{"id":"4a062f27-51f5-4800-b787-9f4cec136d86","arxiv_id":"2502.03951","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A time-derivative moment method applied to NRQCD lattice correlators extracts bottomonium masses and thermal widths, finding increasing width and decreasing mass with temperature.","lead":"This paper extracts the mass and thermal width of bottomonium mesons from lattice QCD correlation functions using time-derivative moments, avoiding the full spectral reconstruction problem. The method yields zero-temperature masses near experiment and indicates that thermal width grows with temperature while the apparent mass increase is a kinematic effect.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The fit forms (17)-(18) drop terms that the Gaussian ansatz of Eq. (2) itself generates, so the extracted mass and width may be biased precisely in the high-temperature and chi_b1 regimes.","rationale":"I read the paper's central claim as a methodological proposal: time-derivative moments, combined with a Gaussian ansatz, can extract mass and thermal width without solving the inverse problem. For this claim to hold, the fit functions used on the lattice data must follow from the stated spectral ansatz under the stated approximations. They do not. The derivation from Eq. (15) to Eq. (17) silently drops the -Gamma0^2 tau term, and the derivation from Eq. (16) to Eq. (18) drops a tau-dependent prefactor that is not constant unless Delta Gamma^2 vanishes. These are not questions about whether the true spectral function is Gaussian; they are internal consistency failures of the method as presented. The reader's weakest_assumption correctly identifies the Gaussian ansatz and the narrow-width limit as fragile, and I partially agree. My concern is more specific: even accepting the Gaussian ansatz, the fitted M(tau) and Gamma^2(tau) are not the expressions that ansatz produces. The paper gives some credit for honesty: Section 5 explicitly flags the chi_b1 width issue, and Figure 1 shows the linear large-tau behavior that the omitted term would cause. The zero-temperature Upsilon agreement is real but weak support because the additive NRQCD energy shift E0 is calibrated to the Upsilon mass. I do not recommend rejection: a corrected derivation or a synthetic-data validation could rescue the method, and the physics results are broadly consistent with prior lattice studies. For that reason the reader's CONDITIONAL verdict remains appropriate, and my concern strengthens the condition that the method must be validated against known inputs before the extracted mass and width are trusted.","tokens_in":7021,"tokens_out":5810,"duration_ms":57771,"concrete_test":"Generate a synthetic correlator from the assumed Gaussian spectral shape, G(tau) = A0 e^{-m0 tau + Gamma0^2 tau^2/2} + A1 e^{-m1 tau + Gamma1^2 tau^2/2}, using the Gen2L tau grid, with m0 near the Upsilon mass, m1 near the first excited state, and Gamma0, Gamma1 values covering 0, 50, 100, and 200 MeV. Apply the paper's bootstrap fitting pipeline with Eq. (17) for the mass and Eq. (18) for the width, and compare the recovered m0 and Gamma0 with the inputs. If the recovered values deviate by more than the quoted statistical plus systematic errors, or if the result depends strongly on the chosen fit window, the method is biased and the central claim fails. The same test using the full Eqs. (15)-(16) should recover the inputs exactly, isolating the internal inconsistency from lattice artifacts.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 derives M_L and Gamma^2_L from a Gaussian spectral function. From Eq. (15), M_L(tau) contains a term -Gamma0^2 tau in addition to the excited-state exponential; Eq. (17) instead fits M(tau)=m0 + A e^{-Delta m tau}. The stated narrow-width condition Gamma^2 tau << m0 only makes this term small relative to m0, not relative to the fit precision or to the excited-state contribution. For Gamma ~ 100 MeV and tau ~ 30 GeV^{-1}, Gamma^2 tau ~ 300 MeV, far larger than the quoted uncertainties and comparable to the thermal shifts the paper reports. Similarly, Eq. (16) gives Gamma^2_L(tau)=Gamma0^2 + (A1/A0)[(-Delta m + Delta Gamma^2 tau)^2 + Delta Gamma^2] e^{-Delta m tau + Delta Gamma^2 tau^2/2}, while Eq. (18) drops the tau-dependent prefactor; this is only justified if Delta Gamma^2 = 0, which contradicts retaining Delta Gamma^2 in the exponent of Eq. (18). The paper's own Section 5 lists the Gamma^2 -> 0 limit for chi_b1 as an outstanding question, and Figure 1 shows linear large-tau behavior in M_L(tau) for chi_b1, which is precisely the signature of the neglected -Gamma0^2 tau term. Thus the central claim that the method gives unbiased mass and width while avoiding the inverse problem rests on an internal inconsistency, not merely on the physical validity of the Gaussian assumption.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a time-derivative-moment method for NRQCD bottomonium correlators: assuming the spectral function is a sum of Gaussians, the authors define logarithmic first and second derivatives of the correlator and fit them to extract the ground-state mass and thermal width of the Upsilon and chi_b1 for temperatures from 47 to 380 MeV. They report zero-temperature masses consistent with experiment after adding the NRQCD energy shift, a thermal broadening of both states, and a mass decrease once a zero-temperature control ('T0 analysis') is applied to separate short-time kinematic effects from genuine thermal effects. The central technical claim is that Eqs. (17)-(18) follow from the Gaussian ansatz under an N=1 narrow-width simplification.","tokens_in":7415,"tokens_out":9392,"duration_ms":85749,"significance":"The method is attractive because it avoids full spectral reconstruction and is computationally cheap, and the paper includes useful control checks: bootstrap errors, comparison of two derivative discretizations (M_L/M_D and Gamma_L/Gamma_D), and a T0 analysis that separates kinematic from thermal effects. If the derivation of the fit forms were consistent, the approach would be a valuable complementary tool to MEM and Bayesian reconstructions for quarkonium at nonzero temperature. However, the manuscript contains a load-bearing gap: the fit forms used for the reported masses and widths do not follow from the stated Gaussian ansatz. The concern is confirmed by comparing Eqs. (15)-(18). The paper does not provide machine-checked code or a full model-systematic error budget for the dropped terms, so the numerical results should be regarded as preliminary until the fit forms are justified or replaced.","major_comments":[{"comment":"The fit form (17) does not follow from the Gaussian ansatz. For N=1, Eq. (15) gives M_L(tau) = -m0 + Gamma0^2 tau + (A1/A0)(-Delta m + DeltaGamma^2 tau) exp(-Delta m tau + DeltaGamma^2 tau^2/2). Equation (17) instead fits M(tau) = m0 + A exp(-Delta m tau), dropping the Gamma0^2 tau term and the tau-dependent prefactor. The stated narrow-width condition Gamma^2 tau << m0 only makes Gamma0^2 tau small relative to m0 ~ 9.5 GeV; it does not make it small relative to the quoted mass uncertainties (about 10 MeV) or the thermal mass shifts reported in Fig. 2 (tens of MeV). For the chi_b1, with Gamma of order 150-400 MeV as shown in Fig. 4 (right), Gamma^2 tau is of order 100 MeV over the fitted window, which is comparable to the magnitude of the reported effects. The linear large-tau trend in the chi_b1 panel of Fig. 1 is the signature of the neglected term. The authors should fit the full expression (15) or quantify the bias from omitting it; otherwise the extracted masses are biased by an uncontrolled approximation.","section":"Sec. 3, Eqs. (15) and (17)"},{"comment":"The width fit is internally inconsistent. With N=1, Eq. (16) gives Gamma^2_L(tau) = Gamma0^2 + (A1/A0)[(-Delta m + DeltaGamma^2 tau)^2 + DeltaGamma^2] exp(-Delta m tau + DeltaGamma^2 tau^2/2). Equation (18) retains DeltaGamma^2 in the exponent but replaces the entire prefactor by a constant A. This is justified only if DeltaGamma^2 = 0, in which case the exponent should be -Delta m tau without the DeltaGamma^2 tau^2/2 term. If DeltaGamma^2 is retained because it is needed to represent the data, the tau-dependence of the prefactor is generically as important as the exponential factor over the fitted range. The Gamma(T) values in Fig. 4 are therefore obtained from a functional form that is not a consequence of the model stated in Eq. (2), and the reported thermal widths inherit an unquantified systematic error.","section":"Sec. 3, Eqs. (16) and (18)"}],"minor_comments":[{"comment":"The printed formula contains a malformed integral and a stray minus sign in the numerator; please correct the typesetting so that the definition of M_D is unambiguous.","section":"Sec. 3, Eq. (5)"},{"comment":"The sentence describing the bootstrap and systematic errors is duplicated with slightly different wording; please consolidate it and specify which systematic estimate is shown in each figure.","section":"Sec. 4"},{"comment":"The legend omits the T=47 MeV entry that appears for the Upsilon; please state whether chi_b1 data at this temperature are unavailable and why.","section":"Fig. 3 (right)"},{"comment":"The text says systematic errors from Gamma_D are estimated, but only Gamma_L and Gamma_L(T0) are plotted for chi_b1; please clarify whether Gamma_D was omitted because it is too noisy, and show it or state the reason.","section":"Fig. 4 (right)"},{"comment":"The zero-temperature Upsilon agreement with experiment is not an independent validation because the additive shift E0 is fixed by the Upsilon(1S) mass; please state this explicitly and present the chi_b1 mass as the genuine prediction.","section":"Sec. 4"},{"comment":"For reproducibility, please report the fit ranges, the number of points, and the fit quality (chi^2 or comparable) for the fits to Eqs. (17) and (18) at each temperature.","section":"Sec. 4"}],"recommendation":"major_revision","confidential_remarks":"This is a proceedings-style contribution from an established collaboration, and the underlying ensemble and analysis infrastructure are solid. The derivation gap in Eqs. (17)-(18) is the main obstacle; it is fixable within the scope of the paper by either fitting the full expressions or providing a quantitative demonstration that the omitted terms are negligible. I do not recommend rejection, but the numerical values should not be quoted as final results until that step is completed. The paper's own conclusion already acknowledges the unresolved narrow-width question for chi_b1."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: the time-derivative-moment idea is worth a look for lattice quarkonium, and the T0 truncation-control analysis is a nice addition. But the paper as written has a load-bearing gap: the fit forms (17) and (18) do not follow from the Gaussian ansatz. Eq (15) contains a -Γ0^2 τ term and Eq (16) contains a τ-dependent prefactor that are simply dropped. The narrow-width condition Γ^2 τ << m0 makes the dropped term small relative to m0, not relative to the fit precision or the excited-state contribution. For the high-temperature Upsilon, Γ ~ 200 MeV and τ out to ~10 GeV^{-1} give Γ^2 τ ~ 0.4 GeV, far larger than the quoted ~10 MeV errors and comparable to the mass shift the paper reports. The chi_b1 plots show exactly the linear large-τ behavior that the neglected term would produce, and the authors themselves flag the chi_b1 width as an open question. The same issue affects the Upsilon at high T, so the central claim that the method avoids the inverse problem and gives unbiased masses/widths is not yet supported.\n\nWhat is genuinely new: applying these standard cumulant relations to NRQCD correlators with a Gaussian spectral-function ansatz, and the careful use of a zero-temperature truncated analysis to separate kinematic from thermal effects. The zero-T Upsilon mass is calibrated via the NRQCD additive shift, so the agreement with experiment is a consistency check, not a prediction; the chi_b1 mass is a more independent test. The data handling (bootstrap errors, ML vs MD as systematics) is standard and transparent.\n\nThe paper is an honest proceedings contribution. It is short, the figures are informative, and the limitations section acknowledges several open points. But the internal inconsistency between the derived equations and the fitted functions is central, not cosmetic. A referee should ask for a corrected derivation: either fit the full expressions, or restrict the fit range so the dropped terms are demonstrably small, or show that the bias is negligible numerically. Without that, the quoted temperature-dependent shifts are unreliable.\n\nI would send this to a referee for the proceedings, because the method is interesting and the data are real, but I would not cite the results in my own work until the fit forms are fixed. It is a candidate for a reading-group discussion on how easy it is to over-simplify moment-based extraction.","headline":"A promising moment-based method for NRQCD quarkonia, but the fit forms drop terms that are not negligible at high T, so the reported mass and width shifts are not yet reliable.","tokens_in":7953,"tokens_out":4748,"would_cite":false,"duration_ms":45055,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Time-derivative moments of the lattice correlator, taken under a Gaussian spectral-function assumption, give the Upsilon and chi_b1 masses and thermal widths without reconstructing the full spectral function.","keywords":["bottomonium","NRQCD","lattice QCD","time-derivative moments","thermal width","spectral function","quark-gluon plasma","heavy quarkonia"],"falsifier":"Take the same lattice ensembles and run the moment method on synthetic correlators generated from a known non-Gaussian spectral function, for instance a skewed or two-peak shape, and from a $\\chi_{b1}$-like large width; if the recovered mass and width deviate from the inputs beyond the quoted errors, the Gaussian Ansatz is biased. A second check is to compare the $\\chi_{b1}$ width extracted here with the maximum-likelihood spectral widths obtained from the same correlators in refs. [8] and [16]; agreement would support the Ansatz, while disagreement would settle against it.","tokens_in":6829,"feed_emoji":"🌡️","tokens_out":15110,"duration_ms":115700,"temperature":0.7,"pith_summary":"This paper claims a new method for extracting the mass and thermal width of heavy quarkonium states from lattice QCD correlators without solving the ill-posed inverse problem of spectral reconstruction. The method assumes the spectral function is a sum of Gaussians, then reads the first and second time-derivative moments of the NRQCD correlator, which give the mass and width directly. On the Upsilon and the chi_b1 it reproduces the experimental zero-temperature masses, shows the thermal width rising with temperature, and — after a zero-temperature temporal-truncation control removes a kinematic effect — finds the mass decreasing with temperature. The result matters because quarkonium melting is a key diagnostic of the quark-gluon plasma produced in heavy-ion collisions.","feed_headline":"Give Upsilon and chi_b1 mass plus thermal width from lattice moments","feed_subtitle":"A derivative-moment method skips spectral reconstruction and tracks how the states melt with temperature.","key_machinery":"The central objects are the time-derivative moments of the Euclidean correlator: the effective mass $M_L(\\tau)=\\partial\\log G(\\tau)/\\partial\\tau$ and the squared thermal width $\\Gamma^2_L(\\tau)=\\partial^2\\log G(\\tau)/\\partial\\tau^2$, together with their discrete-derivative partners $M_D$ and $\\Gamma^2_D$ used to estimate systematic errors. They do the work because a Gaussian spectral function centred at $m_i$ with width $\\Gamma_i$ maps the correlator to $\\exp(-m_i \\tau+\\Gamma_i^2\\tau^2/2)$; the first log-derivative then reads off the mass and the curvature of $\\log G$ reads off the width, sidestepping the inverse problem entirely. The subsequent fit keeps one excited state, assumes the ground state is well separated, and uses the narrow-width condition $\\Gamma^2\\tau\\ll m_0$ to extract $m_0$ and $\\Gamma_0$ at each temperature.","core_discovery":"On FASTSUM's Generation 2L anisotropic lattice ensembles, the authors replace the ill-posed reconstruction of the thermal spectral function $\\rho(\\omega;T)$ by assuming it is a finite sum of Gaussians centred at each state's mass with width set by its thermal broadening. Under that Ansatz the NRQCD correlator becomes a sum of exponentials $\\exp(-m_i \\tau + \\Gamma_i^2 \\tau^2/2)$, so the log-derivatives of the correlator give the ground-state mass $M_L(\\tau)=\\partial\\log G/\\partial\\tau$ and squared width $\\Gamma^2_L(\\tau)=\\partial^2\\log G/\\partial\\tau^2$ without any full inversion. Fitting these moments with one excited state and the narrow-width approximation yields $M(\\Upsilon)=9455(10)$ MeV and $M(\\chi_{b1})=9965(79)$ MeV at zero temperature, consistent with the Particle Data Group values, and a thermal width that grows with temperature for both states. An apparent rise in the $\\Upsilon$ mass with temperature is shown by the truncated zero-temperature ($T_0$) control to be a short-temporal-range kinematic effect; with that control the mass decreases with temperature.","pith_inferences":["The same derivative-moment machinery should transfer to charmonium or other heavy-quark bound states, where the Gaussian Ansatz's bias could be quantified by comparing with full spectral reconstructions on identical ensembles.","A synthetic-data test, generating correlators from known non-Gaussian spectral shapes and running the method on them, would map exactly where the Gaussian-plus-narrow-width assumption distorts the extracted mass and width.","The $T_0$ truncation control is a transferable subtraction scheme: any lattice spectral study whose high-temperature correlators suffer from short temporal extent could apply the same zero-temperature truncation to isolate genuine thermal effects."],"forward_implications":["The method extracts ground-state masses and thermal widths for $\\Upsilon$ and $\\chi_{b1}$ using only derivatives of the correlator, avoiding the ill-posed inverse problem that dominates spectral-function studies.","Zero-temperature masses agree with experiment, so the Gaussian-moment approach reproduces established bottomonium spectroscopy from the same ensembles.","The thermal width $\\Gamma$ increases with temperature for both states, consistent with quarkonium suppression in the quark-gluon plasma.","The apparent upward shift of the $\\Upsilon$ mass with temperature is a kinematic truncation effect, not a thermal one; after the $T_0$ control the mass falls with temperature.","Relaxing the single-excited-state and narrow-width assumptions, by fitting Eqs. (15) and (16) with $N>1$, is a direct route to reduce systematic uncertainties in future applications."],"supporting_citations":[{"why":"It supplies the Gaussian approximation of the spectral function used in Eq. (2).","marker":"[8]"},{"why":"It supports the Gaussian broadening Ansatz for thermal bottomonia that underlies the moment formulae.","marker":"[9]"},{"why":"It provides the standard effective-mass construction that the M_L moment is compared against.","marker":"[11]"},{"why":"It gives the experimental Upsilon and chi_b1 masses used to benchmark the zero-temperature results.","marker":"[13]"},{"why":"It is the previous FASTSUM result of a rising Upsilon mass that this paper reinterprets as a kinematic effect.","marker":"[14]"},{"why":"It explains the apparent mass increase as a short-temporal-range kinematic effect at high temperature.","marker":"[15]"},{"why":"It is the earlier lattice NRQCD study finding an in-medium mass drop, consistent with the T0-controlled result here.","marker":"[16]"}],"fun_headline_variants":["Avoid spectral inversion: derivative moments give Upsilon mass and width","Time-derivative moments extract bottomonium mass and thermal width","No spectral reconstruction: lattice moments track bottomonium melting","Derivative moments give Upsilon and chi_b1 mass, width at finite T","Mass and thermal width from correlator derivatives: bottomonium at finite T"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the true spectral function is a finite sum of well-separated Gaussians and that the ground-state width obeys the narrow-width condition $\\Gamma^2\\tau\\ll m_0$; if the real spectral shape is non-Gaussian or the $\\chi_{b1}$ width is too large for that limit, the extracted mass and width are biased, and the authors single out the $\\chi_{b1}$ case as an outstanding question.","fun_headline_variants_meta":{"raw":{"variants":["Avoid spectral inversion: derivative moments give Upsilon mass and width","Time-derivative moments extract bottomonium mass and thermal width","No spectral reconstruction: lattice moments track bottomonium melting","Derivative moments give Upsilon and chi_b1 mass, width at finite T","Mass and thermal width from correlator derivatives: bottomonium at finite T"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00071,"raw_usage":{"total_tokens":3167,"prompt_tokens":888,"completion_tokens":2279,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":504,"completion_tokens_details":{"reasoning_tokens":2190}},"tokens_in":504,"tokens_out":2279,"duration_ms":14791,"temperature":1.0,"reasoning_tokens":2190,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T00:08:05.603635+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same lattice ensembles and run the moment method on synthetic correlators generated from a known non-Gaussian spectral function, for instance a skewed or two-peak shape, and from a $\\chi_{b1}$-like large width; if the recovered mass and width deviate from the inputs beyond the quoted errors, the Gaussian Ansatz is biased. A second check is to compare the $\\chi_{b1}$ width extracted here with the maximum-likelihood spectral widths obtained from the same correlators in refs. [8] and [16]; agreement would support the Ansatz, while disagreement would settle against it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It gives the experimental Upsilon and chi_b1 masses used to benchmark the zero-temperature results."},{"cited_title":"Bayesian study of relativistic open and hidden charm in anisotropic lattice QCD","cited_arxiv_id":"1802.00667","evidence_quote":"It explains the apparent mass increase as a short-temporal-range kinematic effect at high temperature."}],"review_version":1}