mass and time since quenching that this paper tests at higher redshift.","marker":"Bezanson et al. 2022"},{"why":"Established the gas-age trend at z~0 that motivates the high-redshift comparison.","marker":"French et al. 2018"},{"why":"Provides the HeavyMetal survey spectra, redshifts, stellar masses, and SFRs for the five targets.","marker":"Kriek et al. 2024"},{"why":"Defines the Prospector non-parametric post-starburst star-formation history and the t_quench age measure used for all age comparisons.","marker":"Suess et al. 2022a"},{"why":"Supplies the CO-to-H2 conversion factor alpha_CO=4.0 assumed when converting line fluxes to gas masses.","marker":"Bolatto et al. 2013"},{"why":"Provides the PHIBSS star-forming galaxy scaling relation used to compare depletion times.","marker":"Tacconi et al. 2013"},{"why":"Cited as a recent HI study finding no age-gas trend in low-redshift post-starbursts, used to contextualize the absence of a trend at high redshift.","marker":"Ellison et al. 2025"},{"why":"Supports the turbulence-driven star formation suppression mechanism invoked to explain gas retention in the merger.","marker":"Spilker et al. 2022"}],"fun_headline_variants":["Merging dead pair keeps gas at z~1.4","Quenched merger holds cold gas: quenching and fuel together","Gas-rich quenched pair: merger preserves fuel","Post-starburst pair's gas survives quenching","At z~1.4, dead galaxies still hold cold fuel"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Prospector post-burst ages ($t_{\\rm quench}$) are reliable and directly comparable across the SQuIGGLE and HeavyMetal samples; if those model ages shift, the claimed absence of a gas-age correlation at $z\\sim1.4$ disappears.","fun_headline_variants_meta":{"raw":{"variants":["Merging dead pair keeps gas at z~1.4","Quenched merger holds cold gas: quenching and fuel together","Gas-rich quenched pair: merger preserves fuel","Post-starburst pair's gas survives quenching","At z~1.4, dead galaxies still hold cold fuel"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000517,"raw_usage":{"total_tokens":2607,"prompt_tokens":1144,"completion_tokens":1463,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":760,"completion_tokens_details":{"reasoning_tokens":1382}},"tokens_in":760,"tokens_out":1463,"duration_ms":13727,"temperature":1.0,"reasoning_tokens":1382,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:18:04.412688+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"CO(2-1) observations of a larger sample of $z\\sim1.4$ post-starburst galaxies with spectroscopically measured $t_{\\rm quench}<150$ Myr would test the claim: if such young systems are uniformly gas-poor, the age-gas trend actually persists at high redshift; if they are gas-rich, the trend holds and the HeavyMetal non-detections reflect older ages.","supporting_citations":[{"cited_title":"Low redshift post-starburst galaxies host abundant HI reservoirs","cited_arxiv_id":"2503.03066","evidence_quote":"Cited as a recent HI study finding no age-gas trend in low-redshift post-starbursts, used to contextualize the absence of a trend at high redshift."}],"review_version":1}