{"id":"de9a034b-e4b2-4655-be1a-7f240ffa4327","arxiv_id":"2412.18318","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"No primordial-black-hole gravitational-wave background is found in LVK O1-O3 data, yielding 95% upper limits on the PBH dark-matter fraction for log-normal mass functions.","lead":"Using the first three observing runs of LIGO-Virgo-KAGRA, the authors search for a gravitational-wave background from merging primordial black holes and find no signal. The non-detection produces upper limits on the fraction of dark matter that could be primordial black holes across a broad mass range.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Early-binary suppression formulas are applied to broad log-normal widths (sigma up to 1) that the authors themselves say the formulas are not validated for, so the quoted f_PBH limits rest on an unvalidated model regime.","rationale":"The reader's weakest assumption is the calibrated reliability of the early- and late-binary merger-rate formulas, including the suppression factors in Appendix A. My stress-test sharpens that concern: the paper's own prior allows sigma up to 1, which is a broad mass distribution, while Appendix A explicitly says the suppression expressions need validation for broad distributions and that claims based on these rates may be premature. This is a concrete internal tension rather than a general uncertainty about the PBH binary formation model. It is load-bearing because the f_PBH upper limits are obtained by marginalizing over sigma, and sigma is not constrained by the data, so the unvalidated regime contributes to every quoted limit. The concrete test (re-running with narrow sigma) would settle whether the broad-sigma regime actually affects the numbers. If the limits are insensitive to sigma, the caveat is not numerically important for the headline claims; if they shift substantially, the constraints must be re-reported with the sigma restriction or with a systematic uncertainty from the rate model. I do not think this requires rejection: the statistical analysis and non-detection are solid, and the paper does disclose the caveat in an appendix. The appropriate verdict remains CONDITIONAL, consistent with the reader's original assessment, so I recommend no change to the verdict.","tokens_in":16665,"tokens_out":5051,"duration_ms":49136,"concrete_test":"Fix sigma to 0.05 and 0.1 (rather than the log-uniform prior extending to 1) in the nested-sampling analysis and recompute the 95% upper-limit curves in Fig. 1 and Table II using the same O1-O3 cross-correlation data and likelihood. Compare the f_PBH limits at mu = 30, 100, and 10^3 M_sun with the marginalized results. If any limit shifts upward by more than 30%, the broad-sigma regime is driving the headline constraints, and the paper's own Appendix A caveat invalidates the quoted bounds as unconditional exclusions; if the limits are essentially unchanged, the concern is mitigated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quoted 95% upper limits on f_PBH are obtained after marginalizing over the log-normal width sigma over the full prior range 10^-2 to 1. The early-binary suppression factors S1 and S2 (Appendix A, Eqs. 9-14) are only calibrated for narrow mass distributions: the authors state that 'the validity of the above expressions for a broad mass distribution necessitates detailed analytical investigation or validation through N-body simulations' and that 'making claims based on these merging rates may be premature.' For sigma = 1, the log-normal mass function spans several decades, giving <m^2>/<m>^2 = e ~ 2.7 instead of the narrow-limit value ~1, and the physical processes entering S1 (binary disruption by nearby PBHs, matter fluctuations) are not captured by the same analytic approximations. Because sigma is essentially unconstrained by the data (flat posterior, Appendix D), the central quantities such as f_PBH < 5.3e-3 at 30 M_sun (Rclust=1) average over this unvalidated region of the model. If the true broad-distribution suppression is stronger than Eqs. 9-14 predict, the GWB from early binaries is lower and the inferred f_PBH upper limits are too stringent, potentially by a large factor. The headline improvement over previous constraints is therefore not yet robust until the analysis is restricted to the validated narrow-sigma regime or the suppression factors are tested for broad distributions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper searches for a gravitational-wave background (GWB) from primordial black hole (PBH) binaries using the public LIGO-Virgo-KAGRA O1-O3 cross-correlation data. The signal model includes early binaries formed in radiation domination and late binaries formed by dynamical capture, with a log-normal PBH mass function and an additional astrophysical CBC power-law component. The authors perform a Bayesian analysis, find no significant signal (log Bayes factors ~ -1), and derive 95% upper limits on the PBH dark-matter fraction f_PBH as a function of the mean mass µ, marginalizing over the width σ and the CBC amplitude; for example, f_PBH < 5.3e-3 at µ = 30 M_sun for R_clust = 1. They claim their results improve previous GWB constraints by a factor of 2-10 at 100 M_sun.","tokens_in":16991,"tokens_out":6093,"duration_ms":55943,"significance":"If the quoted bounds are robust, they strengthen the exclusion of PBH dark matter in the roughly 10-300 M_sun mass range using the stochastic background rather than resolved events, and the paper demonstrates a useful full-spectral-shape approach that goes beyond the power-law upper limits usually reported by the LVK. The statistical machinery is standard, the likelihood and data handling follow well-established pygwb procedures, and the injection study in Appendix C is a welcome validation exercise. However, the quantitative bounds rest on the early-binary merger-rate formulas whose suppression factors are acknowledged by the authors themselves to be unvalidated for the broad mass distributions included in their prior; this limits the certainty of the headline numbers and the quoted improvement over previous work.","major_comments":[{"comment":"The analysis marginalizes over σ up to 1 (Table I), yet the early-binary suppression factors S1 and S2 in Eqs. (9)-(14) are, as the authors state in Appendix A, only valid for narrow mass distributions: 'the validity of the above expressions for a broad mass distribution necessitates detailed analytical investigation or validation through N-body simulations,' and 'making claims based on these merging rates may be premature.' Since σ is essentially unconstrained by the data (flat posterior, Figs. 4-6), the quoted limits in Table II and Fig. 1 include models for which the predicted merger rate—and hence the mapping from non-detection to f_PBH—is not reliable. For σ = 1 the mass function spans decades in mass and <m^2>/<m>^2 = e ≈ 2.7, moving substantially away from the narrow-distribution limit, so the suppression factors can be materially different. This is load-bearing because the f_PBH bounds scale inversely with the assumed merger rate; if broad-distribution suppression is stronger than Eqs. (9)-(14) predict, the limits would be too stringent. Please either restrict the analysis to the validated narrow-σ regime (with a clear statement of what 'narrow' means) or provide N-body/analytic validation of the suppression factors for the full prior range.","section":"Appendix A and Sec. II"},{"comment":"The injection study in Appendix C shows that σ and R_clust are not well recovered, and the corner plots in Appendix D confirm an essentially flat posterior for σ. Because the amplitude of the early-binary GWB depends on σ not only through the mass function but also through the suppression factors S1 and S2, marginalizing over σ does not merely integrate over a nuisance parameter; it averages over a range of model predictions whose reliability is not established for the larger σ values. The paper should report how the upper limits change if σ is restricted to, say, σ ≤ 0.3 or σ ≤ 0.5, so the reader can see the sensitivity of the headline f_PBH bounds to the unvalidated regime.","section":"Appendix C and Appendix D"}],"minor_comments":[{"comment":"The sentence 'We concentrate on relatively narrow mass distributions (σ<1)' is misleading, because σ = 1 is not narrow: the log-normal mass function then spans a factor of e^4 ≈ 55 around µ. Please rephrase and make the prior choice in Table I consistent with the stated focus.","section":"Sec. II"},{"comment":"The calibration uncertainty parameter λ is introduced in Eq. (7) but its prior is not specified in the main text; please state the prior used (e.g., the Gaussian calibration-error model from the cited pygwb studies) or refer the reader to the relevant appendix or documentation.","section":"Sec. III, Eq. (7)"},{"comment":"The two-significant-figure entries in Table II give a false impression of precision given the model uncertainty documented in Appendix A; consider quoting one significant figure or adding a note about model dependence.","section":"Table II"},{"comment":"The claim that the paper presents 'for the first time, a direct GWB search specifically targeting PBH mergers' is overstated, since Refs. [25,27,51,55] already interpreted LVK stochastic data in terms of PBH merger backgrounds, and Ref. [55] used a full spectral model rather than a power-law approximation. I suggest softening the novelty claim to emphasize that this is the first combination of early and late formation channels with a full spectral shape.","section":"Conclusion"}],"recommendation":"major_revision","confidential_remarks":"The main technical concern is the mismatch between the stated narrow-distribution validity of the early-binary rate formulas and the broad σ prior that is marginalized over in the headline limits; this is fixable in a revision by restricting σ or by validating the suppression factors. The novelty claim in the conclusion should also be checked against Refs. [25,27,51,55]. The paper is otherwise appropriate for the journal and appears to be conducted in good faith with standard tools."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a competent and useful stochastic-background analysis, but the headline f_PBH limits are conditional on merger-rate formulas that the authors themselves don't trust in the regime they actually analyze. The non-detection is solid; the derived dark-matter exclusions are softer than the abstract implies.\n\nWhat's genuinely new: they run a full Bayesian search over a log-normal PBH mass function with both early and late binary formation channels, using the LVK O1–O3 cross-correlation data, and they marginalize over mean mass, width, f_PBH, and an astrophysical CBC component. That is broader than the earlier single-mass-bin treatments in Refs. [42,55]. The injection study is a nice sanity check, and the recovered Omega_CBC upper limit (~3e-9) is consistent with the LVK result. The statistical machinery is standard and the non-detection is robust: Bayes factors favor noise.\n\nThe load-bearing soft spot is Appendix A. The suppression factors S1 and S2 are written for narrow mass distributions, and the text says explicitly that their validity for broad distributions requires N-body validation and that \"making claims based on these merging rates may be premature.\" Then the main analysis takes sigma up to 1 and marginalizes over it. The posterior for sigma is flat, so the quoted f_PBH limits average over the unvalidated region. If the true suppression is stronger at large sigma, the limits are too strong; if weaker, too weak. Either way, the systematic is not in the error bars. This doesn't undermine the non-detection, but it does mean the abstract's \"constraints on f_PBH\" should carry a model-systematic asterisk.\n\nSecond, the conclusion claims \"for the first time, a direct GWB search specifically targeting PBH mergers.\" That's an overstatement; Refs. [42,55] are direct searches, just with narrower mass models. The first-time claim should be limited to the broad log-normal mass function with both formation channels.\n\nMinor point: fixing R_clust to three reference values is a reasonable choice, and they show the dependence is weak, but at high masses the late-binary channel dominates, so those numbers are conditional on R_clust as well.\n\nOverall, the central argument holds up as far as it goes. The data handling and likelihood are sound, and the paper is useful for anyone working on PBH dark-matter constraints via stochastic backgrounds. It deserves a serious referee. The referee should ask for either a narrow-sigma analysis or a quantification of the suppression-factor systematic, and should push back on the novelty claim. I'd engage with it after those revisions.","headline":"Solid Bayesian stochastic-search paper with a real caveat: the headline f_PBH limits rely on suppression formulas the authors themselves flag as unvalidated for broad mass distributions, so the constraints are model-conditional, not absolute dark-matter exclusions.","tokens_in":17511,"tokens_out":3868,"would_cite":true,"duration_ms":38266,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A Bayesian search of the first three observing runs finds no primordial black hole background and sets 95% upper limits that largely exclude PBH dark matter from 10 to 300 solar masses.","keywords":["gravitational wave background","primordial black holes","stochastic gravitational wave background","dark matter constraints","Bayesian inference","binary black hole mergers","log-normal mass function","early and late binary formation"],"falsifier":"Recompute the 95% upper limits using merger rates calibrated to a large $N$-body simulation of PBH binary formation that includes cluster disruption and accretion; if the resulting limits on $f_{\\rm PBH}$ are weaker than the quoted values (e.g., $f_{\\rm PBH}<5.3\\times10^{-3}$ at $30\\,M_\\odot$), the rate normalization behind the central claim is falsified. Alternatively, a future detector measurement of a PBH GWB below the predicted amplitude for $f_{\\rm PBH}=1$ would show the formulas over-predict the background.","tokens_in":16495,"feed_emoji":"🌌","tokens_out":15703,"duration_ms":112290,"temperature":0.7,"pith_summary":"Using the cross-correlation data from the first three observing runs of the ground-based gravitational-wave detector network, this paper searches for a gravitational-wave background (GWB) produced by binaries of primordial black holes (PBHs), modelling both early- and late-formation channels and a log-normal mass function. No such background is found, and the null result is converted into 95% upper limits on the fraction $f_{\\rm PBH}$ of dark matter in PBHs as a function of mean mass. At 30 solar masses the limit reaches $f_{\\rm PBH} < 5.3\\times10^{-3}$ (for a reference clustering parameter $R_{\\rm clust}=1$), and the constraints improve previous GWB-based bounds by a factor of about 2–10 at 100 solar masses. If correct, these limits strengthen the case that PBHs do not constitute all of the dark matter in the roughly 10–300 solar mass window.","feed_headline":"PBH dark matter largely ruled out from 10 to 300 solar masses","feed_subtitle":"A Bayesian search of O1–O3 data tightens previous stochastic-background limits by 2 to 10 times at 100 solar masses","key_machinery":"The central object is the isotropic gravitational-wave background model $\\Omega_{\\rm GW}(f) = \\Omega_{\\rm CBC}(f) + \\Omega_{\\rm EB}(f) + \\Omega_{\\rm LB}(f)$, where $\\Omega_{\\rm CBC}$ is a fiducial astrophysical compact-binary (CBC) power-law background and $\\Omega_{\\rm EB}$ and $\\Omega_{\\rm LB}$ are the early- and late-binary PBH contributions computed from the differential merger rates in Eqs. (5) and (6). The early-binary rate carries suppression factors $S_1$ and $S_2$ for matter inhomogeneities and Poisson clustering, while the late-binary rate is scaled by a clustering parameter $R_{\\rm clust}$. This spectral shape is compared with the cross-correlation estimator from all available detector baselines through a Gaussian likelihood, and Bayesian nested sampling provides posterior distributions and 95% upper limits on $f_{\\rm PBH}$, $\\sigma$, $\\mu$, and $\\Omega_{\\rm CBC}$.","core_discovery":"The paper's central claim is that, under a log-normal PBH mass distribution with width $\\sigma<1$ and the standard early/late binary merger-rate prescriptions, the absence of a detectable PBH gravitational-wave background in the first three observing runs rules out $f_{\\rm PBH}$ values above roughly $10^{-2}$--$10^{-3}$ for mean masses around 30 solar masses and provides upper limits that improve earlier stochastic-background constraints by a factor of 2–10 at 100 solar masses. The limit is obtained by a full Bayesian fit to the shape of the cross-correlation spectrum rather than by comparing a single power-law bound to a predicted amplitude, and the paper states that this is the first direct GWB search specifically targeting PBH mergers.","pith_inferences":["The reliability of the quoted upper limits rests on the merger-rate formulas; if updated $N$-body simulations show that early binaries are disrupted more easily than the $S_1S_2$ prescription encodes, all $f_{\\rm PBH}$ limits would become weaker.","The same pipeline could be applied to broad mass functions, such as the QCD phase-transition peak, once their merger rates are better understood; the paper explicitly leaves this to future work.","A future positive detection of a GWB whose spectrum deviates from a power law in the way PBH models predict would let this method distinguish primordial from astrophysical binaries, rather than merely set upper limits.","Because the late-binary contribution depends on the clustering parameter $R_{\\rm clust}$ and dominates at high masses, GWB searches of this type could also constrain PBH clustering, not just PBH abundance."],"forward_implications":["If the central claim is correct, PBHs with a narrow log-normal mass distribution cannot make up all of the dark matter for mean masses in the roughly 10–300 $M_\\odot$ range.","The late-binary formation channel dominates the GWB for mean masses above about $3\\times10^2\\,M_\\odot$, so future low-frequency gravitational-wave detectors could extend the $f_{\\rm PBH}$ constraints to heavier PBHs.","The full-spectrum Bayesian approach improves previous power-law-based GWB constraints on $f_{\\rm PBH}$ by a factor of about 2–10 at $100\\,M_\\odot$.","The posteriors show no significant PBH or CBC background, so the data are consistent with noise, with logarithmic Bayes factors near $-1$."],"supporting_citations":[{"why":"Supplies the O1 cross-correlation data used in the stochastic search.","marker":"[56]"},{"why":"Supplies the O2 cross-correlation data and the power-law spectral model.","marker":"[57]"},{"why":"Provides the O3 upper-limit data and the fiducial astrophysical CBC spectrum included in the model.","marker":"[58]"},{"why":"Gives the analytical suppression factors $S_1$ and $S_2$ for early PBH binaries that set the predicted background amplitude.","marker":"[26]"},{"why":"Extends the early-binary suppression treatment and is one of the previous GWB constraints the paper improves upon.","marker":"[27]"},{"why":"Pioneered the gravitational-wave background signature of primordial black hole binaries on which this search builds.","marker":"[50]"},{"why":"Introduces the Bayesian parameter-estimation framework for stochastic backgrounds from which the likelihood is taken.","marker":"[76]"},{"why":"Provides the single-binary inspiral-merger-ringdown energy spectrum used in the $\\Omega_{\\rm GW}$ integral.","marker":"[68]"},{"why":"Supplies the cross-correlation statistic and calibration-uncertainty treatment used to build the estimator.","marker":"[81]"}],"fun_headline_variants":["PBH dark matter nearly excluded in 10-300 solar mass range","Gravitational wave search rules out most PBH dark matter masses","No PBH background: dark matter fraction constrained to 0.1-1%","Bayesian GWB search tightens PBH dark matter limits 10x","O1-O3 data rule out PBH dark matter from 10 to 300 M_sun"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the early- and late-binary merger-rate formulas (Eqs. (5) and (6), with the Appendix A suppression factors) correctly describe how many PBH mergers produce a given $f_{\\rm PBH}$; if the true rates are lower, every quoted upper limit on $f_{\\rm PBH}$ is too strong, and the paper itself cautions that claims based on these rates may be premature.","fun_headline_variants_meta":{"raw":{"variants":["PBH dark matter nearly excluded in 10-300 solar mass range","Gravitational wave search rules out most PBH dark matter masses","No PBH background: dark matter fraction constrained to 0.1-1%","Bayesian GWB search tightens PBH dark matter limits 10x","O1-O3 data rule out PBH dark matter from 10 to 300 M_sun"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000172,"raw_usage":{"total_tokens":1197,"prompt_tokens":790,"completion_tokens":407,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":406,"completion_tokens_details":{"reasoning_tokens":302}},"tokens_in":406,"tokens_out":407,"duration_ms":4220,"temperature":1.0,"reasoning_tokens":302,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T04:48:42.641867+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the 95% upper limits using merger rates calibrated to a large $N$-body simulation of PBH binary formation that includes cluster disruption and accretion; if the resulting limits on $f_{\\rm PBH}$ are weaker than the quoted values (e.g., $f_{\\rm PBH}<5.3\\times10^{-3}$ at $30\\,M_\\odot$), the rate normalization behind the central claim is falsified. Alternatively, a future detector measurement of a PBH GWB below the predicted amplitude for $f_{\\rm PBH}=1$ would show the formulas over-predict the background.","supporting_citations":[{"cited_title":"Upper limits on the stochastic gravitational-wave background from advanced ligo’s first observing run","cited_arxiv_id":null,"evidence_quote":"Supplies the O1 cross-correlation data used in the stochastic search."},{"cited_title":"Search for the isotropic stochastic back- ground using data from advanced ligo’s second observing run","cited_arxiv_id":null,"evidence_quote":"Supplies the O2 cross-correlation data and the power-law spectral model."},{"cited_title":"Upper limits on the isotropic gravitational-wave background from advanced ligo and advanced virgo’s third observing run","cited_arxiv_id":null,"evidence_quote":"Provides the O3 upper-limit data and the fiducial astrophysical CBC spectrum included in the model."},{"cited_title":"Stochas- tic gravitational-wave background due to primordial binary black hole mergers","cited_arxiv_id":null,"evidence_quote":"Pioneered the gravitational-wave background signature of primordial black hole binaries on which this search builds."},{"cited_title":"Inspiral- merger-ringdown waveforms for black-hole binaries with nonprecessing spins","cited_arxiv_id":null,"evidence_quote":"Provides the single-binary inspiral-merger-ringdown energy spectrum used in the $\\Omega_{\\rm GW}$ integral."},{"cited_title":"Detection prospects of gravitational waves from su (2) axion inflation","cited_arxiv_id":null,"evidence_quote":"Supplies the cross-correlation statistic and calibration-uncertainty treatment used to build the estimator."}],"review_version":1}