{"id":"a6660734-c966-435a-8cdf-daeae4ddeb00","arxiv_id":"2504.16807","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Using five lattice spacings with physical quark masses, the authors obtain Gamma(h_c->eta_c gamma)=0.604(24) MeV and the first lattice QCD estimate Gamma(h_b->eta_b gamma)=46.0(4.8) keV.","lead":"This lattice QCD paper computes the decay rates of the h_c and h_b quarkonium states into eta_c plus a photon and eta_b plus a photon, finding 0.604(24) MeV for charm and 46.0(4.8) keV for bottom. The bottom rate is the first direct QCD-based determination, relevant for searches for new particles that could mix with quarkonium states.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unquantified quark-disconnected diagrams are the most load-bearing gap: they affect both the direct h_c form factor and the entire h_b ratio ladder, with no numerical estimate of their size.","rationale":"The reader identified both the heavy-mass extrapolation and the neglected disconnected diagrams as weak assumptions. I agree that the disconnected diagrams are a real unquantified systematic, and I think they are more load-bearing than the extrapolation. The heavy-mass extrapolation is extensively cross-checked in Section III.C: three distinct NRQCD-motivated ansätze, multiple fit variants with and without power corrections, data cuts that drop the lightest one or two simulated masses, and a BAIC combination. The spread among ansätze (F_b^1 from −0.285 to −0.298) is reflected in the quoted 0.290(15), and the differences at the physical point are small because the heaviest simulated point (m_ηH ≈ 6.6 GeV) already sits close to the apparent asymptotic scaling. In contrast, the disconnected diagrams are dropped without any numerical estimate for the transition form factor; the paper's only quantitative support (Ref. [39]) concerns masses, not matrix elements. Since the h_c width is quoted at 4% precision and the h_b width inherits the connected-only approximation through the entire ratio ladder, a few-percent disconnected contribution would be comparable to or larger than the stated errors. This does not invalidate the paper, but it does mean the central results should be treated as CONDITIONAL on the disconnected contributions being negligible, exactly as the reader concluded. My concrete test would settle this directly, and I therefore recommend no change to the reader's verdict.","tokens_in":25773,"tokens_out":16043,"duration_ms":142921,"concrete_test":"Compute the full disconnected contribution to C^1_3pt for h_c→η_cγ on the D96 and E112 ensembles using the Hansen–Lupo–Tantalo method (Ref. [61]) or hierarchical probing for the light and charm loops in Fig. 1, and form the ratio to the connected F_c^1. If |F_c^1_dis/F_c^1_conn| is below 1.5% (i.e., less than half the quoted 1.9% error on F_c^1), the assumption is validated; if it is 5% or larger, the quoted uncertainties must be inflated and the comparison with experiment revisited. For the h_b channel, repeat the computation at m_H = λ^5 m_c on the E112 ensemble to check that the disconnected ratio remains negligible and does not alter the extrapolation to the b-quark mass.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claims—both the precise h_c width and the first lattice h_b width—are computed with only the quark-connected contribution to the three-point function. In Section II, diagrams (b)–(e) of Fig. 1 are dropped for F_c^1, with the justification that they are Zweig or SU(3) suppressed; footnote 2 cites Ref. [39] for charmonium masses, but no bound is given for the transition matrix element itself. The quoted error on F_c^1 is 0.010/0.522 ≈ 1.9%, and the h_c width error is 4%; a disconnected contribution of only 2–3% in the matrix element would shift the width by 4–6%, comparable to or larger than the quoted uncertainty. For the h_b channel, the same approximation enters not only through the normalization F1(m_c) in Eq. (42) but also through each connected-only ratio r_λ in Eq. (39), so a disconnected systematic could affect both the absolute scale and the mass dependence of the ladder. The paper's own Section V states that 'their impact can and should be studied separately,' acknowledging that this is an unquantified systematic. Because the new h_b result is the paper's main novelty, and the h_c result is presented as a significant precision improvement, the lack of any numerical estimate of the disconnected contribution leaves the error budget incomplete.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents lattice QCD computations of the hadronic matrix elements for the electric-dipole radiative decays h_c -> eta_c gamma and h_b -> eta_b gamma. The charmonium calculation is performed directly at the physical charm quark mass using N_f=2+1+1 ETMC ensembles at five lattice spacings, yielding F_c^1 = -0.522(10) and Gamma(h_c -> eta_c gamma) = 0.604(24) MeV. The bottomonium calculation uses the ratio method with heavy quark masses m_H = lambda^{n-1} m_c up to roughly 3 m_c, followed by a continuum extrapolation and then a heavy-quark extrapolation to m_b using three NRQCD-inspired functional forms. The authors obtain F_b^1 = -0.290(15) and Gamma(h_b -> eta_b gamma) = 46.0(4.8) keV, which they identify as the first lattice QCD estimate of this quantity.","tokens_in":26008,"tokens_out":6464,"duration_ms":65613,"significance":"If the quoted uncertainties are reliable, the charmonium result is a clear improvement over previous lattice determinations, and the bottomonium result would be a valuable first lattice-QCD-based prediction for a quantity currently known only from quark models and pNRQCD. The calculation benefits from physical sea quark masses, five lattice spacings, small cutoff effects, and a transparent continuum-limit analysis using the BAIC. The ratio method is used thoughtfully to reduce statistical and discretization errors. However, two load-bearing systematic issues remain: the quark-disconnected contributions are dropped without a numerical estimate, and the h_b result rests on an extrapolation in the heavy-quark mass across a region with no lattice data. These issues do not invalidate the central claims, but they need to be addressed before the quoted error bars can be taken at face value.","major_comments":[{"comment":"The neglect of quark-disconnected diagrams is an unquantified systematic that affects both central results. The quoted uncertainty on F_c^1 is about 1.9% (Table II and Eq. (24)), while the h_c width error is 4%; the h_b result inherits the same approximation both through F1(m_c) in Eq. (42) and through each ratio r_lambda in Eq. (39). The Zweig/SU(3) suppression arguments given in Section II (footnote to Fig. 1) indicate that the disconnected contribution should be small, but they do not provide a numerical bound. The reference cited there, Ref. [39], constrains charmonium masses, not the transition matrix element. Since Section V states that the impact 'can and should be studied separately,' the error budget is incomplete as it stands. Please provide a numerical estimate of the disconnected contribution, or add a conservative systematic uncertainty to both Gamma(h_c -> eta_c gamma) and Gamma(h_b -> eta_b gamma).","section":"Section II, Fig. 1; Section V"},{"comment":"The heavy-quark extrapolation is the load-bearing step for the first h_b result. The lattice data extend only to m_H around 3 m_c (Table IV), while the physical point is at m_b/m_c about 4.58, so the fit is extrapolated across an additional factor of roughly 1.5 in m_H with no data in that window. The three ansaetze give F_b^1 = -0.285(10), -0.298(20), and -0.293(15), and the BAIC combination gives -0.290(15); this spread is an estimate of model variation only among fits that share the same NRQCD-motivated scaling assumption. If the effective heavy-quark velocity scaling changes outside the fitted window, all three forms could be biased in the same direction, and the quoted 5% error would be an underestimate. Please quantify the extrapolation uncertainty in a way that does not rely entirely on the BAIC weight, for example by showing the sensitivity to the fit range or by assigning an additional model-selection uncertainty based on the envelope of the ansaetze.","section":"Section III.C, Eqs. (51), (53), (55)"}],"minor_comments":[{"comment":"The caption says 'all four gauge ensembles' but Table I and the text describe five ensembles; this should be corrected.","section":"Fig. 3 caption"},{"comment":"There is a duplicated definite article in 'except for the the coarsest lattice'; please fix this typo.","section":"Abstract"},{"comment":"The label 'BES-IIII' in the left panel should read 'BESIII'.","section":"Fig. 16"},{"comment":"The expression for the smearing radius r_0(m_H) should match the definition given in Section II for r_0; please check whether the square root over n(m_H) has been omitted in the displayed formula.","section":"Eq. (30)"},{"comment":"It would be helpful to state explicitly whether the A48 ensemble, with m_pi about 175 MeV, is treated as a physical-mass ensemble in the continuum extrapolation or whether a separate light-quark-mass systematic is assigned.","section":"Section II.B"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid ETMC contribution and the charmonium part is in good shape. The main risk is the h_b error budget, particularly the unquantified disconnected diagrams and the model dependence of the m_H -> m_b extrapolation. I believe these can be addressed in a revision without changing the overall approach, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The new thing here is real: this is the first lattice QCD estimate of the h_b -> eta_b gamma form factor, F_b^1 = -0.290(15), with a width of 46.0(4.8) keV. The h_c -> eta_c gamma width, 0.604(24) MeV, is a factor of 2.3 more precise than the previous lattice result and agrees with experiment. The setup is good: physical sea quark masses with Nf = 2+1+1, five lattice spacings down to about 0.05 fm, clear plateaus, and continuum extrapolations that look controlled, especially for the charmonium form factor where chi^2/dof is about 0.2.\n\nThe bottomonium part is more delicate. The data go up to about three times the charm mass, and the results at the physical b mass come from three NRQCD-inspired fit forms. The authors are appropriately agnostic and fold the spread across forms into the quoted error, so the 5% uncertainty on F_b^1 reflects real model variation within their set. Still, the assumption that these forms hold from m_H ~ 3 m_c to m_b is exactly that: an assumption. It is a reasonable one for a first estimate, and the consistency among the ansaetze helps, but it is not a controlled extrapolation in the same way the a^2 continuum limit is.\n\nThe more load-bearing gap, and the one the stress-test note correctly flags, is the disconnected diagrams. They are dropped everywhere, for both the direct h_c form factor and for each rung of the h_b ratio ladder. The arguments for suppression are plausible, and the paper cites a lattice study showing tiny effects on charmonium masses, but no numerical bound is given for the transition matrix elements themselves. For the h_c width the quoted error is about 4%; a disconnected contribution of only a few percent in the matrix element would move the width by more than that. The paper itself says in Section V that the impact of these diagrams can and should be studied separately, which is honest but also an admission that the error budget is incomplete.\n\nOverall the central claims are very likely correct, and the charmonium result in particular is clean and publishable. The bottomonium result is a valuable first step but should be treated as a first lattice estimate, not a benchmark, until the disconnected contribution is quantified or convincingly bounded.\n\nYes, send this to a serious referee. I would cite it. It deserves peer review, and the right referee will push for a numerical estimate of the disconnected effect and a sharper discussion of the m_H -> m_b extrapolation, but the paper is not fatally flawed.","headline":"First lattice QCD result for h_b -> eta_b gamma plus a sharpened h_c width; solid charm part, bottom part rests on unquantified disconnected diagrams and a phenomenological heavy-mass extrapolation.","tokens_in":26648,"tokens_out":1779,"would_cite":true,"duration_ms":20282,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["12.38.Gc","13.40.Hq"],"model":"deepseek-v4-flash","headline":"Lattice QCD gives Γ(h_c→η_c γ)=0.604(24) MeV and the first prediction Γ(h_b→η_b γ)=46.0(4.8) keV.","keywords":["lattice QCD","radiative decay","charmonium","bottomonium","transition form factor","heavy quark extrapolation","NRQCD scaling","twisted mass fermions"],"falsifier":"Perform the same computation with a lattice spacing fine enough to set a bottom quark directly (or with a relativistic bottom action) and compare the resulting $F_b^{1}$ with −0.290(15); alternatively, measure Γ(h_b→η_b γ) experimentally, since the paper's value implies Γ(h_b)=88(13) keV through the measured branching fraction, and a future direct measurement outside that window would rule out the extrapolation.","tokens_in":25534,"feed_emoji":"⚛️","tokens_out":7505,"duration_ms":65327,"temperature":0.7,"pith_summary":"The paper argues that lattice QCD can compute the electric-dipole radiative decays of both charmonium and bottomonium from first principles. For the charmonium decay h_c→η_c γ it works directly at the physical charm mass and obtains a form factor $F_c^{1}$=-0.522(10), giving Γ(h_c→η_c γ)=0.604(24) MeV, in agreement with experiment and 2.3 times more precise than previous lattice values. For the bottomonium decay h_b→η_b γ, the b quark is too heavy for direct simulation, so the authors compute for six heavy-quark masses from m_c up to about 3m_c and extrapolate with NRQCD-motivated scaling laws, obtaining the first lattice QCD estimate $F_b^{1}$=-0.290(15) and Γ(h_b→η_b γ)=46.0(4.8) keV. These quantities matter because radiative quarkonium transitions probe the nonperturbative hyperfine structure of QCD and because observed rates involving η_b could be altered by pseudoscalar particles mixing with it.","feed_headline":"First lattice prediction: h_b→η_b γ width is 46.0 keV","feed_subtitle":"First-principles QCD now predicts the bottomonium photon decay and sharpens the charmonium decay by 2.3×.","key_machinery":"The load-bearing object is the single on-shell transition form factor F_1 ≡ F_1(0) that encodes all nonperturbative QCD in the decay rate, extracted from a three-point correlation function of η_c and h_c interpolating operators with the electromagnetic current, and placed at q²=0 with twisted boundary conditions. For the b case, the paper computes ratios r_λ(m_H)=F_1(m_H)/F_1(m_H/λ) at consecutive heavy-quark masses (the ratio method, introduced in Ref. [33]) to suppress statistical errors and cutoff effects, reconstructs F_1(m_H) from F_1(m_c), and extrapolates to m_b with several NRQCD-motivated scaling ansätze, including F_1√m_{η_H}=const and variants with Coulombic or string-potential velocity scaling, combined through a Bayesian information criterion.","core_discovery":"The central discovery is that the transition matrix elements for h_c→η_c γ at q²=0 can be computed with sub-2% precision, and that the same connected-diagram computation extended to fictitious heavy-quark mesons follows a clean scaling law F_1(m_H)√m_{η_H} ≈ const that permits a controlled extrapolation to the physical b quark. In the continuum limit the authors report Γ(h_c→η_c γ)=0.604(24) MeV from $F_c^{1}$=-0.522(10) and, for the first time in lattice QCD, Γ(h_b→η_b γ)=46.0(4.8) keV from $F_b^{1}$=-0.290(15). The h_c result agrees with the experimental measurement, and it improves on previous lattice determinations by using physical sea quark masses and five lattice spacings. The paper also notes that combining its widths with measured branching fractions yields total widths of Γ(h_c)=1.007(78) MeV and Γ(h_b)=88(13) keV, more precise than direct measurements.","pith_inferences":["If the dropped quark-disconnected diagrams shift these electric-dipole form factors no more than they shift charmonium masses (a few MeV out of ~3 GeV), the central values stand; the paper leaves that magnitude unquantified, so the natural next check is to compute them with the method cited for future work.","The same ratio-method ladder could be extended to a direct bottom-quark simulation on finer or anisotropic lattices, turning the m_H→m_b extrapolation into an interpolation and testing the stability of F_b^1.","A precise future measurement of Γ(h_b→η_b γ) would not only test QCD but also sharpen the inferred total width of h_b, which is the quantity that searches for pseudoscalar admixtures, such as axion-like particles or extra Higgs states, would use."],"forward_implications":["The h_c→η_c γ rate becomes the most precise lattice determination to date, sharpening the comparison with experiment and making the combined lattice-plus-experiment total width Γ(h_c)=1.007(78) MeV the best available constraint on that state.","Combining Γ(h_b→η_b γ)=46.0(4.8) keV with the measured branching fraction yields Γ(h_b)=88(13) keV, a prediction of a total width that has not been directly measured.","The clean observation F_1√m_{η_H}≈const over the simulated mass range provides a benchmark that potential-NRQCD descriptions of E1 quarkonium transitions should reproduce.","The b-quark form factor, being the first lattice value, gives a first-principles point of comparison for quark-model and pNRQCD predictions, two of which currently sit about 2σ from the lattice width."],"supporting_citations":[{"why":"Supplies the previous lattice determination of F_c^1 and the Gaussian-smearing, twisted-mass methodology that this paper follows and improves.","marker":"[30]"},{"why":"Introduces the ratio method used to build r_λ(m_H) and suppress heavy-quark discretization and statistical errors.","marker":"[33]"},{"why":"Provides the NRQCD scaling law Γ∝|k|³/(m_H² v_H²) that motivates the heavy-mass extrapolation ansätze.","marker":"[49]"},{"why":"Gives the pNRQCD prediction of h_b→η_b γ used as the main theoretical comparison for the extrapolated form factor.","marker":"[52]"},{"why":"Reports the experimental branching fraction and total width used to compare Γ(h_c→η_c γ) and to form the combined total-width estimate.","marker":"[53, 54]"},{"why":"Reports the measured branching fraction for h_b→η_b γ used with the lattice width to predict Γ(h_b).","marker":"[19, 20]"},{"why":"Documents the N_f=2+1+1 ensembles, renormalization constants, and mixed-action setup that provide all lattice data.","marker":"[36]"},{"why":"Earlier quenched lattice result for Γ(h_c→η_c γ) used as one of the comparison points.","marker":"[29]"}],"fun_headline_variants":["First lattice QCD width for h_b→η_b γ: 46.0 keV","Lattice QCD sharpens h_c decay, predicts h_b decay","2.3× more precise lattice QCD rate for h_c→η_c γ","First lattice QCD prediction: h_b→η_b γ width 46 keV","Lattice QCD: h_c→η_c γ rate matches experiment, h_b predicted"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extrapolation from the simulated heavy quark masses (up to about three times the charm mass) to the physical b quark assumes that the scaling laws fitted to those masses keep working all the way to the b quark.","fun_headline_variants_meta":{"raw":{"variants":["First lattice QCD width for h_b→η_b γ: 46.0 keV","Lattice QCD sharpens h_c decay, predicts h_b decay","2.3× more precise lattice QCD rate for h_c→η_c γ","First lattice QCD prediction: h_b→η_b γ width 46 keV","Lattice QCD: h_c→η_c γ rate matches experiment, h_b predicted"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0007,"raw_usage":{"total_tokens":3253,"prompt_tokens":1128,"completion_tokens":2125,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":744,"completion_tokens_details":{"reasoning_tokens":2016}},"tokens_in":744,"tokens_out":2125,"duration_ms":14168,"temperature":1.0,"reasoning_tokens":2016,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:55:05.541803+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform the same computation with a lattice spacing fine enough to set a bottom quark directly (or with a relativistic bottom action) and compare the resulting $F_b^{1}$ with −0.290(15); alternatively, measure Γ(h_b→η_b γ) experimentally, since the paper's value implies Γ(h_b)=88(13) keV through the measured branching fraction, and a future direct measurement outside that window would rule out the extrapolation.","supporting_citations":[],"review_version":1}