{"id":"99a0adb9-ecc4-44a8-92df-6a090a772c17","arxiv_id":"2504.17431","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":28,"one_line_summary":"A coupled-channel fit to e+e- to D meson pairs finds that G(3900) is a dynamically generated P-wave D Dbar* resonance.","lead":"This paper analyzes electron-positron collision data to understand a broad enhancement called G(3900), a particle-like bump near 3.9 GeV. The authors conclude it is not an ordinary charmonium state but a loosely bound composite of two charmed mesons, and they predict additional exotic states that future experiments could detect.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Regulator dependence of the cutoff scheme is the weakest link: the paper concedes that omitting counterterms makes the extracted G(3900) pole cutoff-dependent, so the qualitative dynamical-generation claim is not yet fully secured.","rationale":"The paper makes a credible effort: it fits eight measured cross sections in two model variants, finds a G(3900) candidate pole in both, and uses pole trajectories under g -> 0 to separate dynamical poles from renormalized charmonia. Agreement with Refs. [29,32] on the qualitative molecular assignment is independent support. My concern is not that the calculation is internally inconsistent; it is that the single quantity on which the claim pivots, the near-3.9 GeV pole, is extracted in a scheme whose cutoff dependence the authors explicitly acknowledge. Changing Lambda, altering the form-factor shape, or adding a single counterterm could in principle move the pole, change its Riemann sheet, or change the dynamically-generated classification, and the paper provides no scan or estimate for any of those effects. Since the central claim is qualitative and the two models already disagree by about 50 MeV in the pole position, the conclusion should remain conditional until this regulator sensitivity is quantified. The reader's verdict already captures exactly this weakness, so no change to the verdict is needed.","tokens_in":37692,"tokens_out":4564,"duration_ms":51460,"concrete_test":"Re-run the Model II fit, and as a cross-check Model I, with Lambda = 0.40, 0.50, and 0.60 GeV, and also with a dipole form factor (1 + q^2/Lambda^2)^-2 at Lambda = 0.50 GeV, tracking the candidate G(3900) pole on all nearby Riemann sheets and repeating the g -> 0 trajectory classification. If the near-3.9 GeV pole persists on a consistent sheet with its real part stable to within about 30 MeV and its width to within about 30 MeV, the dynamical-generation identification is regulator-robust; if the pole shifts by more than about 50 MeV, changes sheets, or disappears under any of these choices, the present regulator scheme does not yet support the claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that G(3900) is a dynamically generated P-wave D\\bar D^* state—rests on a pole extracted from one cutoff-regulated separable contact interaction, with the Gaussian cutoff fixed at Lambda = 0.50 GeV and no higher-order counterterms. The paper states this limitation explicitly in Section III: 'The description of the data thus relies on the choice of the cutoff as a consequence of omitting these necessary counter terms.' That is exactly the quantity on which the conclusion pivots. The two models already give materially different candidates for the same physical state: 3832.52 +/- 74.53i MeV on the (-, +, -, +, +, +) sheet in Model I versus 3883.91 +/- 46.53i MeV on the (-, -, +, +, +, +) sheet in Model II, and the decision to emphasize Model II is a goodness-of-fit judgment rather than a regulator-independent extraction. The pole-trajectory analysis in Fig. 5 is a sensible internal diagnostic for separating dynamically generated poles from renormalized bare charmonia, but it only varies the charmonium couplings; it does not test sensitivity to Lambda, to the form-factor shape, or to the omitted counterterms. Because P-wave contact interactions in this class of LSE are UV-sensitive, the existence, Riemann-sheet assignment, mass, and width of the 'dynamically generated' pole are not shown to be invariant under changes of the regulator. The qualitative conclusion may survive, but the load-bearing evidence for it is precisely the pole whose regulator dependence is conceded.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper performs a global fit to the BESIII, Belle, and BaBar cross-section data for e+e- -> D Dbar, D Dbar* + c.c., and D* Dbar* in the energy region 3.7-4.25 GeV, using a coupled-channel Lippmann-Schwinger equation with P-wave contact interactions built from the heavy-quark spin symmetry doublet (D, D*) and its antiparticle in an SU(3) flavor extension of earlier work. Three bare charmonia (psi(1D), psi(3S), psi(2D)) are included in Model I and two in Model II. The main claim is that a pole near 3.9 GeV, at 3832.57(+0.91,-0.79) - 74.53(+0.68,-2.15)i MeV in Model I and 3883.91(+0.38,-0.46) - 46.53(+1.22,-1.22)i MeV in Model II, is dynamically generated rather than a renormalized bare charmonium, based on pole-trajectory analysis that switches off the bare couplings. The same fitted parameters are used to predict dynamically generated states in the JPC = 1-+ channel, which are proposed as search targets in single-photon e+e- processes.","tokens_in":38451,"tokens_out":3817,"duration_ms":41610,"significance":"If the central claim is correct, the paper provides a quantitative identification of G(3900) as a P-wave D Dbar*/Dbar D* hadronic molecule rather than a conventional charmonium state, with the BESIII lineshape explained as a threshold-coupled-channel effect. The work is a serious contribution: it uses up-to-date experimental cross sections, implements a consistent SU(3) generalization of the HQSS contact formalism, solves the full coupled-channel LSE with an explicit production amplitude, and provides a non-circular pole-trajectory diagnostic for distinguishing dynamically generated poles from renormalized bare states. The predictions for JPC = 1-+ exotic states are falsifiable and give the paper additional value beyond the G(3900) interpretation. The main limitation, acknowledged in the text, is that the pole extraction is cutoff-dependent because higher-order counterterms are omitted; this is the key issue that determines whether the dynamical-generation classification is robust.","major_comments":[{"comment":"The paper explicitly states that 'The description of the data thus relies on the choice of the cutoff as a consequence of omitting these necessary counter terms.' Since the central conclusion that G(3900) is dynamically generated rests entirely on the pole obtained with Lambda = 0.50 GeV and a Gaussian separable regulator, the manuscript should demonstrate that the existence, Riemann-sheet assignment, mass, and width of this pole are stable under reasonable variations of Lambda and of the form-factor shape, or under inclusion of the minimal higher-order counterterms. Without such a study, the pole trajectory analysis in Fig. 5 varies only the bare charmonium couplings and does not address the regulator sensitivity that is the load-bearing uncertainty of the claim.","section":"Section III, discussion after Eq. (66) and the renormalization paragraph near the end of Section III A"},{"comment":"The two models give materially different candidates for the same physical state: 3832.52 +/- 74.53i MeV on the (-, +, -, +, +, +) sheet in Model I versus 3883.91 +/- 46.53i MeV on the (-, -, +, +, +, +) sheet in Model II. The difference in the real part is about 50 MeV and the widths differ by roughly 28 MeV, both much larger than the quoted statistical uncertainties. The choice to emphasize Model II is justified by fit quality and residual behavior, but the manuscript does not provide a quantitative estimate of the model systematic uncertainty or a criterion that would decide when the difference between models undermines the extraction. This should be addressed before the pole parameters are presented as the definitive G(3900) resonance parameters.","section":"Tables II and III and the accompanying discussion"},{"comment":"The reduced chi-square values are 2.17 (Model I) and 2.66 (Model II), and the standardized residuals in panels (c) and (d) of Fig. 3 are not randomly distributed around zero, as the authors themselves note. This moderate fit quality means that the pole parameters, including the dynamically generated G(3900) candidate, are not tightly constrained by the data alone. The paper should quantify the impact of the residual non-randomness, for example by adding a systematic uncertainty associated with the inconsistent BESIII and Belle data sets or by fitting with and without the problematic channels, to show that the dynamical-generation conclusion survives such changes.","section":"Section III, chi2/d.o.f. values in Table I and residual plots in Fig. 3"},{"comment":"The distinction between dynamically generated and renormalized bare charmonium poles is made by switching off g0_2D, g0_1D, and g0_3S and observing which poles move toward the bare masses. This is a valid self-consistency test, but it tests only one combination of parameters. The same classification should be checked against variations of the contact LECs and of the regulator, since a dynamically generated pole in a P-wave contact theory can move across Riemann sheets or vanish when the short-distance physics is changed. A short appendix with such sensitivity scans would substantially strengthen the central claim.","section":"Section III A, pole trajectory method and Fig. 5"}],"minor_comments":[{"comment":"There is a typo in 'Moldel I or II' in the paragraph following Eq. (26); it should read 'Model I or II'.","section":"Section II B"},{"comment":"The phrase 'one we get' should be 'one gets' or 'one obtains the physical production amplitude'.","section":"Section II C, sentence after Eq. (37)"},{"comment":"The text writes 'Breit-Winger' where 'Breit-Wigner' is intended.","section":"Section III A"},{"comment":"The fit parameters are reported with statistical uncertainties but no correlation matrix is given; given the large number of correlated LECs and bare parameters, a correlation matrix or at least a statement about the strongest correlations would help assess the stability of the pole extraction.","section":"Tables VII and VIII"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope and the formalism is competently constructed, with the strongest weak point being the regulator dependence of the dynamically generated pole, which the authors explicitly concede. I do not see circularity in the pole-trajectory classification, and the paper's predictions for the 1-+ channel are valuable. My recommendation of major revision is driven by the requirement to make the central claim robust against regulator and model choices; this is a fixable deficiency rather than a fatal error."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a competent coupled-channel analysis that gives the best-supported case so far for G(3900) as a dynamically generated P-wave D Dbar* molecule, and it adds a useful diagnostic for telling dynamical poles from renormalized bare states. The regulator dependence is real and the paper admits it; I do not think that kills the result, but it does mean the quantitative pole parameters should not be quoted without a caveat.\n\nWhat is new: previous work already identified G(3900) as a P-wave molecule. This paper extends the HQSS contact scheme to SU(3) with strange channels, fits e+e- cross sections in eight channels, and uses pole trajectories under variation of the bare charmonium couplings to distinguish poles that run to bare masses from poles that stay put. That diagnostic is a genuinely useful idea for this class of problems. The two models (three vs two bare charmonia) both produce a pole near 3.9 GeV with a large width, and the conclusion is stable across them. The 1-+ predictions follow from the same LECs and give concrete search channels.\n\nThe soft spot, stated by the authors themselves, is that the whole extraction sits in one cutoff-regularized separable contact interaction with Lambda = 0.50 GeV and no higher-order counterterms. The paper says the description of data relies on the cutoff choice. The pole-trajectory analysis does not scan Lambda, form-factor shape, or counterterm structure; it only varies the charmonium couplings. Models I and II give poles differing by about 50 MeV in real part and 28 MeV in imaginary part, and the choice to emphasize Model II is based on fit quality rather than regulator independence. So the quantitative resonance parameters are not yet on solid ground. The chi2/dof (2.17 and 2.66) is moderate, with visible residual structure in some channels, and no code or data tables are provided, so the fit cannot be independently reproduced. None of these flaws overturns the qualitative conclusion, which agrees with external analyses, but they should be quantified before the numbers are treated as definitive.\n\nFor a reader in hadron spectroscopy, this is worth engaging with. The formalism is laid out carefully, the comparison to previous work is fair, and the pole-trajectory method is likely to be reused. I would send it to peer review; the referee should ask for a regulator-sensitivity study, a second form-factor shape, and fit outputs beyond the parameter table. With that, the paper would be a solid reference for the molecular interpretation.","headline":"Solid coupled-channel case for G(3900) as a dynamically generated P-wave molecule, with a useful pole-trajectory diagnostic; the quantitative pole is regulator-dependent and needs a sensitivity study before it is quoted.","tokens_in":38994,"tokens_out":2222,"would_cite":true,"duration_ms":23517,"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":"This paper claims that the broad G(3900) bump seen in $e^+e^-\\to D\\bar D$ is a P-wave dynamically generated state of the $(D,D^*)$ meson doublet, not an ordinary charmonium, and supports that claim with a global coupled-channel fit.","keywords":["G(3900)","dynamically generated state","P-wave hadronic molecule","heavy quark spin symmetry","Lippmann-Schwinger equation","charmoniumlike states","electron-positron annihilation","open charm meson pairs"],"falsifier":"Repeat the coupled-channel fit with a different regulator, for example varying $\\Lambda$ from 0.4 to 0.6 GeV or using a sharp cutoff, and check whether a pole near 3830 to 3890 MeV persists on the relevant unphysical Riemann sheet; if it disappears or moves to one of the bare charmonium masses, the claim that G(3900) is dynamically generated fails.","tokens_in":1938,"feed_emoji":"⚛️","tokens_out":1792,"duration_ms":84585,"temperature":0.7,"pith_summary":"The paper analyzes the most precise electron-positron annihilation cross sections into open-charm meson pairs in the energy region 3.7 to 4.25 GeV and asks what the bump called G(3900) actually is. Its answer is that the bump is a P-wave state generated dynamically by the rescattering of $D$ and $D^*$ mesons, rather than a conventional quark-antiquark charmonium state or a mere threshold artifact. The evidence comes from a global fit whose scattering amplitudes, obtained by solving the Lippmann-Schwinger equation, contain a pole near 3.9 GeV in two different model settings, one with three and one with two bare charmonia. If the claim is right, G(3900) joins the family of hadronic molecules formed from charmed meson pairs, and the same dynamics predicts exotic $1^{-+}$ partners that could be searched for in radiative electron-positron processes.","feed_headline":"G(3900) is a P-wave D-D* molecule, not a plain charmonium","feed_subtitle":"A global fit to e+e- open-charm data finds a dynamically generated pole that reproduces the 3.9 GeV bump.","key_machinery":"The engine of the argument is a separable $P$-wave contact interaction between the $(D,D^*)$ doublet and its antiparticle, built in the heavy quark spin symmetry limit and decomposed into SU(3) flavor singlet, octet, and isospin-triplet channels. This interaction feeds a Lippmann-Schwinger equation with twelve open-charmed channels and either two or three bare vector charmonia, regulated by a Gaussian form factor with cutoff $\\Lambda=0.50$ GeV. The decisive diagnostic is the pole-trajectory plot: each pole of the $T$-matrix is followed as the bare charmonium couplings are gradually reduced to zero, so that poles that remain stationary are identified as dynamically generated while poles that migrate to the bare mass are identified as renormalized charmonia.","core_discovery":"The central claim is that G(3900) is a dynamically generated P-wave state of the $(D,D^*)$ heavy-quark-spin doublet and its antiparticle. When the scattering amplitudes for $e^+e^-\\to D\\bar D$, $e^+e^-\\to D\\bar D^*+\\text{c.c.}$, and $e^+e^-\\to D^*\\bar D^*$ are fitted globally, a pole near 3.9 GeV appears regardless of whether three bare charmonia ($\\psi(1D)$, $\\psi(3S)$, $\\psi(2D)$) or only two are included. Pole-trajectory analysis shows that this pole stays essentially fixed as the charmonium couplings are switched off, which is the paper's criterion for a dynamically generated state, whereas the ordinary charmonium poles move back to their bare masses. The paper reports a candidate pole at roughly $3832.6-74.5i$ MeV in Model I and $3883.9-46.5i$ MeV in Model II, and it argues that the Model II result is the more reliable one because Model I shows signs of overfitting.","pith_inferences":["If the mechanism is correct, the same $P$-wave contact dynamics should generate partner molecules in the bottom sector, such as $B\\bar B^*$ and $B^*\\bar B^*$ states near their thresholds, offering a test at a different heavy-quark mass.","The pole-trajectory diagnostic could be applied to other vector charmoniumlike states to decide, case by case, which are renormalized quarkonia and which are dynamically generated hadronic molecules.","A decisive next step would be to repeat the fit with a different regulator or with the omitted higher-order counterterms included; if a pole near 3.9 GeV survives such changes, the molecule interpretation would cease to be regulator-dependent."],"forward_implications":["The broad structure near 3.9 GeV should be classified as a $P$-wave $D\\bar D^*/\\bar D D^*$ hadronic molecule rather than a missing charmonium state, so searches for a conventional charmonium assignment there are unnecessary.","Two bare charmonia, $\\psi(1D)$ and $\\psi(3S)$, are sufficient to describe the data in the energy region considered; the additional $\\psi(2D)$ in Model I is redundant and its fitted pole has no physical significance.","Within this framework the $\\psi(4040)$ pole is also dynamically generated, with its bare input shifted to a higher pole around 4.23 to 4.28 GeV, which changes how its width should be interpreted near threshold.","The same low-energy constants predict several dynamically generated $1^{-+}$ exotic states, which should be looked for in electron-positron annihilation accompanied by a single photon.","For near-threshold states such as G(3900) and $\\psi(4040)$, Breit-Wigner masses and widths are not reliable resonance parameters; pole positions on the relevant Riemann sheets are the more appropriate quantities."],"supporting_citations":[{"why":"Supplies the near-threshold $e^+e^-\\to D\\bar D$ cross-section data that enter the global fit.","marker":"[1]"},{"why":"Supplies the precise $e^+e^-\\to D\\bar D$ Born cross-section data whose peak near 3.9 GeV defines the G(3900) input.","marker":"[4]"},{"why":"Supplies the $e^+e^-\\to D^{*+}D^{*-}$ and $e^+e^-\\to D^{*+}D^-$ cross-section data that constrain the $D\\bar D^*$ and $D^*\\bar D^*$ channels.","marker":"[5]"},{"why":"A K-matrix analysis that describes the same data without an additional bare pole; it serves as the benchmark against which the dynamically generated pole interpretation is compared.","marker":"[20]"},{"why":"An earlier identification of G(3900) as a $P$-wave $D\\bar D^*/\\bar D D^*$ resonance, which the present pole analysis supports and extends with explicit dynamical-generation evidence.","marker":"[29]"},{"why":"Provides the $P$-wave heavy-quark-spin-symmetry contact formalism and the heavy-light basis decomposition that this work extends to SU(3) flavor symmetry.","marker":"[31]"},{"why":"A global coupled-channel analysis whose G(3900) pole position is compared with the ones found in Models I and II.","marker":"[32]"},{"why":"Provides the regularized two-point loop function used in the Lippmann-Schwinger equation for the $P$-wave scattering amplitudes.","marker":"[38]"}],"fun_headline_variants":["G(3900) is a P-wave D-D* molecule, not plain charmonium","Global fit reveals G(3900) as dynamically generated P-wave state","G(3900) pole persists in e+e- data without bare charmonia","G(3900) emerges from D-D* interactions, not a simple charmonium","Study: G(3900) is a P-wave D-D* resonance, not a charmonium"],"cache_read_input_tokens":40576,"weakest_assumption_plain":"The classification leans on a single Gaussian cutoff of 0.50 GeV standing in for all omitted higher-order effects; the paper itself concedes that without the omitted counterterms the description of the data relies on that cutoff, so if the near-3.9 GeV pole vanishes or moves to a charmonium mass under a different regulator, the claim would lose its support.","fun_headline_variants_meta":{"raw":{"variants":["G(3900) is a P-wave D-D* molecule, not plain charmonium","Global fit reveals G(3900) as dynamically generated P-wave state","G(3900) pole persists in e+e- data without bare charmonia","G(3900) emerges from D-D* interactions, not a simple charmonium","Study: G(3900) is a P-wave D-D* resonance, not a charmonium"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000685,"raw_usage":{"total_tokens":3174,"prompt_tokens":1076,"completion_tokens":2098,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":692,"completion_tokens_details":{"reasoning_tokens":1983}},"tokens_in":692,"tokens_out":2098,"duration_ms":14699,"temperature":1.0,"reasoning_tokens":1983,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:40:54.404588+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the coupled-channel fit with a different regulator, for example varying $\\Lambda$ from 0.4 to 0.6 GeV or using a sharp cutoff, and check whether a pole near 3830 to 3890 MeV persists on the relevant unphysical Riemann sheet; if it disappears or moves to one of the bare charmonium masses, the claim that G(3900) is dynamically generated fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the near-threshold $e^+e^-\\to D\\bar D$ cross-section data that enter the global fit."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the precise $e^+e^-\\to D\\bar D$ Born cross-section data whose peak near 3.9 GeV defines the G(3900) input."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the $e^+e^-\\to D^{*+}D^{*-}$ and $e^+e^-\\to D^{*+}D^-$ cross-section data that constrain the $D\\bar D^*$ and $D^*\\bar D^*$ channels."}],"review_version":1}