{"id":"d8e6c48e-0407-4a9b-b476-326e834d8946","arxiv_id":"2603.16991","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Between z≈4.4 and 2.7, a simulated protocluster's cold circumgalactic gas is destroyed from the inside out by a merger and AGN feedback, while its Lyman-alpha halo shifts from filamentary to spherical.","lead":"A high-resolution simulation of a massive galaxy cluster while it is still forming (z≈3) shows the gas around it heating from the inside out, turning cold filaments into a hot, X-ray-emitting medium. The paper predicts how this transition should appear in Lyman-alpha and H-alpha emission and in magnesium/oxygen absorption, giving telescopes like MUSE and JWST specific things to look for.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No feedback-off control run: the causal claim that AGN feedback drives the inside-out CGM-to-ICM transition is inferred from timing and energetics, not demonstrated.","rationale":"I read the paper as an honest, well-executed single-halo simulation study. Its descriptive results—the inside-out heating, the MgII destruction, the Lyα morphology shift—are supported within the simulation framework. The reader's verdict of CONDITIONAL is appropriate because the causal claim about AGN feedback is central and currently rests on correlation, not on a controlled experiment. My concern is more specific than the reader's 'subgrid model fidelity' worry: even if every Fable subgrid parameter is accepted, the paper never tests whether the AGN feedback is responsible for the transition. The high-resolution versus fiducial comparison in Section 6.2.1 isolates resolution, not feedback; no control run isolates AGN. Thus the abstract's statement that AGN feedback 'drives' the transformation is an interpretation. This does not require rejection: the case study is valuable, the caveats are partly acknowledged, and a single feedback-off run would resolve the issue. The agreement is partial because the reader identified subgrid fidelity as the weakest assumption, whereas I see the missing causal control as more load-bearing; both point to the same general region of the argument. I therefore recommend no change to the CONDITIONAL verdict.","tokens_in":27243,"tokens_out":3561,"duration_ms":41597,"concrete_test":"Run a paired control of the same zoom-in initial conditions, with identical shock refinement and resolution, but with AGN feedback disabled (or at minimum radio-mode bubble injection switched off) through z=2.7. Then compare volume-weighted density, temperature, and metallicity profiles in the 50–250 kpc annulus, plus MgII covering fractions, against the fiducial run. If the profiles flatten comparably and the MgII covering fraction declines similarly without AGN feedback, the central causal claim is unsupported. Quantify with, e.g., Δlog10 T(100 kpc) between z=4.4 and z=2.7 in both runs; require at least a ~0.3 dex difference to retain the attribution.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is causal: 'a major merger and AGN feedback drive an inside-out transformation' (Abstract; also Section 4.1 and Conclusions item 4). For a single simulated halo, the evidence presented is a temporal correlation—radio-mode feedback turns on at z≲3.5 and injects ~3e60 erg—plus a comparison between the high-resolution shock-refined run and the fiducial Fable run. But that comparison changes only the CGM resolution; it does not vary the feedback implementation. There is no run with AGN feedback disabled, no run with radio-mode bubble injection switched off, and no run with modified feedback coupling radius or duty cycle. Consequently, the observed flattening of density/temperature/metallicity profiles in the 50–250 kpc range and the destruction of intermediate-column MgII absorbers could be caused instead by the 1:2.5 merger shock, by normal virialization as the halo grows, by resolution-dependent cooling, or by a combination. The authors sometimes hedge (Section 4.1 says 'most likely due to a combination of merger activity and radio-mode AGN feedback'), but the abstract and conclusions assert causation. This is sharper than the generic subgrid-fidelity concern: even granting the Fable subgrid model exactly, the unique causal role of AGN feedback is untested. The single-black-hole setup further weakens the attribution—merger-driven BH growth is excluded by construction, so injected AGN feedback energy is lower than in fiducial Fable, while the radiative AGN luminosity in the post-processing uses the Fable black-hole mass. This makes the asserted AGN-driven story less secure, not more.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a cosmological zoom-in simulation of a massive cluster progenitor (M_200 ≈ 7×10^13 M_sun at z=2.7) with enhanced CGM resolution (shock-refined Arepo cells, target 3.2×10^4 M_sun). Using the Fable/Illustris subgrid model, the authors follow the halo from z=8 to z=2.7 and analyze the evolution of density, temperature, metallicity, and gas-phase fractions. They post-process the simulation with Trident for MgII and OVII absorption and with the colt Monte Carlo radiative transfer code for Lyα and Hα emission, including variations in dust-to-metal ratio, AGN ionisation, and resonant scattering. The central claims are that a 1:2.5 merger and AGN feedback drive an inside-out transformation of the CGM into a hot, enriched, X-ray-dominated ICM; that intermediate-column MgII absorbers are destroyed while dense satellite-associated gas survives; and that Lyα haloes evolve from filamentary to spherical while remaining detectable without AGN photoionisation. The authors compare their mock Lyα profiles to QSOMuseum data and find an underpredicton in the central ~50 kpc, which they interpret as unresolved ISM photon escape/AGN conversion. The paper is explicitly framed as a single-object case study with acknowledged caveats in §7.3.","tokens_in":27593,"tokens_out":3735,"duration_ms":43573,"significance":"If the central evolutionary narrative holds, the paper provides a valuable, high-resolution case study of the CGM-to-ICM transition at z~3, a regime that is poorly constrained observationally and numerically. The explicit resolution comparison (fiducial Fable vs. high-resolution CGM) and the systematic exploration of dust, scattering, and AGN ionisation in the radiative-transfer post-processing are genuine strengths. The mock absorption and emission predictions are not fitted to the observed Lyα data; the comparison is predictive rather than circular. However, the paper's general significance is limited by the use of a single halo and by the fact that the causal attribution to AGN feedback is not tested with a control run. The authors are appropriately cautious in parts of §4.1 and §7.3, but the abstract and conclusions state the causal claim more strongly than the available evidence supports.","major_comments":[{"comment":"The paper's central causal claim—that AGN feedback drives the inside-out CGM-to-ICM transformation—is supported only by a temporal correlation and a resolution comparison. The comparison with the fiducial Fable run (Fig. 9, §6.2.1) changes CGM resolution but not the feedback implementation. There is no run with AGN feedback disabled, no radio-mode-off control, and no variation of duty cycle or coupling radius. The contemporaneous 1:2.5 merger shock (§4.1) and the normal growth of the halo are equally capable of producing the observed flattening and heating in the 50–250 kpc range (Fig. 2). The 3×10^60 erg injection quoted in §4.2 is an internal model output, not an independent constraint. I request either a control experiment (even a short restart with radio-mode feedback disabled) or a systematic softening of the causal wording in the abstract and conclusions to 'concurrent with' or 'mo","section":"§4.2 and Abstract / Conclusion 4"},{"comment":"The single-black-hole setup weakens the AGN attribution in a specific way. The authors state that because the black hole cannot grow by mergers, its mass at z=2.7 is ~7 times lower than in fiducial Fable, with 'an equivalent decrease in the injected feedback energy.' Yet the paper still attributes the ICM transition to this same AGN feedback. The Eddington luminosity used for AGN photoionisation in colt is that of the Fable black hole (§3.2.1), so the radiative bracketing is reasonable, but the mechanical feedback that drives the claimed redistribution is not bracketed. The reader is left without a quantitative comparison of injected AGN energy versus merger-shock energy or virialization heating. Please add this comparison or, better, a run where the black hole follows the fiducial Fable growth history.","section":"§2 and §7.3"},{"comment":"The generalization from one object to a universal 'inside-out' transition is not supported by the presented evidence. The authors explicitly acknowledge in §7.3 that they 'only focus on an individual object' and in §4.1 that a broader suite is needed to determine ubiquity, but the abstract and Conclusion 4 state the inside-out transition as an unconditional result. Since the merger history (1:2.5), the single BH, and the particular resolution setup are all specific to this halo, the paper should either present a small ensemble or explicitly reframe the abstract and conclusions as a case study with model-dependent predictions. Otherwise the headline claim overreaches the data.","section":"§4.1 and §7.3"}],"minor_comments":[{"comment":"The text says Trident 'assumes ionisation equilibrium with the metagalactic UVB' and later notes local stellar sources are similar, but the relative importance of collisional ionisation in the hot halo is not quantified. A sentence on where collisional ionisation matters for OVII would help the absorption predictions.","section":"§3.1"},{"comment":"In Fig. 10, the color/line-style legend is dense and the difference between 'Observed, wAGN ionisation, DTM = 0' and the no-AGN case is not visible at the plotted scale. Consider using a zoomed inset or tabulating the central SB values for each model.","section":"§6.2.2"},{"comment":"The phrase 'shock refined run' and 'shock refinement scheme' is used interchangeably; defining 'ShockRef512' once in §2 would avoid confusion with the earlier Bennett & Sijacki (2020) nomenclature.","section":"§2"},{"comment":"The discussion of future work mentions better-resolved multiphase ISM models, but the authors do not discuss potential effects of a higher metal cooling rate or non-equilibrium ionisation on MgII destruction. This would be a useful caveat given the rapid timescales claimed.","section":"§7.3"}],"recommendation":"major_revision","confidential_remarks":"This is a competent and useful case study, and the mock-observation pipeline is a strength. However, the abstract and conclusions assert a causal role for AGN feedback that the simulation design—single halo, single BH, no feedback-off control—cannot establish. The requested control run may be expensive, but the authors could alternatively reframe the paper's contribution as a high-resolution case study with model-dependent predictions. I would lean toward accepting after that reframing. The paper fits the journal's scope well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real case study, not a field-changing result. If you work on high-z CGM or protoclusters, it is worth reading for the resolution comparison and the mock maps.\n\nWhat is actually new: the ShockRef512 zoom is continued to z = 2.7 with 200–300 pc cells in the CGM, and the paper shows explicitly that Ly-alpha surface brightness profiles are boosted by a factor of 2.8 when CGM resolution is increased. The mock MgII/OVII absorption and Ly-alpha/H-alpha emission from the same run, including comparisons to QSOMuseum data, are useful and mostly well executed. The inside-out flattening of density, temperature, and metallicity profiles between z ~ 4.4 and 2.7 is clearly presented and internally consistent. The authors are also honest: Section 7.3 flags the single-object limitation, the single black hole, the simplified ISM treatment, and the central Ly-alpha underprediction. They do not hide the central tension with observations.\n\nThe soft spot worth naming is the causal claim. The abstract and conclusions say a major merger and AGN feedback drive the inside-out transformation. What the simulation actually shows is a temporal correlation — radio-mode feedback turns on around z ~ 3.5 and injects ~3e60 erg — plus a resolution comparison that does not vary feedback physics. There is no feedback-off run, no radio-mode-off run, and no variation of coupling radius or duty cycle. So the unique causal role of AGN feedback is not demonstrated. The 1:2.5 merger shock, normal virialization, and resolution-dependent cooling could plausibly do much of the work. To their credit, the authors hedge in Sections 4.1 and 4.2 ('most likely due to a combination'), but the abstract overstates the case. That is a proportionality problem, not a fatal flaw.\n\nOther caveats are the usual ones: single halo, Fable subgrid feedback, assumed dust-to-metal ratio of 0.1, AGN luminosity set by Eddington from the Fable black hole mass. All are stated, and the mock observables are not fitted to the target data, which is good. The simulation data and code are not public, which limits reproducibility but is common for this type of work.\n\nI agree with the reader's conditional verdict. If I were the editor, I would send this to peer review rather than desk reject it. The referee report should ask for softened causal language or, better, a control run with feedback disabled from z ~ 4. I would bring this to reading group and teach it as a case study, not as proof that AGN feedback births the ICM.","headline":"A genuinely useful single-halo case study of the z ~ 3 CGM-to-ICM transition with honest mock observables; the AGN-causality claim is asserted more strongly than the run design can support, but the paper deserves a serious referee.","tokens_in":28155,"tokens_out":1939,"would_cite":true,"duration_ms":23833,"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":"A simulated protocluster shows the intracluster medium is born from the inside out as AGN feedback and a major merger destroy cold gas and reshape Lyman-alpha haloes between redshift 4.4 and 2.7.","keywords":["intracluster medium","circumgalactic medium","protocluster","AGN feedback","Lyman-alpha halo","MgII absorbers","OVII absorption","cosmological zoom-in simulation"],"falsifier":"Measure the radial distribution of strong MgII absorbers around a statistical sample of z∼3 protoclusters (e.g., using MUSE or Keck spectroscopy): if intermediate-column absorbers persist within 250 kpc in a majority of merging systems, the proposed inside-out destruction of cold gas would be contradicted.","tokens_in":27144,"feed_emoji":"🌌","tokens_out":3879,"duration_ms":36431,"temperature":0.7,"pith_summary":"This paper aims to establish how a galaxy cluster's hot, X-ray-emitting intracluster medium (ICM) first forms out of the cold, filamentary gas that feeds galaxies at high redshift. Using a high-resolution cosmological simulation of a single massive protocluster, the authors argue that between z≈4.4 and 2.7 a major merger and AGN feedback heat the halo from the inside out, flattening density, temperature, and metallicity profiles. This internal heating destroys intermediate-column MgII absorbers, pushes oxygen to higher ionisation states, and transforms Lyman-alpha halo morphology from filamentary to spherical. The work matters because it identifies a concrete, physically driven pathway for the CGM-to-ICM transition at an epoch—z∼3—that is just becoming observable with current instruments.","feed_headline":"Mergers and AGN heat galaxy-cluster gas from the inside out","feed_subtitle":"A high-resolution protocluster simulation links destroyed cold gas to reshaping Lyman-alpha haloes by z≈2.7","key_machinery":"The central object is a high-resolution cosmological zoom-in simulation of a single massive protocluster—a z=0 ~10^15 Msun cluster progenitor—evolved to z=2.7 with a factor-of-512 increase in CGM mass resolution (average cell radius 200–300 pc). The analysis connects this simulation to observables through post-processing: photoionisation-equilibrium calculations yield mock MgII and OVII absorption, while Monte Carlo radiative transfer generates Ly-alpha and H-alpha emission maps including dust attenuation, resonant scattering, and optional AGN photoionisation. The load-bearing mechanism is the coupling of AGN feedback (thermal quasar-mode injection with a 25 Myr duty cycle and radio-mode bub","core_discovery":"The authors claim that the transition from a multiphase circumgalactic medium to a hot intracluster medium is driven primarily by internal processes—specifically a 1:2.5 major merger and AGN feedback in the quasar and radio modes—rather than by external accretion alone. In their simulated protocluster, gas is redistributed outward from 50–250 kpc, the volume-filling phase heats above 3×10^7 K, and metallicity flattens to the canonical ~0.3 Z_sun value from the inside out. The observable signatures follow: intermediate-column MgII absorbers are rapidly destroyed, leaving only dense clumps associated with satellites; OVII covering fractions first grow then drop as gas becomes over-ionised; and","pith_inferences":["If this inside-out transformation is generic rather than a quirk of one merger history, then surveys targeting z∼3 massive haloes should find a systematic deficit of strong MgII absorbers inside roughly 250 kpc, while OVII covering fractions rise then fall—a signature testable with stacked quasar spectra behind known protoclusters.","The authors' single-halo focus leaves the specific role of the 1:2.5 merger unquantified; a natural extension is a small ensemble of protocluster zoom-in simulations spanning different merger chronologies to separate merger-driven from feedback-driven heating.","The near-independence of the extended Ly-alpha profile from AGN photoionisation suggests that enormous Ly-alpha nebulae around quasars may be powered more by stellar emission, photon escape, and resonant scattering than by the quasar's own ionising photons; comparing radial Ly-alpha and H-alpha profiles in a broader sample could test this.","The predicted compact H-alpha haloes and the central Ly-alpha/H-alpha ratio gradient could be directly observed with JWST NIRSpec, providing a relatively clean test of whether unresolved ISM structure is the cause of the central Ly-alpha deficit."],"forward_implications":["Between z≈4.4 and 2.7, the halo's gas heats from the inside out; density, temperature, and metallicity profiles flatten as material is redistributed to larger radii.","Intermediate-column MgII absorbers (column densities between roughly 10^11 and 10^13 cm^-2) are preferentially destroyed, leaving a clumpier cold-gas distribution concentrated around satellite galaxies, while the densest satellite gas survives.","OVII becomes more widespread as the hot halo grows, then decreases in covering fraction in the inner 50–250 kpc as gas is heated beyond the OVII ionisation stage into the X-ray regime.","Extended Lyman-alpha haloes persist even in the absence of AGN photoionisation, and their morphology shifts from filamentary to more spherical as cold inflows are disrupted, with covering fractions consistent with the evolution from a typical nebula to an ELAN-like object during a major heating episode.","The fiducial dust-to-metal ratio of 0.1 underpredicts central Ly-alpha emission, implying more efficient nuclear Ly-alpha production or lower dust attenuation; H-alpha haloes are dimmer and more compact and could be probed with JWST.","Higher CGM resolution boosts predicted Ly-alpha surface brightness by about a factor of 2.8, showing that many existing cluster simulations may underestimate cold gas and Ly-alpha emission."],"fun_headline_variants":["Merger and AGN heat protocluster gas from the inside out","Inside-out heating from merger and AGN births the ICM in simulation","Merger and AGN reshape Lyman-alpha halo while heating protocluster gas","Inside-out gas heating by merger and AGN destroys MgII absorbers"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The subgrid feedback model—how AGN energy is injected into the surrounding gas—is assumed to deposit that energy at the right radii and times, so if the prescription misplaces energy, the inside-out transformation could be a numerical artifact rather than a physical trend.","fun_headline_variants_meta":{"raw":{"variants":["Merger and AGN heat protocluster gas from the inside out","Inside-out heating from merger and AGN births the ICM in simulation","Merger and AGN reshape Lyman-alpha halo while heating protocluster gas","Inside-out gas heating by merger and AGN destroys MgII absorbers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000973,"raw_usage":{"total_tokens":4070,"prompt_tokens":939,"completion_tokens":3131,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":683,"completion_tokens_details":{"reasoning_tokens":3051}},"tokens_in":683,"tokens_out":3131,"duration_ms":22972,"temperature":1.0,"reasoning_tokens":3051,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T17:57:30.672951+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the radial distribution of strong MgII absorbers around a statistical sample of z∼3 protoclusters (e.g., using MUSE or Keck spectroscopy): if intermediate-column absorbers persist within 250 kpc in a majority of merging systems, the proposed inside-out destruction of cold gas would be contradicted.","supporting_citations":[],"review_version":1}