{"id":"d776463d-812c-451c-aad5-490da12d1b39","arxiv_id":"1011.6661","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"An exact analytic expression is obtained for the leading-color two-loop QCD coefficient of gluon-fusion heavy-quark pair production, valid for all Mandelstam invariants and all heavy-quark masses.","lead":"This paper computes, for the first time with exact dependence on the top-quark mass and on the collision kinematics, the leading-color part of the two-loop QCD correction to gluon-fusion production of a heavy-quark pair. It is one of the last missing analytic ingredients needed for next-to-next-to-leading-order predictions of top-quark pair production at the LHC.","discovery_kind":"unclear","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The exact-mass finite part is validated only by two limits (m²/s→0 and IR poles); the interior, where the new result lives, rests on an unpublished reduction tool (Reduze 2, Ref. [37]) and internal symmetry checks.","rationale":"The paper is a solid, standard two-loop calculation and the reader's weakest_assumption is the right one. I would sharpen two points. First, the small-mass check against [23] is stronger than the reader implies: it is a functional match over t/s, not a single number, so it constrains the angular dependence of the massless limit well. Second, the threshold expansion (Eqs. 5.7–5.11) is derived from the exact result, so its finiteness is self-consistency, not independent evidence. What remains unvalidated is the finite-mass finite part at generic angle, and the reduction rests on an unpublished tool (Reduze 2, Ref. [37]), which the citation rules treat as a legitimate target. The shipped C++ code is a reproducibility aid but implements the same expression, so it does not independently validate. None of this warrants rejection: the scope is honestly stated (leading color, one channel), the MIs are published, and two independent published checks pass. It warrants keeping confidence at moderate and, ideally, an independent reduction cross-check. I would keep ACCEPT.","tokens_in":15244,"tokens_out":16238,"duration_ms":295525,"concrete_test":"Independently recompute one interior benchmark point from Table 1 — e.g. row 2 (s/m²=43, t/m²=-21, μ/m=1.7), whose ε⁰ coefficient is quoted as -707.01281 — using a different reduction code (FIRE or Reduze 1) and independent MI evaluation, without using the shipped C++ evaluator. Agreement to the quoted precision would settle the concern; a discrepancy would localize an error in the reduction or in the s<0→physical continuation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (abstract:s1) is an analytic expression valid for arbitrary s, t, m. The evidence is a standard method, published MIs [25,26,42–48], two limit checks, t↔u symmetry, threshold finiteness, and a shipped evaluator. The two published checks are limits: [23] is the m²/s→0 limit, which does probe the t/s (angular) dependence but only in the massless limit; [35] is the IR pole coefficients (1/ε⁴…1/ε), which are universal and do not constrain the finite part. Neither reaches finite m²/s at generic angle. The internal checks are necessary but not sufficient: a wrong rational coefficient in a weight-3 or weight-4 HPL that is symmetric under t↔u and finite at threshold would survive both. The reduction chain uses Reduze 2, cited as 'to be published' (Ref. [37]), so the reduction is not independently verified by any artifact in the paper. A sign or boundary-term error in the analytic continuation from s<0 to the physical region, or a wrong finite-part coefficient, would survive both limit checks. This is the load-bearing soft spot: the paper's own validation does not reach the region where the result is new. Note the small-mass check is stronger than a single number — it is a functional match over t/s — but it still stops at m²/s→0.","agreement_with_reader":"agree"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The manuscript presents the analytic calculation of the leading-color (N_c^3) coefficient A in the color decomposition of the two-loop interference term for gg → t tbar, Eq. (2.6), retaining the full dependence on s, t and the heavy-quark mass m. The reduction is performed with QGRAF and Reduze 2, the master integrals are taken from the authors' previous work and evaluated with the differential-equation method, and the result is expressed in terms of one- and two-dimensional harmonic polylogarithms of weight up to four. The UV renormalization is carried out in a mixed scheme (on-shell for the heavy-quark mass and wave functions, MS for α_s). The authors validate the result against the small-mass limit of Ref. [23] and the IR pole coefficients of Ref. [35], note the t ↔ u symmetry and threshold finiteness, and provide a C++ evaluator, a text file with the full expression, and expansions in the threshold and small-mass limits.","tokens_in":15598,"tokens_out":19819,"duration_ms":363203,"significance":"If correct, this is the first exact-mass analytic result for the leading-color coefficient of gg → t tbar at two loops, and it is a genuine ingredient toward a fully differential NNLO description of top-quark pair production. The paper is the third in a series and the novelty is incremental with respect to Refs. [25,26], but the result is new and useful. The shipped numerical code and the two independent limit checks are real strengths, and the threshold and small-mass expansions add practical value.","major_comments":[{"comment":"The central result of the paper, the full analytic expression for the coefficient A, is not contained in the manuscript; the authors state in §5 that it is 'too long to be explicitly presented in written form' and is supplied as a text file with the arXiv submission. Please confirm that the file is actually included, document its format and the normalization convention (Eq. (5.1)), and consider depositing it in a permanent repository so that the result can be checked independently of the arXiv source. The same applies to the C++ evaluator: a short usage note and a few benchmark outputs would make the code verifiable.","section":"§5, Eq. (5.1) and the arXiv text file","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The two external checks reported in §5 are limits: the small-mass expansion compared with Ref. [23] and the IR pole coefficients compared with Ref. [35]. Neither probes the finite part of A at finite m²/s and generic scattering angle, which is precisely the region where the new result lives; the t ↔ u symmetry and threshold finiteness are necessary but not sufficient. A sign or boundary-term error in the analytic continuation from s < 0 to the physical region, or a wrong finite-part coefficient in a weight-3 or weight-4 HPL, would survive both checks. Please add at least one independent check in the interior, for example a numerical evaluation of a subset of the master integrals at a finite-mass phase-space point by an independent method (sector decomposition or Mellin-Barnes), or a comparison of the discontinuity of A across the physical cut with the corresponding product of tree and one-loop amplitudes. Relatedly, the reduction is performed with Reduze 2, cited as 'to be published' (Ref. [37]); please clarify its status and whether the reduction can be reproduced with the public Reduze [40].","section":"§5, validation","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The term (11/6)² C_A² ε in the numerator appears to be a typographical error for (11/6)² C_A² / ε. As printed, the expression lacks the 1/ε² pole required by the standard MS two-loop coupling renormalization and contains a spurious finite piece; the correct form is (1/(4ε))[(11/6)² C_A² / ε − (17/12) C_A²], consistent with β0 = (11/6)C_A and β1 = (34/3)C_A². Please check.","section":"Eq. (4.13)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The ε⁰ term appears to carry mass dimension −2 while the ε and ε² terms are dimensionless; please confirm the normalization of A0 or supply the missing factor.","section":"Eq. (2.4)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that the IR poles 'find numerical agreement' with Ref. [35] should be quantified: please state the precision of the comparison and the kinematic points at which it was performed.","section":"§5, IR pole comparison","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The threshold-expansion coefficients are given numerically; please state the precision and, where possible, identify the transcendental constants that enter, so that the expansion can be reproduced independently.","section":"§5, Eqs. (5.7)–(5.11)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The phrase 'valid for any value of the Mandelstam invariants s and t' is stronger than what is demonstrated: the calculation is performed in the Euclidean region s < 0 and continued to the physical region. Please state this explicitly in the abstract or in §2.","section":"Abstract","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The counting 'additional 116 (out of 218) diagrams for the process initiated by ghosts' should be explained more clearly, including how the ghost contributions enter the interference with the tree-level amplitude.","section":"§3, ghost sector","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper should comment on the numerical stability of the HPL evaluation near the boundaries β → 0 and φ → 0, 1, where the arguments of the HPLs approach the branch points.","section":"§5, numerical stability","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The statement that A 'receives contributions from planar Feynman diagrams only' is correct but could be misread as identifying A with the full planar contribution; planar diagrams also contribute to the N_c B coefficient. Please clarify that A is the N_c³ coefficient specifically.","section":"§5, Eq. (2.6)","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a solid, standard two-loop calculation from an experienced group. The main issues are the availability of the central result (supplementary file) and the fact that the validation does not reach the interior region. I would not block publication on these, but I would like the authors to address them. The paper is the third in a series and the novelty is incremental but real; it is within the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"Quick read on this one: it's the gluon-fusion counterpart to the authors' earlier exact-mass calculation in the quark-antiquark channel, and it delivers what it promises — the leading-color coefficient A of Eq. (2.6) as an analytic expression in s, t, m, with no mass hierarchy assumed. That coefficient is new. The prior gluon-fusion two-loop result (Czakon–Mitov–Moch) was small-mass only, and the exact-mass qqbar result is a different channel. So this is a real ingredient for NNLO ttbar, not a reshuffling of existing work. The method is the standard chain — QGRAF, Reduze 2, Laporta reduction, differential equations, HPLs up to weight four — and the master integrals are the published ones from their earlier papers. Nothing exotic, nothing hand-wavy in the setup. The scope is stated honestly in the conclusions: leading color, planar diagrams only, with the non-planar and fermionic pieces and the elliptic sunrise subtopology explicitly deferred. The classifier's 'no limitations found' is wrong; the paper is quite explicit about all of this, and about the threshold-expansion constants being numerical because they couldn't find a closed form for the HPL values at x=1. That's an honest admission, not a hidden gap. The soft spot is the one the stress test names, and it's real but not fatal. Validation is two limits: the small-mass expansion against [23] and the IR pole coefficients against [35]. The first is a functional match in t/s, which is more than a single number, but it still sits at m²/s→0. The second is universal and says nothing about the finite part. Symmetry under t↔u and finiteness at threshold are necessary, not sufficient — a wrong weight-three or weight-four coefficient could survive both. The reduction also runs through Reduze 2, cited as 'to be published', so the reduction chain isn't independently reproduced by anything in the paper. The obvious missing check is a numerical comparison in the interior once someone has gluon-fusion numerics; until then, the shipped C++ evaluator is the only handle, and it's the authors' own code. Worth a serious referee, mainly to have someone re-derive a couple of the finite-part coefficients independently. I'd cite it.","headline":"A solid, standard two-loop calculation that delivers a genuinely new exact-mass color coefficient; the only real soft spot is that the interior of the kinematic region is checked only by limits.","tokens_in":16201,"tokens_out":1089,"would_cite":true,"duration_ms":18494,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["12.38.Bx","14.65.Ha","13.85.-t"],"model":"selfhosted/deepseek-v4.1-flash","headline":"The leading-color two-loop correction to gluon-fusion top-quark pair production is now known exactly in the quark mass, for all energies and scattering angles.","keywords":["heavy-quark pair production","gluon fusion","two-loop QCD","leading color coefficient","harmonic polylogarithms","Laporta algorithm","differential equations","top quark"],"falsifier":"An independent numerical evaluation of the two-loop $gg$ interference at one of the paper's own benchmark points, for instance $s/m^2 = 5$, $t/m^2 = -1.25$ with $mu = m$, compared against the quoted coefficients of the $epsilon$ expansion; a mismatch in the finite part would show that the analytic continuation or the reduction chain fails away from the two checked limits.","tokens_in":16671,"feed_emoji":"⚛️","tokens_out":2680,"duration_ms":56861,"temperature":0.7,"pith_summary":"The paper computes one specific piece of the two-loop QCD correction to $gg$ fusion into a heavy-quark pair: the coefficient of the leading color factor $N_c^3$, called $A$ in the authors' decomposition of the two-loop interference term. What makes the result distinctive is that it keeps the heavy-quark mass $m$ exact, so it is claimed valid for any values of the Mandelstam invariants $s$ and $t$, not only in the limit where $s, |t|, |u|$ dwarf $m^2$. This matters because a fully differential next-to-next-to-leading-order description of top-quark pair production needs the mass dependence everywhere, including near threshold and at moderate energies where the small-mass approximation is unreliable. The authors validate the expression by expanding it to the known small-mass result and by checking its infrared pole coefficients against an independent prediction, and they supply a C++ code that evaluates it numerically. The stated price is that the full analytic expression is too long to print and ships as a text file, and that the coefficients of the threshold expansion are given numerically because the authors found no satisfactory closed form for the transcendental constants that appear.","feed_headline":"Exact two-loop QCD result for gluon-fusion top pairs","feed_subtitle":"The leading-color piece now holds for any quark mass, energy and scattering angle, not just the small-mass limit.","key_machinery":"The load-bearing object is the harmonic polylogarithm basis: every master integral needed here was already known from the authors' earlier quark-antiquark annihilation calculation or follows from it by swapping $t$ with $u$, so the new work is a reduction of the two-loop interference onto that fixed basis rather than a new class of integrals. The reduction is done with a distributed implementation of the Laporta algorithm, and the master integrals are evaluated by the differential equation method. The claim of exactness in $m$ rests on this basis being complete for the leading-color diagrams, which are all planar; the authors note explicitly that non-planar and fermionic diagrams, and some heavy-quark-loop diagrams containing a three-propagator sunrise sub-topology, fall outside it and would need elliptic integrals or numerical methods.","core_discovery":"The paper reports an analytic expression for the coefficient $A$ multiplying $N_c^3$ in the two-loop interference of $gg$ fusion with the Born amplitude for $gg$ into a heavy-quark pair, retaining the complete dependence on $s$, $t$, the renormalization scale $m$ and the heavy-quark mass $m$. The calculation covers the 300 of 789 two-loop diagrams that carry the leading color structure, together with the corresponding ghost-initiated diagrams, and renormalizes them with on-shell wave functions and heavy-quark mass and a modified-minimal-subtraction strong coupling. The result is expressed in one- and two-dimensional harmonic polylogarithms of maximum weight four, computed first in the unphysical region $s<0$ and continued analytically to the physical region. Agreement is reported with the earlier small-mass expansion of the same coefficient and with the independently derived infrared pole coefficients, and the expression is shown to be symmetric under $t$ exchange with $u$. A threshold expansion in the quark velocity $beta$ is given through order $beta^5$, with all logarithms of $beta$, $xi$ and $1-2xi$ cancelling, which the authors read as evidence that the coefficient is finite at threshold.","pith_inferences":["Because the leading-color coefficient is the planar piece, and because the authors already possess the machinery for the quark-antiquark channel, the same pipeline plausibly extends to several of the non-leading color coefficients at modest additional cost; the real barrier is the elliptic sunrise topology, not color bookkeeping.","The threshold expansion being polynomial in $\\beta^2$ and in $\\xi(1-\\xi)$, with all logarithms cancelling, is a structural statement that could be checked against the general factorization predictions for massive two-loop amplitudes, and would be a sharper test than either of the two limit checks the paper performs.","A natural next check that the paper does not perform is a direct numerical comparison at a point with finite $m^2/s$ and generic scattering angle against an independent numerical two-loop evaluation; the benchmark table the authors provide is exactly the set of points such a comparison would use."],"forward_implications":["The leading-color part of the $gg \\to t\\bar{t}$ two-loop amplitude is available at any mass, energy and scattering angle, removing the small-mass approximation from that color structure.","The supplied C++ evaluator makes the expression usable inside a future NNLO partonic Monte Carlo without re-deriving the integrals.","The threshold expansion through order $\\beta^5$ can be used to organize logarithmically enhanced leading-color terms near the top-pair production threshold.","The cancellation of all $\\ln\\beta$ terms in that expansion supports the expectation that the coefficient is finite at threshold, which constrains threshold resummation for this channel.","The result fixes one gauge-invariant sub-set of the two-loop $gg \\to t\\bar{t}$ matrix elements, leaving the fermionic and non-planar gluonic pieces as the remaining obstacles to a full NNLO prediction."],"supporting_citations":[{"why":"Supplies the small-mass expansion of the same leading-color coefficient that the new exact expression is checked against.","marker":"[23]"},{"why":"Supplies the independently derived infrared pole coefficients used as the second validation of the result.","marker":"[35]"},{"why":"Provides the master-integral basis and the functional HPL list from the quark-antiquark channel that this calculation reuses.","marker":"[26]"},{"why":"Earlier paper in the series establishing the calculational method and part of the master integrals.","marker":"[25]"},{"why":"The Laporta algorithm used to reduce the scalar loop integrals to master integrals.","marker":"[27]"},{"why":"The differential equation method by which the master integrals are evaluated analytically.","marker":"[28]"},{"why":"The reduction code that performs the integral reduction; cited as still to be published.","marker":"[37]"},{"why":"The numerical routines for evaluating multiple polylogarithms on which the supplied C++ code depends.","marker":"[50]"},{"why":"Source of the two-loop wave-function and mass renormalization constants used in the on-shell scheme.","marker":"[52]"}],"fun_headline_variants":["Leading-color two-loop heavy-quark pair production, any mass","Analytic leading-color two-loop result for gluon-fusion heavy quarks","Heavy-quark pair production: leading-color two-loop coefficient","Leading-color two-loop result holds for any s, t and quark mass","Threshold expansion through beta^5 for gluon-fusion heavy quarks"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The claim rests on the reduction and the analytic continuation being correct in the interior of the kinematic region, yet the only checks performed are two limits: the small-mass expansion and the infrared pole coefficients, neither of which probes finite $m^2/s$ at a generic angle, and the reduction itself is carried out with a code the paper cites as not yet published.","fun_headline_variants_meta":{"raw":{"variants":["Leading-color two-loop heavy-quark pair production, any mass","Analytic leading-color two-loop result for gluon-fusion heavy quarks","Heavy-quark pair production: leading-color two-loop coefficient","Leading-color two-loop result holds for any s, t and quark mass","Threshold expansion through beta^5 for gluon-fusion heavy quarks","Leading-color only: two-loop heavy-quark production, any mass","Two-loop leading-color piece for heavy quarks at any energy and angle","Gluon-fusion heavy quarks: leading-color two-loop analytic expression","Heavy-quark pair production keeps full mass dependence at two loops","Leading-color two-loop coefficient agrees with IR pole checks"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":1850,"prompt_tokens":857,"completion_tokens":993,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":843}},"tokens_in":857,"tokens_out":993,"duration_ms":18939,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":843,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-10T18:30:42.787914+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"An independent numerical evaluation of the two-loop $gg$ interference at one of the paper's own benchmark points, for instance $s/m^2 = 5$, $t/m^2 = -1.25$ with $mu = m$, compared against the quoted coefficients of the $epsilon$ expansion; a mismatch in the finite part would show that the analytic continuation or the reduction chain fails away from the two checked limits.","supporting_citations":[{"cited_title":"Heavy-quark production in gluon fusion at two loops in QCD","cited_arxiv_id":"0707.4139","evidence_quote":"Supplies the small-mass expansion of the same leading-color coefficient that the new exact expression is checked against."},{"cited_title":"Two-loop divergences of massive scattering amplitudes in non-abelian gauge theories","cited_arxiv_id":"0908.3676","evidence_quote":"Supplies the independently derived infrared pole coefficients used as the second validation of the result."},{"cited_title":"Two-Loop Planar Corrections to Heavy-Quark Pair Production in the Quark-Antiquark Channel","cited_arxiv_id":"0906.3671","evidence_quote":"Provides the master-integral basis and the functional HPL list from the quark-antiquark channel that this calculation reuses."},{"cited_title":"Two-Loop Fermionic Corrections to Heavy-Quark Pair Production: the Quark-Antiquark Channel","cited_arxiv_id":"0806.2301","evidence_quote":"Earlier paper in the series establishing the calculational method and part of the master integrals."},{"cited_title":"The analytical value of the electron (g-2) at order alpha^3 in QED","cited_arxiv_id":"hep-ph/9602417","evidence_quote":"The Laporta algorithm used to reduce the scalar loop integrals to master integrals."},{"cited_title":"Differential Equations for Feynman Graph Amplitudes","cited_arxiv_id":"hep-th/9711188","evidence_quote":"The differential equation method by which the master integrals are evaluated analytically."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The reduction code that performs the integral reduction; cited as still to be published."},{"cited_title":"Numerical evaluation of multiple polylogarithms","cited_arxiv_id":"hep-ph/0410259","evidence_quote":"The numerical routines for evaluating multiple polylogarithms on which the supplied C++ code depends."},{"cited_title":"The three loop on-shell renormalization of QCD and QED","cited_arxiv_id":"hep-ph/0005131","evidence_quote":"Source of the two-loop wave-function and mass renormalization constants used in the on-shell scheme."}],"review_version":1}