{"id":"a727ec81-8889-425b-a723-970b432ea862","arxiv_id":"2412.14248","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"Short episodes of super-Eddington accretion triggered by major mergers can produce the overmassive black holes seen by JWST in the early Universe.","lead":"This paper uses a computer model of galaxy and black hole growth to explain why early galaxies contain black holes that seem too heavy for their stars. It finds that short, repeated bursts of very fast eating, triggered by galaxy collisions, can make black holes grow faster than their host galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The key assumption that major mergers trigger ~1 Myr super-Eddington bursts (Sec. 2.4, Eq. 4) is built into CAT rather than tested; Sec. 6 admits merger triggering is poorly explored, so the overmassive-BH explanation rests on an unverified causal link.","rationale":"The paper is a good-faith consistency scenario: it uses a calibrated semi-analytic model to show that episodic super-Eddington accretion can reproduce the overmassive BH population, the short duty cycles, and the bright end of the JWST AGN LF. It has genuine supporting evidence: the SE model populates the observed MBH-Mstar region while the EL model does not (Fig. 1), the evolutionary tracks match the dormant BH GN-1001830 (Sec. 4), and the predicted LF is in qualitative agreement with Akins et al. (2024) modulo SED assumptions. The main vulnerability is not internal inconsistency but external validity: the SE bursts are inserted by hand in Sec. 2.4 with a fixed rate and a stopping time tau_dyn/100, and the paper itself (Sec. 6) concedes that merger-triggered super-Eddington accretion is poorly explored in simulations. The cited hydro simulations are mostly isolated galaxies, so they do not validate the merger-trigger mechanism. This is a load-bearing concern because the abstract's central claim names major mergers as the trigger. However, it does not warrant rejection: the scenario is plausible and testable, and the paper is appropriately framed as 'can be explained.' The reader's conditional verdict is appropriate; I would keep it, possibly with an explicit request for a hydrodynamical test of merger-triggered bursts or a re-analysis of existing simulations. No change to the verdict.","tokens_in":21116,"tokens_out":8535,"duration_ms":75737,"concrete_test":"Extract from the high-resolution hydro simulations already cited (Sassano et al. 2023, Massonneau et al. 2023, Lupi et al. 2024a, Gordon et al. 2024) the directly measured statistics of super-Eddington accretion episodes: their durations, peak Eddington ratios, and whether each episode is triggered by a major merger. Compare these distributions with the CAT prescription (Mdot = 0.017 Mgas/10 Myr, Δt = τdyn/100, merger mass ratio >1/10). If the simulations show that major mergers do not preferentially trigger such bursts, or that the bursts are shorter/weaker by more than a factor of ~3, then the central explanation for overmassive BHs loses its physical foundation. Alternatively, run a new idealized major-merger simulation at z~6 with a seed BH and measure the same quantities.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. 2.4 the SE scenario sets the accretion rate to Mdot = ϵBH Mgas/τaccr (Eq. 4) during major mergers, and stops the burst after Δt = τdyn/100 or when Mgas/MBH < 10. Because τdyn/100 is typically ~1 Myr at z~6, the quoted 'predicted' burst durations of 0.5–3 Myr and duty cycles of 1–4% are largely direct outputs of this assumed stopping rule, not independent predictions. The central claim therefore depends on two unverified physical premises: that major mergers actually drive such short, intense accretion episodes, and that the episodes have roughly the assumed duration and rate. The paper's own Sec. 6 states that the role of major mergers in directly triggering/enhancing BH growth 'is still poorly explored with simulations,' and the hydro simulations cited for burst durations (Sassano et al. 2023; Massonneau et al. 2023; Lupi et al. 2024a; Gordon et al. 2024) are mostly isolated, non-merging gas-rich galaxies. If real merger-triggered super-Eddington accretion is substantially shorter, weaker, or absent, the predicted overmassive population, short duty cycles, and dormant fraction would not follow. This is a correctness risk in the external validity of the scenario, not an internal inconsistency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses the semi-analytical Cosmic Archaeology Tool (CAT) to test whether the overmassive black holes (MBH/Mstar up to ~0.1-0.4) found by JWST at z=4-7 can be produced by episodic super-Eddington (SE) accretion. CAT, built on GALFORM merger trees with Pop III and direct-collapse seeding, is run in two accretion modes: an Eddington-limited (EL) Bondi prescription (Eq. 3) and an SE mode in which major mergers (mass ratio > 1/10) drive accretion at Mdot = eps_BH Mgas/tau_accr (Eq. 4) for Delta t = tau_dyn/100 or until Mgas/MBH < 10 (Sec. 2.4). The authors find that the SE model populates the MBH-Mstar region of the JWST sample whereas the EL model does not; that more than 85% of the matched overmassive systems descend from light Pop III seeds; that BH-galaxy co-evolution starts only at z<8; that SE bursts last 0.5-3 Myr with duty cycles of a few percent; and that the SE model partially reproduces the JWST/LRD bolometric luminosity function at 5<z<7. They also identify synthetic counterparts for the dormant BH in JADES GN-1001830 and argue that such systems converge to the local MBH-Mstar relation within ~0.5 Gyr. The paper presents this as a consistency scenario and explicitly acknowledges that the SE prescription is simplified and that the merger-trigger link is not yet tested by simulations.","tokens_in":21549,"tokens_out":14891,"duration_ms":118512,"significance":"If the scenario is correct, the main implications are: (i) the overmassive BH population at z=4-7 does not require abundant heavy seeds, since light seeds can reach comparable masses through repeated SE bursts; (ii) the model predicts a largely dormant overmassive population with an AGN duty cycle of a few percent, which is testable with JWST samples; and (iii) the Little Red Dot population may be the active, transient phase of this growth mode. A clear strength is that the SE parameters were calibrated to quasar number counts in earlier work (Trinca et al. 2022, 2024), not to the JWST overmassive targets, so the MBH-Mstar match is not a trivial re-fit of the target data. The paper is also unusually candid about its limitations (Sec. 6). The main caveat is that the headline quantitative results (burst durations, duty cycles) are largely inherited from the assumed stopping rule of Sec. 2.4, and the merger-trigger hypothesis lacks direct simulation support; the result is therefore best read as a proof-of-concept consistency model rather than a unique explanation of the data.","major_comments":[{"comment":"The predicted burst durations and duty cycles are not independent outputs of the model. In Sec. 2.4 the SE phase is terminated by the condition Delta t = tau_dyn/100 (or Mgas/MBH < 10), and for the halos relevant at z ~ 5-6, tau_dyn/100 ~ 1-2 Myr, which directly produces the 0.5-3 Myr distribution reported in Sec. 3.3 (Fig. 4, median ~ 1 Myr). Over such a short interval the accretion of Eq. (4) removes only Delta MBH ~ 0.017 Mgas (Delta t/10 Myr) ~ 0.002 Mgas, so the gas-exhaustion condition is rarely the active stopper; the duration is essentially the assumed tau_dyn/100. Consequently, the abstract and conclusions present as model 'suggestions' (0.5-3 Myr bursts, 1-4% duty cycle) quantities that are largely prescribed. I recommend that the paper (i) state explicitly that these numbers are consequences of the adopted stopping rule, and (ii) add a robustness test varying Delta t (e.g., tau_dyn/30 and tau_dyn/300) and the product eps_BH/tau_accr, showing how the overmassive population, the duty cycle, and the LF normalization respond. This would convert a potential circularity into a quantitative sensitivity statement.","section":"Sec. 2.4 and Sec. 3.3"},{"comment":"The causal link between major mergers and SE accretion, which is the backbone of the central claim, is not directly supported by the simulations cited for the burst durations. As acknowledged in Sec. 6, the hydrodynamic simulations invoked (Sassano et al. 2023; Massonneau et al. 2023; Lupi et al. 2024a; Gordon et al. 2024; Shi et al. 2024) model isolated gas-rich galaxies hosting 10^3-10^4 Msun seeds, not merger-driven gas inflows around 10^6-10^8 Msun BHs. The statement in Sec. 6 that the role of major mergers 'is still poorly explored with simulations' is therefore in tension with the abstract's causal phrasing ('triggered by major galaxy mergers') and with the consistency claim made in Sec. 3.3. I ask the authors to separate explicitly in Sec. 3.3 what the simulations validate (the duration of SE accretion in dense, clumpy environments) from what remains a working assumption (that major mergers trigger such phases at the assumed rate and duration), and to soften the abstract and Sec. 7 wording accordingly unless a new test of the merger-trigger hypothesis is added.","section":"Sec. 3.3 and Sec. 6"},{"comment":"The claim that the SE model 'closely reproduces the high number density of accreting BHs with Lbol = 10^45-10^47 erg/s' (last bullet of Sec. 7) is not supported by Fig. 6: the text of Sec. 5 states that both the EL and SE luminosity functions drop by more than two orders of magnitude below the Akins et al. (2024) number densities at Lbol > 10^46.5 erg/s, and the agreement in the SE panel is limited to roughly Lbol ~ 10^45-10^46 erg/s. The conclusion bullet should be moderated to a partial agreement, or the paper should explicitly compute the LF under the stellar-contamination correction it invokes to reconcile the bright end. As written, the bullet overstates the quantitative support that the LF comparison provides.","section":"Sec. 5 and Sec. 7 (last bullet)"}],"minor_comments":[{"comment":"Duty-cycle values are quoted inconsistently: the abstract and Sec. 7 say 1-4%, Sec. 3.3 and Fig. 4 give 0.5%-6% as a function of MBH, and Sec. 6 says 1-6%. Please harmonize the numbers and specify that the duty cycle is mass-dependent.","section":"Abstract, Sec. 3.3, Sec. 6, Sec. 7"},{"comment":"There is a typo: 'the impact of jest remains debated' should read 'jets'. The author list also contains a formatting artifact ('Va liante'), presumably from the LaTeX source.","section":"Sec. 2.5"},{"comment":"The statement that 'less than 15% of the whole population of overmassive BHs at these epochs descends from heavy seeds' refers to the 78-system sample selected in Sec. 3.1 (10/78 ~ 13%); the text should clarify that this is a property of the JADES-Medium-matched selection, not of the full predicted population.","section":"Sec. 3.1 and Sec. 6"},{"comment":"The many free parameters (alpha, eps_SF, eps_BH, tau_accr, Jcrit, Zcrit, eps_w,SN, eps_w,AGN, and the assumed spin a=0.572) are introduced piecemeal; a summary table of parameters with adopted values and calibration targets would make the model setup and its calibration much more transparent.","section":"Sec. 2"},{"comment":"The comparison with the JWST sample in Fig. 1 is qualitative. A quantitative statistic (e.g., the fraction of observed objects falling within the predicted SE locus, or a two-dimensional two-sample test between the EL and SE distributions and the data) would strengthen the claim that the SE model 'better reproduces' the observed population than the EL model.","section":"Sec. 3 and Fig. 1"},{"comment":"The rationale for using a spin-dependent radiative efficiency (a=0.572) in the SE model but a fixed eps_r = 0.1 in the EL model is not given; since this asymmetry directly affects the luminosity boost used in Sec. 5, one justifying sentence is needed.","section":"Sec. 2.5"}],"recommendation":"major_revision","confidential_remarks":"This is a well-written consistency-scenario paper from a group with a substantial prior record on the CAT model; the incremental content beyond Trinca et al. (2022, 2023, 2024) is the targeted JWST overmassive-BH analysis, seed-origin statistics, duty-cycle estimates, and the GN-1001830 evolutionary tracks, which is a reasonable contribution for this journal. My main editorial concern is the gap between the cautious wording inside the text and the stronger claims in the abstract and conclusions (burst durations presented as findings, LF agreement overstated); all three major comments are fixable within a revision, and I would not regard the underlying scenario as disproved. The manuscript is well within the scope of the journal, and I see no novelty-disclosure problem, since the SE scenario and its parameters were published earlier and are properly cited."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper's core claim is that the overmassive BHs JWST sees at z~4-7 are a natural outcome of short, repeated super-Eddington bursts triggered by major mergers. The genuinely new pieces are the assembly histories, the duty-cycle and dormant-fraction estimates, and the concrete match to GN-1001830. The LF comparison in the Lbol ~ 1e45-1e47 range is also a real success for the scenario. The SE model itself was developed in Trinca et al. 2022/2023, so the novelty here is the application, not the machinery.\n\nThe main soft spot is the one the stress-test note flags. The burst durations of 0.5-3 Myr and the 1-4% duty cycles are not independent predictions; they are direct consequences of the stopping condition Δt = τdyn/100, which at z~6 is about 1 Myr. The paper acknowledges that the SE prescription is simplified, and Sec. 6 concedes that merger triggering of BH growth is poorly explored with simulations. The hydro evidence cited for such short bursts comes mostly from isolated, gas-rich galaxies, not merger-driven events. So the causal link from major mergers to these short intense episodes is assumed rather than demonstrated.\n\nThat said, the paper is careful to frame itself as a consistency scenario, not a unique explanation. Given that framing, the argument holds up: the SE model produces an overmassive population, low duty cycles, and a boosted LF, all in qualitative agreement with JWST. The parameters were not calibrated to the overmassive sample, so the circularity concern is mild. The sample selection does guarantee some of the overmassive property, but the authors are transparent about it.\n\nOther issues are moderate: the comparison with observed data is qualitative in places, and there is no public code or catalog, which makes independent checks harder. The LF comparison relies on the LRD SED decomposition, which is itself debated.\n\nBottom line: this is a solid, honest application of an established SAM to a hot observational puzzle. It deserves a serious referee. The quantitative predictions should be treated as scenario outputs rather than independent forecasts, but the paper will be useful to anyone working on high-z BH formation and the JWST overmassive population.","headline":"Solid consistency scenario connecting JWST overmassive BHs to episodic super-Eddington accretion, but the headline burst durations and duty cycles are built into the assumed stopping rule rather than independently derived.","tokens_in":22084,"tokens_out":2870,"would_cite":true,"duration_ms":25345,"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":"Short, merger-driven super-Eddington bursts can explain the overmassive black holes JWST sees at high redshift.","keywords":["supermassive black holes","super-Eddington accretion","overmassive black holes","galaxy mergers","high-redshift galaxies","active galactic nuclei","JWST","semi-analytical galaxy formation model"],"falsifier":"A decisive test would be measuring the AGN duty cycle of overmassive black holes at $z\\sim5$–$7$ (for example through variability or clustering): if a large fraction of these systems are found to be active more than a few percent of the time, or if actively accreting overmassive black holes are found in hosts that show no recent major-merger signatures, the merger-triggered super-Eddington scenario would be ruled out.","tokens_in":20885,"feed_emoji":"🕳️","tokens_out":8744,"duration_ms":62563,"temperature":0.7,"pith_summary":"JWST has found supermassive black holes at redshifts 4 to 7 that are far heavier relative to their host galaxies than the local scaling relation predicts. This paper argues that these \"overmassive\" black holes arise naturally if early black hole seeds grow not by steady accretion but by short, repeated bursts of super-Eddington accretion, each triggered by a major galaxy merger. Using the Cosmic Archaeology Tool semi-analytical model, the authors show that such bursts last only about 0.5 to 3 million years, so these black holes are active only 1 to 4 percent of the time and most are dormant at any given moment. If correct, this resolves the apparent tension between JWST detections and local scaling relations, and it explains the abundance of faint active galactic nuclei and the puzzling \"little red dots\" as a transient, merger-triggered phase.","feed_headline":"JWST's heavy black holes arise from short super-Eddington bursts","feed_subtitle":"A merger-driven growth model reproduces JWST's heavy black holes with duty cycles of only 1-4%.","key_machinery":"The load-bearing mechanism is the super-Eddington accretion prescription of Section 2.4: during a major galaxy merger (halo mass ratio $\\mu>1/10$), the black hole accretes at $\\dot{M} = \\epsilon_{\\rm BH} M_{\\rm gas}/\\tau_{\\rm accr}$ with $\\epsilon_{\\rm BH}=0.017$ and $\\tau_{\\rm accr}=10$ Myr, and the enhanced phase is terminated after $\\Delta t = \\tau_{\\rm dyn}/100$ or when the gas-to-black-hole mass ratio falls below 10. This prescription is what converts each major merger into a short, intense growth episode; it is what lets light seeds catch up with heavy seeds, sets the burst durations and duty cycles, and ultimately determines where the model lands in the $M_{\\rm BH}$–$M_{\\rm star}$ plane.","core_discovery":"The central claim is that the population of overmassive black holes observed by JWST at $4<z<7$, with black-hole-to-stellar-mass ratios $M_{\\rm BH}/M_{\\rm star}$ as high as $\\sim0.1$–$1$, is produced naturally when early black hole seeds grow through short, repeated episodes of super-Eddington accretion triggered by major galaxy mergers. In the CAT semi-analytical model, a major merger (halo mass ratio $\\mu>1/10$) drives gas onto the nuclear black hole at a rate $\\dot{M} = \\epsilon_{\\rm BH} M_{\\rm gas}/\\tau_{\\rm accr}$ with $\\epsilon_{\\rm BH}=0.017$ and $\\tau_{\\rm accr}=10$ Myr, stopping after $\\Delta t=\\tau_{\\rm dyn}/100$ or when $M_{\\rm gas}/M_{\\rm BH}<10$. With this prescription, the model reproduces the observed $M_{\\rm BH}$–$M_{\\rm star}$ locus, predicts that over $85\\%$ of overmassive systems descend from light Population III remnant seeds rather than heavy direct-collapse seeds, and finds that black-hole–galaxy co-evolution begins only at $z<8$, after about $30\\%$ of the final stellar mass has formed outside the black hole's host. The predicted bursts last $0.5$–$3$ Myr with a median around $1$ Myr, giving duty cycles of $1$–$4\\%$; most overmassive black holes are therefore dormant ($\\lambda_{\\rm Edd}<0.05$), consistent with the observed dormant system GN-1001830, and the luminosity boost of the active fraction matches the JWST AGN luminosity function, including the little red dots.","pith_inferences":["If the duty cycle is truly $1$–$4\\%$, the number of overmassive black holes in the JWST fields is several tens of times larger than the number caught in a bright phase; targeted variability or stacking searches for dormant nuclei could test this directly.","The same episodic growth mode may operate at later cosmic epochs, since overmassive black holes with similar properties have been reported at $z\\sim1$–$3$; if so, the mechanism is a generic feature of black hole growth rather than a high-redshift special case.","If super-Eddington accretion changes the geometry of the accretion disk, single-epoch virial black hole mass estimates could be biased, so the inferred $M_{\\rm BH}/M_{\\rm star}$ ratios and the apparent agreement with the model may shift once such biases are included.","The model's merger-triggering assumption could be stress-tested with high-resolution simulations that follow gas clump dynamics; if minor mergers or disk instabilities also trigger bursts, the predicted duty cycles and mass-ratio distribution would change."],"forward_implications":["Most overmassive black holes at $z=4$–$7$ should be dormant, so surveys should uncover many more inactive systems like GN-1001830 than bright AGNs.","The bright AGN phase and the little red dots are transient: each source shines for only about a million years per burst, so the same black hole can be observed in very different accretion states.","Light Population III remnant seeds can account for most overmassive black holes once super-Eddington bursts are allowed, erasing information about the seeding channel by $z\\sim8$.","Black-hole–galaxy co-evolution begins late (at $z<8$), implying early black hole growth is largely decoupled from the assembly of the host's stellar mass.","Overmassive systems converge to the local $M_{\\rm BH}$–$M_{\\rm star}$ relation in roughly 0.5 Gyr once bursts stop, explaining why such extreme ratios are not seen locally."],"supporting_citations":[{"why":"Supplies the Cosmic Archaeology Tool semi-analytical model and its calibration, including the super-Eddington accretion prescription used throughout.","marker":"Trinca et al. 2022"},{"why":"Provides the parameterization and calibration of the super-Eddington accretion efficiency and timescale adopted in the model.","marker":"Pezzulli et al. 2016"},{"why":"Provides the JWST sample of overmassive black holes at z=4-7 that the model aims to reproduce.","marker":"Maiolino et al. 2024b"},{"why":"Reports the dormant overmassive black hole GN-1001830 at z=6.67, the key observational benchmark for the predicted dormant population.","marker":"Juodžbalis et al. 2024"},{"why":"Defines the local M_BH-M_star scaling relation against which the overmassive offset is measured.","marker":"Reines & Volonteri 2015"},{"why":"Provides the bolometric luminosity function of little red dots that the super-Eddington model matches at L_bol~10^45-10^47 erg/s.","marker":"Akins et al. 2024"},{"why":"High-resolution hydrodynamical simulation cited by the paper as independent support for super-Eddington burst durations of a few Myr.","marker":"Sassano et al. 2023"}],"fun_headline_variants":["Super-Eddington burst cycles drive JWST's early overmassive black holes","Mergers trigger short super-Eddington episodes that build heavy black holes","Brief super-Eddington feasts grow the early Universe's overmassive BHs","Episodic super-Eddington accretion powers the first heavy black holes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results hinge on the assumption that major galaxy mergers actually drive brief, intense super-Eddington accretion episodes of roughly the assumed efficiency and duration; if real bursts are much weaker, shorter, or triggered by something other than major mergers, the predicted overmassive population, the ~1 Myr burst lengths, and the 1-4% duty cycles would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Super-Eddington burst cycles drive JWST's early overmassive black holes","Mergers trigger short super-Eddington episodes that build heavy black holes","Brief super-Eddington feasts grow the early Universe's overmassive BHs","Episodic super-Eddington accretion powers the first heavy black holes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000618,"raw_usage":{"total_tokens":2996,"prompt_tokens":1202,"completion_tokens":1794,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":818,"completion_tokens_details":{"reasoning_tokens":1709}},"tokens_in":818,"tokens_out":1794,"duration_ms":13525,"temperature":1.0,"reasoning_tokens":1709,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:23:30.036194+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be measuring the AGN duty cycle of overmassive black holes at $z\\sim5$–$7$ (for example through variability or clustering): if a large fraction of these systems are found to be active more than a few percent of the time, or if actively accreting overmassive black holes are found in hosts that show no recent major-merger signatures, the merger-triggered super-Eddington scenario would be ruled out.","supporting_citations":[],"review_version":1}