{"id":"4d6587c3-94c9-46bd-8df0-c23d5150b1e4","arxiv_id":"1908.07802","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"The double-virtual gamma*gamma* to eta_c transition form factor is computed from light-front wave functions and is predicted to violate factorization strongly while depending only weakly on the asymmetry of the two photon virtualities.","lead":"This paper derives a light-front wave function description of the transition form factor for charmonium states eta_c(1S) and eta_c(2S) produced by two off-shell photons. It predicts that this form factor does not factorize into single-photon parts and depends mainly on the average photon virtuality, an effect that could be tested at Belle II.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Qbar^2 scaling and near-omega-independence are kinematically enforced by the common z-peaked nonrelativistic LFWF input; the model-independence claim is asserted but not shown across potentials.","rationale":"The reader's CONDITIONAL verdict is well placed. My read agrees that the central claim is forward-looking and not circularly fitted to the BaBar data, and the LFWF derivation has useful internal checks: the normalization reduces to the known two-photon width in the nonrelativistic limit, and the Bose symmetry of the representation is verified numerically. However, the most load-bearing weakness is more specific than the reader's phrasing: the Terentev kinematic map in Eq. (2.12) is actually exact for an equal-mass on-shell two-body state, so the concern is the nonrelativistic Schrodinger wave-function input and the single-Fock truncation, which all five potentials share. That shared construction enforces a narrow z-distribution around z=1/2, and Eq. (2.40) shows that this directly produces the claimed Qbar^2 scaling and near-omega-independence. The paper's own admission that the all-potential R comparison is not shown means the 'almost model independent' claim is supported only by an assertion, not by displayed evidence. This justifies the conditional acceptance: the derivation may well be correct, but the headline robustness requires an external probe of the z-distribution or at least a quantified all-potential comparison. I therefore keep the verdict at CONDITIONAL and recommend no change from the reader's assessment.","tokens_in":16501,"tokens_out":23521,"duration_ms":212730,"concrete_test":"Recompute F(Q_1^2,Q_2^2) and R for eta_c(1S) at fixed Qbar^2 = 50 GeV^2 while scanning omega, using at least one LFWF not built from Eq. (2.12) (e.g., the covariant light-front quark model of Ref. [18], or the same u(p) with an explicitly boosted relativistic spinor factor), and plot R for all five potentials. If the spread in R across potentials or the shift from the relativistic LFWF exceeds about 20%, or if a visible omega-dependence appears, then the claimed model-independent factorization breaking and Qbar^2 scaling are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline scaling is largely built into the LFWF-construction ansatz rather than being an independent output. For each of the five potentials, the rest-frame Schrodinger wave function u(p) is mapped to the LFWF through Eq. (2.12), p_z = (z-1/2) M_{c\\bar c}. Because u(p) is a nonrelativistic wave function peaked at p/m_c ~ 0.3, the resulting LFWF is sharply peaked at z = 1/2. Inserting z = 1/2 into Eq. (2.40) makes both denominators functions of (Q_1^2 + Q_2^2)/4 (plus k^2 and m_c^2), so the claimed scaling in Qbar^2 and the near independence from omega are largely enforced by the narrow z-distribution. The five-potential scan cannot expose this, because all five potentials are processed through the same nonrelativistic construction and the same single-c\\bar c Fock truncation. The paper itself states that the R pattern \"looks very similar for different potentials (not shown explicitly here)\" in Section III.B, so the \"almost model independent\" claim is unquantified. If the true eta_c LFWF has a broader z-distribution, from relativistic corrections or higher Fock states, the terms in Eq. (2.40) proportional to (1-z)^2 Q_2^2 and z^2 Q_2^2 no longer combine into Qbar^2 only, and both R(Q_1^2,Q_2^2) and the omega-dependence can change. The kinematic map Eq. (2.12) itself is exact for equal-mass on-shell two-body states, so the load-bearing assumption is the nonrelativistic Schrodinger wave-function input and the single-Fock truncation, rather than the Terentev boost per se.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives a light-front wave function (LFWF) representation of the gamma* gamma* -> eta_c(1S,2S) transition form factor F(Q1^2,Q2^2) for two spacelike photons. The LFWFs are obtained by solving the Schroedinger equation for five different c-cbar potentials and converting the rest-frame momentum-space wave functions to light front via the Terentev prescription. The authors compare the single-virtual limit to BaBar data for eta_c(1S), discuss the delayed onset of the Brodsky-Lepage asymptotic behavior, present two-dimensional distributions of F(Q1^2,Q2^2), and introduce a factorization-breaking ratio R(Q1^2,Q2^2). The central claims are that factorization is strongly broken, that R is almost model independent across the five potentials, and that the form factor scales approximately with Qbar^2 = (Q1^2+Q2^2)/2, with almost no dependence on the asymmetry parameter omega.","tokens_in":16944,"tokens_out":5491,"duration_ms":58008,"significance":"If the advertised model independence holds, the paper provides a concrete, falsifiable prediction for double-tag measurements at Belle II: the gamma* gamma* -> eta_c form factor depends mainly on the average virtuality Qbar^2 and not on the virtuality asymmetry. The derivation from Eq. (2.38) to Eq. (2.40) is transparent, the normalization is checked against the two-photon decay width, and the comparison with BaBar data is a genuine forward calculation with no parameters fitted to the transition form factor itself. The main weakness is that the headline claims of model independence and Qbar^2 scaling rest on a shared nonrelativistic LFWF construction and are not quantitatively demonstrated across the potentials.","major_comments":[{"comment":"The abstract's claim that factorization breaking is \"almost model independent\" is not supported by the presented evidence. Only the Buchmueller-Tye potential is plotted in Fig. 9, and the text states \"The factorization breaking pattern looks very similar for different potentials (not shown explicitly here)\" without quantification. Since this is a central claim, please show R(Q1^2,Q2^2), or a projection such as R as a function of Qbar^2 at fixed omega, for all five potentials, and quantify the spread across potentials. Without this, the headline claim is an assertion rather than a demonstrated result.","section":"Section III.B, Fig. 9"},{"comment":"The Terentev prescription is stated to be valid for weakly bound nonrelativistic systems, but charmonium is only moderately nonrelativistic, with typical p/m_c around 0.3-0.5. Because all five potentials are converted through the same nonrelativistic mapping, the resulting LFWFs are all sharply peaked at z approximately 1/2. Inserting z approximately 1/2 into Eq. (2.40) makes the denominators depend mainly on Q1^2+Q2^2, so the claimed Qbar^2 scaling and near-omega-independence are largely enforced by the construction rather than emerging independently from the dynamics. The five-potential scan cannot expose this limitation because it varies only the shape of u(p) within the same ansatz. Please add a quantitative discussion of the sensitivity of R(Q1^2,Q2^2) and of the omega-dependence to a broader z-distribution, for example by comparing with a relativistic treatment or by varying m_c over a wider range, and qualify the model-independence claim accordingly.","section":"Section II.A, Eqs. (2.12)-(2.20), and Eq. (2.40)"},{"comment":"The factorization-breaking ratio R is defined using the single-virtual form factors F(Q1^2,0) and F(0,Q2^2) computed in the same model. While this is a sensible normalization, it means that model dependence in the single-virtual form factor largely cancels in R. The statement that R is almost model independent should therefore be accompanied by a statement about how much of the model dependence is removed by this normalization, and the residual spread across potentials should be shown explicitly. Otherwise the reader cannot distinguish a genuine dynamical prediction from a consequence of the ratio construction.","section":"Section III.B, Eq. (3.3)"}],"minor_comments":[{"comment":"The notation switches between Q_i^2 = -q_i^2 at the beginning of Section II and q1^2, q2^2 in the final form factor expressions. Please state explicitly that q_i^2 in Eqs. (2.38) and (2.40) denotes the spacelike virtuality Q_i^2, or introduce a consistent sign convention, to avoid confusion in the denominators.","section":"Section II.C, Eqs. (2.38) and (2.40)"},{"comment":"The sentence \"The factorization breaking pattern looks very similar for different potentials (not shown explicitly here)\" should either be replaced by a quantitative statement with a figure or removed, since it is precisely the evidence needed for the main claim.","section":"Section III.B"},{"comment":"The phrase \"rate of approaching\" should be \"rate of approach\" or \"approach to.\"","section":"Fig. 5 caption"},{"comment":"Reference [14] is written as \"Phys. Rev. D97, 0094034 (2018)\"; the article number appears to be a typo and should read 094034.","section":"References"},{"comment":"The sentence \"The reader is asked to notice much broader range of Q^2 in the figure compared to that in previous figures\" is informal and should be rephrased.","section":"Section III.B, paragraph before Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The calculation is competently done and the double-virtual predictions are potentially useful for Belle II, but the abstract and conclusions overstate the degree of demonstrated model independence. The referee report focuses on the missing quantification of R across the five potentials and on the shared nonrelativistic Terentev ansatz that largely enforces the Qbar^2 scaling. I do not see a fatal internal inconsistency, so rejection is not warranted; a revision that shows the cross-potential spread and carefully qualifies the model-independence claim would bring the paper in line with its claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this paper derives the double-virtual gamma*gamma* -> eta_c(1S,2S) transition form factor in a light-front quark model, which is a real extension of the single-virtual LFQM results in Refs. [17,18]. It also introduces a factorization-breaking ratio R and argues that factorization is strongly broken while the form factor scales predominantly with Qbar^2 and is nearly independent of the asymmetry omega. That is a concrete, forward-looking prediction Belle II could test.\n\nThe derivation itself is solid. I traced Eqs. (2.38) through (2.40); the azimuthal integration is consistent, and the q2=0 limit reduces to the single-virtual result. The normalization at Q^2=0 reproduces the known NRQCD width formula, which is a good check. The paper uses five published potential-model wave functions and fits nothing to the BaBar TFF data, so the circularity burden is genuinely low, and the single-virtual comparison to BaBar is honest about model spread. Credit where due: this is a useful new observable and a cleanly presented formalism.\n\nThe soft spots are real but not fatal. First and most important: the \"almost model independent\" claim is asserted, not shown. The text says the R pattern \"looks very similar for different potentials (not shown explicitly here)\", but only the Buchmueller-Tye potential is plotted. That matters because all five wave functions are sharply peaked at z = 1/2 by construction: the Terentev mapping takes nonrelativistic Schrodinger wave functions and forces a narrow z-distribution. Once z is pinned near 1/2, Eq. (2.40) almost automatically makes the denominators depend on Q1^2 + Q2^2, so the Qbar^2 scaling and weak omega dependence are at least partly kinematically enforced. The five-potential scan does not address this; it only varies the radial function under the same narrow-z assumption. A referee should ask for R for all potentials, with the spread quantified, and ideally a test with a broader-z model or a relativistic correction.\n\nSecond, the remark that modifying mc to 1.3 GeV \"leads to a much better agreement\" with BaBar is ad hoc. It is not a fit, but it reads like post-hoc tuning. This should either be justified or dropped. Minor point: the discussion of the delayed BL asymptotics is fine, but the claim that evolution is irrelevant is a statement about these wave functions, not a general theorem.\n\nWho this is for: hadron phenomenologists working on charmonium form factors, double-tag measurements at Belle II, and hadronic light-by-light. The central derivation is worth refereeing seriously, even though the model-independence claim needs work. I would send it to peer review with a request to quantify the R spread and address the z-width sensitivity.","headline":"The double-virtual eta_c form factor is genuinely new and the derivation is clean, but the headline factorization-breaking measure is under-quantified across potentials and the omega-independence is at least partly built into the narrow nonrelativistic LFWF.","tokens_in":17490,"tokens_out":1985,"would_cite":true,"duration_ms":120355,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["12.38.Bx","13.85.Ni","14.40.Pq"],"model":"deepseek-v4-flash","headline":"The paper claims that the double-virtual $\\eta_c$ transition form factor does not factorize into single-virtual pieces; at large $Q^2$ it depends only on the average virtuality, with the breaking ratio reaching about 2.2.","keywords":["transition form factor","eta_c meson","light-front wave function","factorization breaking","charmonium potentials","double-tag measurement","photon fusion","Schrödinger equation"],"falsifier":"Measure the double-tag $e^+e^-\\to e^+e^-\\eta_c$ form factor at fixed average virtuality $(Q_1^2+Q_2^2)/2$ while varying the asymmetry $(Q_1^2-Q_2^2)/(Q_1^2+Q_2^2)$; the paper predicts almost no change, whereas a factorized vector-meson-dominance model predicts a visible dependence. Alternatively, recompute the same five potentials with a fully covariant light-front wave function instead of the Terentev mapping; if the ratio $R$ changes substantially, the model-independence claim would fail.","tokens_in":16316,"feed_emoji":"⚛️","tokens_out":16690,"duration_ms":566464,"temperature":0.7,"pith_summary":"This paper tries to establish that the double-virtual transition form factor $F(Q_1^2,Q_2^2)$ for $\\gamma^*\\gamma^*\\to\\eta_c(1S),\\eta_c(2S)$ is not the product of two single-virtual form factors, and that at large photon virtualities it depends mainly on the average virtuality $(Q_1^2+Q_2^2)/2$. Using light-front wave functions built from five different charm-anticharm potentials, the authors define a factorization-breaking ratio $R=\\tilde F(Q_1^2,Q_2^2)/[\\tilde F(Q_1^2,0)\\tilde F(0,Q_2^2)]$ and find that it grows to about 2.2 for the ground state while remaining almost the same across all five potentials. They also find that the onset of the asymptotic Brodsky-Lepage behavior is strongly delayed, with $Q^2F(Q^2)$ flattening below the asymptotic value for virtualities up to 50 GeV$^2$. This matters because double-tag measurements in high-luminosity $e^+e^-$ collisions can directly test factorization, and the same amplitude, up to color factors, enters $g^*g^*\\to\\eta_c$ production at hadron colliders.","feed_headline":"Double-virtual eta_c form factor breaks factorization by ~2.2","feed_subtitle":"At large virtualities only the average photon virtuality matters, so double-tag measurements can test the prediction directly.","key_machinery":"The load-bearing object is the light-front radial wave function $\\psi(z,k)$ obtained from the Terentev mapping. In practice, the rest-frame momentum $p$ of the quark is related to the light-front variables by $\\mathbf p^2=\\frac14(M_{c\\bar c}^2-4m_c^2)$ and $p_z=(z-\\frac12)M_{c\\bar c}$, with $\\psi(z,k)=\\frac{\\pi}{\\sqrt{2M_{c\\bar c}}}\\frac{u(p)}{p}$. This one mapping converts all five potential-model Schrödinger wave functions into the same light-front representation, and the hard-scattering integral (2.40) turns them into the form factor. The factorization-breaking measure $R$ defined in Eq. (3.3) quantifies how far the result is from the factorized form $\\tilde F(Q_1^2,0)\\tilde F(0,Q_2^2)$.","core_discovery":"The paper claims that for two spacelike photons, the transition form factor $F(Q_1^2,Q_2^2)$ of $\\eta_c(1S)$ and $\\eta_c(2S)$ is strongly factorization-breaking, with the ratio $R=\\tilde F(Q_1^2,Q_2^2)/[\\tilde F(Q_1^2,0)\\tilde F(0,Q_2^2)]$ rising to about 2.2 for the ground state and being nearly identical for five potential-model wave functions. The explicit light-front representation is obtained by convoluting a hard $\\gamma^*\\gamma^*\\to c\\bar c$ amplitude with a wave function $\\psi(z,k)$,\n\\[\nF($Q_1^{2}$,$Q_2^{2}$)=$e_c^{2}$\\sqrt{N_c}\\,4m_c\\int\\frac{dz\\,dk}{z(1-z)8\\$pi^{2}$}\\psi(z,k)\\left\\{\\frac{1-z}{\\sqrt{($k^{2}$-$m_c^{2}$-z(1-z)$Q_1^{2}$-(1-z)^$2Q_2^{2}$)^2+$4k^{2}$($m_c^{2}$+z(1-z)$Q_1^{2}$)}}+\\frac{z}{\\sqrt{($k^{2}$-$m_c^{2}$-z(1-z)$Q_1^{2}$-$z^{2}$$Q_2^{2}$)^2+$4k^{2}$($m_c^{2}$+z(1-z)$Q_1^{2}$)}}\\right\\},\n\\]\nwhere $\\psi(z,k)$ comes from the Terentev mapping of the rest-frame Schrödinger wave function. The normalized form factor is almost independent of the asymmetry parameter $(Q_1^2-Q_2^2)/(Q_1^2+Q_2^2)$, so $F$ scales with $(Q_1^2+Q_2^2)/2$. The single-virtual limit reproduces the shape of existing data, while $Q^2F(Q^2)$ does not approach the Brodsky-Lepage value $\\frac{8}{3}f_{\\eta_c}$; the paper argues that QCD evolution of the distribution amplitude is negligible below 100 GeV$^2$.","pith_inferences":["If the near-flat dependence on asymmetry survives, double-tag data will cleanly distinguish light-front potential models from factorized parametrizations, because the latter predict a visible dependence on asymmetry.","The similarity of $R$ across five potentials hints that the breaking is driven mainly by the hard quark propagators and the narrow peak of the wave function near $z=1/2$, rather than by the long-distance potential; varying the width of the wave function in transverse momentum would test this.","Applying the same construction to the bottomonium analogue would separate kinematic effects from relativistic corrections: a similar $R$ would point to kinematics, while a smaller $R$ would expose the relativistic corrections already present in charmonium."],"forward_implications":["For $\\eta_c(1S)$, the factorization-breaking ratio $R$ rises to about 2.2 at large $Q_1^2,Q_2^2$, so a product of single-virtual form factors underestimates the double-virtual form factor by more than a factor of two.","The form factor is nearly determined by the average virtuality; changing the asymmetry at fixed average moves $F$ very little, in contrast to factorized vector-meson-dominance models.","The Brodsky-Lepage asymptotic plateau $Q^2F(Q^2)\\to \\frac{8}{3}f_{\\eta_c}$ is not reached below $Q^2\\approx 50$ GeV$^2$, and QCD evolution of the distribution amplitude changes the result negligibly below 100 GeV$^2$.","For $\\eta_c(2S)$, the factorization breaking is weaker, reaching about 1.8 in the explored virtuality range, and the approach to asymptotics is even slower.","Because the $\\gamma^*\\gamma^*\\to\\eta_c$ amplitude is the same as $g^*g^*\\to\\eta_c$ up to color factors, future double-tag measurements would also constrain $\\eta_c$ production at hadron colliders."],"supporting_citations":[{"why":"Supplies the Terentev prescription that maps rest-frame Schrödinger wave functions into light-front wave functions.","marker":"[26]"},{"why":"Provides the five Schrödinger-equation potential-model wave functions used throughout the paper.","marker":"[20]"},{"why":"Supplies the hard-scattering spinor formalism and the asymptotic Brodsky-Lepage limit used as the benchmark.","marker":"[38]"},{"why":"Gives the earlier single-virtual light-front quark-model result to which the present representation reduces.","marker":"[18]"},{"why":"Provides the single-virtual eta_c(1S) form-factor data used for comparison.","marker":"[21]"},{"why":"One of the five potentials whose near-identical results support the model-independence claim.","marker":"[31]"},{"why":"Another potential, used for the representative two-dimensional form-factor and factorization-breaking plots.","marker":"[37]"},{"why":"Quantifies the asymmetry dependence for the pion case, the comparison that makes the weak asymmetry dependence of eta_c noteworthy.","marker":"[44]"}],"fun_headline_variants":["Double-virtual eta_c breaks factorization by 2.2","Eta_c double-virtual form factor: factorization broken at 2.2","2.2 factorization breaking in double-tag eta_c form factor","Double-tag eta_c: factorization breaks with factor 2.2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Terentev mapping from the nonrelativistic rest-frame wave function to the light-front wave function is accurate enough for charmonium, whose internal quark momentum is only moderately nonrelativistic, so the predicted size and shape of the factorization breaking depend on that mapping.","fun_headline_variants_meta":{"raw":{"variants":["Double-virtual eta_c breaks factorization by 2.2","Eta_c double-virtual form factor: factorization broken at 2.2","2.2 factorization breaking in double-tag eta_c form factor","Double-tag eta_c: factorization breaks with factor 2.2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000796,"raw_usage":{"total_tokens":3641,"prompt_tokens":1218,"completion_tokens":2423,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":834,"completion_tokens_details":{"reasoning_tokens":2347}},"tokens_in":834,"tokens_out":2423,"duration_ms":111909,"temperature":1.0,"reasoning_tokens":2347,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:56:54.842296+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the double-tag $e^+e^-\\to e^+e^-\\eta_c$ form factor at fixed average virtuality $(Q_1^2+Q_2^2)/2$ while varying the asymmetry $(Q_1^2-Q_2^2)/(Q_1^2+Q_2^2)$; the paper predicts almost no change, whereas a factorized vector-meson-dominance model predicts a visible dependence. Alternatively, recompute the same five potentials with a fully covariant light-front wave function instead of the Terentev mapping; if the ratio $R$ changes substantially, the model-independence claim would fail.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the five Schrödinger-equation potential-model wave functions used throughout the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the earlier single-virtual light-front quark-model result to which the present representation reduces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the single-virtual eta_c(1S) form-factor data used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"One of the five potentials whose near-identical results support the model-independence claim."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Another potential, used for the representative two-dimensional form-factor and factorization-breaking plots."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Quantifies the asymmetry dependence for the pion case, the comparison that makes the weak asymmetry dependence of eta_c noteworthy."}],"review_version":1}