{"id":"ed12c5a0-1be3-4ff2-b775-af47a095c561","arxiv_id":"2607.25707","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"First lattice QCD extraction of the h_c to gamma eta and gamma eta-prime electric dipole and longitudinal form factors finds E1(0) suppressed relative to experiment, and predicts h_c to l+l- eta(eta') Dalitz distributions.","lead":"This paper reports the first lattice QCD calculation of the form factors and decay rates for the charmonium meson h_c decaying to a photon plus an eta or eta-prime meson. It finds the computed rates sit below measured values, repeating a pattern seen for J/psi decays on the same lattices, and it makes predictions for the unmeasured Dalitz decays h_c to lepton pairs plus eta or eta-prime.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Excited-state contamination in the Eq. 6 three-point fits is only partially tested; a bias in J would propagate directly to E1(0) and the claimed suppression, so the conditional verdict remains appropriate pending a stronger test.","rationale":"The central claim is a first lattice determination of E1 and C1 for h_c -> gamma eta(eta-prime) and a suppression relative to experiment. For that claim to hold, the matrix elements J must be unbiased. The most vulnerable step is the time-dependence fit, because it relies on a model assumption rather than a direct spectral decomposition. Other concerns, such as the single lattice spacing, unphysical pion mass, and parameterization spread in E1(0), affect precision and physical interpretation but not the internal validity of the extraction; even the largest parameterization spread (roughly 0.0128-0.0197 GeV for h_c -> gamma eta) preserves the qualitative suppression and the eta-prime/eta ratio, so it is not the load-bearing issue. The paper deserves credit for the ground-state versus excited-state h_c consistency check, the overlapping momentum data, the multiple parameterizations, and the comparison with the J/psi calculation; these make the extraction believable but do not eliminate the need for a stronger excited-state test. Therefore the conditional verdict is appropriate and unchanged.","tokens_in":18923,"tokens_out":10736,"duration_ms":104764,"concrete_test":"Re-analyze a representative subset of the h_c -> gamma eta and h_c -> gamma eta-prime correlation functions, especially those with n_eta = [220], [221], [300] and n_eta-prime = [200], [210], [211], using (i) fit forms with two source and two sink exponentials and (ii) a GEVP-based three-point analysis with a 2x2 operator basis (J/psi and h_c in the helicity +/-1 irreps; eta and eta-prime in the pseudoscalar channel). If the AIC-averaged J shifts by more than the quoted statistical error for any of these configurations, the extracted E1(0) values are biased and the suppression claim needs revision; if shifts are negligible, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The extraction of J from C(t, Delta t) uses Eq. 6, a constant plus at most one source and one sink exponential, with the AIC model average choosing among time-windows and these limited fit forms. For h_c helicity +/-1 channels, the h_c is an excited state above J/psi; the optimized operator should suppress J/psi, but any residual overlap with J/psi, psi-prime, or multi-particle states such as eta pi cannot be represented if there are two comparable exponentials, and the fit will compromise over time-windows. The same limitation applies to eta-prime, which is itself an excited state in the pseudoscalar channel. Appendix A compares h_c-as-ground-state (helicity 0) with h_c-as-excited-state (helicity +/-1) for h_c -> gamma eta and finds agreement; this is a useful test, but it uses the same Eq. 6 fit family for both cases, so a common bias from the fit form would not appear, and it is not shown for eta-prime. Appendix B checks simultaneous versus independent Delta t fits, again within the same form family. No multi-particle operators are included. Since J feeds directly into Gamma and then into E1, C1, and E1(0), a bias here would change the central values and the claimed suppression relative to experiment. The concern is not that the paper is wrong, but that the key systematic is not fully excluded by the presented tests.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a lattice QCD calculation of the radiative transition form factors for h_c -> gamma eta and h_c -> gamma eta'. It uses anisotropic three-flavor ensembles with m_pi ~ 391 MeV, optimized operators to access the h_c in boosted irreps where it can be an excited state, three-point correlation functions at several source-sink separations with simultaneous time-window fits and AIC model averaging, and a linear-system solution to separate the electric dipole (E1) and longitudinal (C1) form factors at each Q^2. The authors report best estimates |E1^{eta}(0)| = 0.0195(19) GeV and |E1^{eta'}(0)| = 0.0556(30) GeV, which imply radiative widths below the BESIII values, with a suppression pattern similar to their earlier J/psi -> gamma eta(eta') calculation, and an eta'/eta ratio of 2.9(3) compatible with the J/psi ratio. They also present first lattice results for psi' -> gamma eta(eta'), finding signals statistically compatible with zero.","tokens_in":19278,"tokens_out":9245,"duration_ms":85759,"significance":"This is, to my knowledge, the first lattice QCD determination of the h_c radiative multipole form factors, and it demonstrates a viable method for extracting matrix elements of an excited charmonium state in flight and for separating two form factors at fixed Q^2. The internal validation is in many respects careful: overlapping momentum points agree, the E1/C1 systems are overconstrained at most Q^2, ground-state and excited-state h_c extractions are compared in Appendix A, simultaneous and independent Delta-t fits are compared in Appendix B, and the Q^2 dependence is tested against several parameterizations. If the systematic concerns raised below are resolved, the results will provide a valuable nonperturbative benchmark for charmonium radiative transitions and for the ongoing discussion of the discrepancy between lattice determinations and BESIII for J/psi -> gamma eta(eta'). The paper is clearly a strong technical contribution from an experienced collaboration.","major_comments":[{"comment":"The extraction of the matrix elements J is the pivot of the entire analysis, but the excited-state systematics are not fully controlled. Equation (6) allows only one source and one sink exponential in addition to a constant; for the h_c helicity +/-1 channels the h_c is an excited state above the J/psi, and the eta' is itself an excited state in the pseudoscalar channel, so residual contamination from the J/psi, psi', or multi-particle states such as eta-pi cannot be represented if the contamination has more than one comparable exponential. Appendix A checks ground-state versus excited-state h_c extractions for h_c -> gamma eta, and Appendix B checks simultaneous versus independent Delta-t fits, but both use the same fit family, so a common bias would not be visible, and no analogous check is shown for h_c -> gamma eta'. Because J feeds directly into E1(Q^2), C1(Q^2), and E1(0), this is a load-bearing systematic. The authors should either add fits with additional exponentials or operators, or at least demonstrate stability of E1(0) under a wider class of fit forms and time windows, with the h_c -> gamma eta' channel explicitly included.","section":"Section III, Eq. (6), and Appendices A and B"},{"comment":"The quoted 'best estimate' uncertainties do not appear to cover the spread among the parameterizations. For h_c -> gamma eta, the z-poly order-1 fit gives |E1(0)| = 0.0128(11) GeV while the best estimate is 0.0195(19) GeV and the dipole and z-poly order-2 fits give about 0.0197 GeV; the spread is roughly 0.007 GeV, several times the quoted error of 0.0019 GeV. For h_c -> gamma eta', the z-poly order-1 value is 0.0445(21) versus the best estimate 0.0556(30), again a spread of about 0.011 GeV. The text says that variation over parameterization choice is considered in the best estimate, but neither the combination algorithm nor the error budget is specified. Please state explicitly how the central values and errors are constructed (e.g., AIC-weighted average, an envelope over the fits, or a stated choice of representative parameterization), and either enlarge the quoted errors to cover the parameterization dependence or provide a quantitative justification for excluding the low z-poly order-1 values.","section":"Section V, Figure 13, and Appendix D"}],"minor_comments":[{"comment":"The axis label 'square_i K_E1 K_C1' in the upper panel appears garbled; it should indicate the kinematic factors K_E1 and K_C1, and the lower panel's 'K_i' should be 'K_i'.","section":"Figure 5"},{"comment":"Figure 13 labels entries as 'z poly-4' although Appendix D states that z-polynomial fits were considered up to cubic order; please clarify whether a quartic fit was performed and, if so, tabulate its parameters.","section":"Figure 13 and Appendix D"},{"comment":"The statement that the 30 largest AIC values are retained and used in a model average would benefit from a specification of the total number of fits considered and the exact weighting formula.","section":"Section III"},{"comment":"The two exponential forms are defined in Appendix D, but the figure captions in Section IV refer to 'exp fit order 1' and 'exp fit order 2' without explicit definitions; please label them consistently in the main text.","section":"Section IV and Appendix D"},{"comment":"The eigenvalue plateaus for the h_c are visibly noisier and shorter than those for the J/psi; a quantitative statement of the selected t-range used in the GEVP and the resulting overlap factors would help the reader judge the quality of the optimized-operator projection.","section":"Section II, Figure 1"},{"comment":"The conclusion that the discrepancy is 'most likely' due to properties of the eta and eta' on this particular lattice is a plausible conjecture but is not tested here; consider softening the wording to reflect the alternatives listed in the introduction (quark-mass dependence, discretization, topological sampling).","section":"Section V"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a technically advanced lattice calculation with an interesting first result, but the two load-bearing issues (excited-state control and the error budget for E1(0)) should be addressed before publication. I do not see a fundamental flaw; a revision that adds an eta' consistency test and a transparent combination of parameterizations would make the paper acceptable. The paper fits the scope of the journal and the citation practice appears normal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First lattice QCD results for h_c -> gamma eta(eta-prime), with a clean separation of E1 and C1 and a believable central value. I think the paper deserves serious refereeing, and conditional acceptance is about right.\n\nWhat's new: no previous lattice calculation of these transitions. The authors extend their J/psi technology to access h_c as an excited state, separate E1 and C1 via overconstrained linear systems at fixed Q2, and introduce simultaneous Delta-t fitting. The extra psi-prime -> gamma eta(eta-prime) results, statistically compatible with zero, are honest scaffolding rather than overclaimed. The internal consistency is good: overlapping momentum points agree, ground-state and excited-state h_c extractions agree for the eta channel, and the parameterizations give E1(0) values within a modest spread. The eta'/eta ratio, 2.9(3), compared to 3.3(3) for J/psi on the same lattices, and the dimensionless E1/F ratio agreement with experiment, are the most persuasive structural checks. The qualitative suppression relative to experiment is credible as a real lattice effect tied to eta/eta' properties, not a numerical accident.\n\nSoft spots: the paper is one lattice spacing, m_pi ~ 391 MeV, so the size of the suppression can't be pinned down quantitatively until there's a continuum extrapolation and physical quark masses. The chi-squared cut on Q2 points, and the exclusion of the most timelike points, are reasonable but post hoc; it would be better to see retained and discarded points in the final fits. The stress-test concern about Eq. 6 is real: a constant plus at most one source and one sink exponential cannot capture multi-particle contamination, and Appendix A's ground-state vs excited-state check uses the same fit family and is shown only for eta, not eta-prime. That said, the overall consistency across Delta-t and between independent and simultaneous fits limits how large that bias can plausibly be. This is a moderate systematic, not a load-bearing flaw. No data/code release is a disappointment, though the paper is detailed enough to reproduce with effort.\n\nWho is this for: lattice QCD practitioners and charmonium phenomenologists; anyone interested in the eta/eta' anomaly. I would send it to a serious referee. I'd suggest requiring a clearer statement of the excited-state systematics, ideally a multi-particle operator test or at least an eta-prime Appendix A analogue, and public data if the journal allows.","headline":"First lattice QCD results for h_c radiative decays, internally consistent and credible; the main caveat is excited-state systematics from the three-point fit forms.","tokens_in":19891,"tokens_out":2201,"would_cite":true,"duration_ms":20050,"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":"The first lattice-QCD calculation of the h_c radiative decays finds both real-photon decay rates below the measured values, with the η′/η amplitude ratio matching the J/ψ pattern.","keywords":["lattice QCD","charmonium","h_c meson","radiative decay","electric dipole form factor","longitudinal form factor","eta-prime meson","Dalitz decay"],"falsifier":"Compute $|E_1(0)|$ for both channels on a second lattice ensemble with a different spacing, with physical pion mass, or with well-characterized gauge-field topology sampling: if the values rise toward the BESIII decay widths, the suppression is an artefact of these lattices, while if they stay low the suppression is a physical feature of the $\\eta/\\eta'$ system. A cheaper probe is to redo the three-point fits with multi-particle operators added and check whether the extracted matrix elements near $Q^2 = 0$ move outside their quoted errors, which would put the blame on the time-window fits rather than on the $\\eta$–$\\eta'$ physics.","tokens_in":18658,"feed_emoji":"⚛️","tokens_out":19037,"duration_ms":139918,"temperature":0.7,"pith_summary":"The paper reports the first lattice QCD determination of the form factors controlling the radiative decays $h_c \\to \\gamma\\eta$ and $h_c \\to \\gamma\\eta'$, separating the electric dipole ($E_1(Q^2)$) and longitudinal ($C_1(Q^2)$) multipoles for this $J^{PC}=1^{+-}$ charmonium state. Best estimates of $|E_1(0)| = 0.0195(19)$ GeV and $0.0556(30)$ GeV translate through the decay-rate formula into radiative widths that lie significantly below the BESIII measurements — the same suppression previously found for $J/\\psi \\to \\gamma\\eta^{(\\prime)}$ on the same lattices — while the $\\eta'/\\eta$ amplitude ratio of $2.9(3)$ matches the $J/\\psi$ value of $3.3(3)$. The paper reads this as evidence that the discrepancy lives in the $\\eta$–$\\eta'$ sector of this ensemble, most plausibly in how the gauge-field topology sampling affects their flavor-singlet content, and it supplies concrete predictions for the as-yet-unmeasured Dalitz decays $h_c \\to \\ell^+\\ell^-\\eta^{(\\prime)}$.","feed_headline":"First lattice QCD h_c decay rates fall short of experiment","feed_subtitle":"Both channels echo the earlier J/ψ suppression, pointing the discrepancy at the η–η′ mesons.","key_machinery":"The load-bearing object is the covariant decomposition of the vector-current matrix element into two invariant form factors, the electric dipole $E_1(Q^2)$ and the longitudinal $C_1(Q^2)$, with kinematic weights fixed by the helicities and momenta; by construction the longitudinal piece vanishes for a real photon, so the decay rate is controlled by $E_1(0)$ alone. On the lattice, each computed matrix element is a known linear combination of these two unknowns, so measuring many momentum, irrep, and helicity combinations at the same $Q^2$ and solving the overconstrained system $\\Gamma = K \\cdot F$ separates the multipoles. The inputs to that system come from three-point correlation functions fitted simultaneously across four source–sink separations $\\Delta t/a_t = \\{16, 20, 24, 28\\}$, using fit forms of a constant plus at most one source and one sink exponential, combined by AIC model averaging; optimized charmonium operators from a two-point generalized eigenvalue problem are what isolate the $h_c$ in irreps where it is an excited state above the $J/\\psi$. Dipole, exponential, and conformal-$z$ polynomial parameterizations then carry the discretely sampled form factors to $Q^2 = 0$.","core_discovery":"An axial meson such as the $h_c$ ($J^{PC} = 1^{+-}$) decaying to a pseudoscalar through the vector current is described by two invariant form factors, an electric dipole $E_1(Q^2)$ and a longitudinal $C_1(Q^2)$, whose kinematic weights depend on the helicities selected. Using three-point correlation functions on three-flavor lattices with $m_\\pi \\sim 391$ MeV, optimized operators that isolate the $h_c$ even in irreps where it sits above the $J/\\psi$, and an overconstrained linear system $\\Gamma = K \\cdot F$ solved at each virtuality, the paper determines both form factors for $h_c \\to \\gamma\\eta$ at 53 retained $Q^2$ points and for $h_c \\to \\gamma\\eta'$ at 63 points. The central results are the real-photon couplings $|E_1^{h_c \\to \\gamma\\eta}(0)| = 0.0195(19)$ GeV and $|E_1^{h_c \\to \\gamma\\eta'}(0)| = 0.0556(30)$ GeV, which through Eqn. (2) give radiative decay rates significantly below the BESIII experimental values. The paper argues the suppression is common to both $J/\\psi$ and $h_c$ decays to $\\eta^{(\\prime)}$, because the $\\eta'/\\eta$ ratio $2.9(3)$ is compatible with the same-lattice $J/\\psi$ ratio $3.3(3)$ and a dimensionless electric-versus-magnetic dipole comparison agrees with experiment for both mesons; the likely common cause is a property of the $\\eta$ and $\\eta'$ on this lattice, most plausibly their SU(3) flavor-singlet content and its sensitivity to gauge-field topology. First results for $\\psi' \\to \\gamma\\eta^{(\\prime)}$, extracted with the same operators, are compatible with zero.","pith_inferences":["If the suppression is really a topology-sampling artefact, the extracted $E_1(0)$ values should shift when the gauge ensemble is binned or reweighted by topological charge; that correlation is directly testable on the existing configurations.","The same mechanism would predict a similar suppression for any light-meson final state produced through a gluonic intermediate state on these lattices, which could be checked by computing other flavour-singlet-dominated radiative channels.","The simultaneous-$\\Delta t$ fitting with AIC model averaging is a transferable recipe for three-point function analyses beyond radiative transitions, wherever multiple source–sink separations are computed but currently fitted independently.","A lattice with physical pion mass and well-sampled topology that recovers the experimental rates would confirm the paper's conclusion, while one that still falls short would instead indicate a genuinely physical suppression of the gluonic production of $\\eta/\\eta'$ — with consequences for how charmonium radiative decays are used as a light-meson factory."],"forward_implications":["The real-photon decay rates $\\Gamma(h_c \\to \\gamma\\eta)$ and $\\Gamma(h_c \\to \\gamma\\eta')$ computed on these lattices lie significantly below the BESIII measurements, so on this ensemble the lattice description of the $\\eta$–$\\eta'$ system, rather than the charmonium transition, is the prime suspect for the gap.","The $\\eta'/\\eta$ amplitude ratio of $2.9(3)$ for $h_c$ is compatible with the $3.3(3)$ found for $J/\\psi$ on the same lattices, consistent with both amplitudes being governed by the SU(3) flavor-singlet content of the final-state mesons produced through a gluonic intermediate state.","A dimensionless ratio of the $h_c$ electric-dipole strength to the $J/\\psi$ magnetic-dipole strength agrees with experiment for both $\\eta$ and $\\eta'$, supporting the paper's picture that the charmonium annihilation factorizes from the $\\eta/\\eta'$ production.","The timelike $Q^2$ dependence of both $E_1$ and $C_1$ determines the Dalitz decays $h_c \\to \\ell^+\\ell^-\\eta^{(\\prime)}$, with $E_1$ dominant for the $\\eta$ final state but both multipoles significant across the range, giving angular and $q^2$ distributions that experiment can test.","First lattice results for $\\psi' \\to \\gamma\\eta^{(\\prime)}$ are statistically compatible with zero across the sampled $Q^2$ range, so the data cannot yet constrain the measured $\\psi'$ radiative widths."],"supporting_citations":[{"why":"The BESIII measurements of the h_c → γη and h_c → γη′ decay widths; the experimental benchmark the lattice rates fall below.","marker":"[2]"},{"why":"The companion letter reporting the lattice calculation of J/ψ → γη^(′), the process whose approach and suppression pattern this work extends.","marker":"[3]"},{"why":"The full method paper for J/ψ → γη^(′) on the same lattices: supplies the optimized-operator technology, the improved vector current, the Wigner–Eckart averaging, and the comparison ratio 3.3(3).","marker":"[4]"},{"why":"The spectrum calculation whose large operator basis underlies the optimized operators used to isolate the h_c, ψ′, η, and η′.","marker":"[5]"},{"why":"The source of the covariant matrix-element decomposition adopted in Eqn. (1), first presented for the equivalent 0⁺ → 1⁻ case.","marker":"[8]"},{"why":"The redefinition of the longitudinal form factor C_1 that fixes its normalization in this paper.","marker":"[9]"},{"why":"The AIC model-averaging procedure used to combine time-window fits and assign the quoted uncertainties on the matrix elements.","marker":"[12]"}],"fun_headline_variants":["Lattice QCD h_c decay rates lag experiment like J/ψ","First h_c radiative rates from lattice: suppressed vs data","h_c → γη(′) widths: lattice vs experiment shortfall","New lattice h_c decays mirror J/ψ suppression","Lattice h_c decays to γη and γη′ fall short"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results assume that the fitted time dependence of the three-point correlation functions — a constant plus at most one source and one sink exponential, combined by AIC model averaging — removes all contamination from states other than the desired one, including multi-particle states and the nearby J/ψ and ψ′, an assumption that the consistency checks in Appendices A and B never probe with multi-particle operators.","fun_headline_variants_meta":{"raw":{"variants":["Lattice QCD h_c decay rates lag experiment like J/ψ","First h_c radiative rates from lattice: suppressed vs data","h_c → γη(′) widths: lattice vs experiment shortfall","New lattice h_c decays mirror J/ψ suppression","Lattice h_c decays to γη and γη′ fall short"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00028,"raw_usage":{"total_tokens":1754,"prompt_tokens":1135,"completion_tokens":619,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":751,"completion_tokens_details":{"reasoning_tokens":530}},"tokens_in":751,"tokens_out":619,"duration_ms":5345,"temperature":1.0,"reasoning_tokens":530,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:25:16.556607+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $|E_1(0)|$ for both channels on a second lattice ensemble with a different spacing, with physical pion mass, or with well-characterized gauge-field topology sampling: if the values rise toward the BESIII decay widths, the suppression is an artefact of these lattices, while if they stay low the suppression is a physical feature of the $\\eta/\\eta'$ system. A cheaper probe is to redo the three-point fits with multi-particle operators added and check whether the extracted matrix elements near $Q^2 = 0$ move outside their quoted errors, which would put the blame on the time-window fits rather than on the $\\eta$–$\\eta'$ physics.","supporting_citations":[],"review_version":2}