{"id":"00dddd9d-4e66-4496-87d6-06835251351d","arxiv_id":"2607.15214","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"NNLO QCD corrections to pion- and kaon-induced exclusive Drell-Yan are obtained by analytic continuation and found to be large, often exceeding NLO.","lead":"This paper computes two-loop (NNLO) quantum chromodynamics corrections to exclusive Drell-Yan scattering induced by pion and kaon beams, and finds the corrections are large and positive. The results are aimed at upcoming J-PARC experiments that plan to use meson beams to map the internal structure of the proton.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (10) is the linchpin of the NNLO claim, but its continuation is only stated, not proved; a sign mismatch with Eq. (7) and unaddressed xi-singularity effects could invalidate the results.","rationale":"The reader correctly puts the paper on a conditional footing, pointing at leading-twist accuracy in the moderate-Q region and at the unverified imported kernel. I agree with those concerns. But the single most load-bearing point is more specific and internal: the all-order continuation identity, Eq. (10), is the mechanism that manufactures the NNLO coefficients. The published text does not derive it beyond a conjugation statement, and it leaves a visible sign tension with Eq. (7)'s +i pi timelike logarithm. It also does not address contour effects from changing the skewness sign in the GPD convolution. If Eq. (10) is defective, the numerical NNLO corrections - and hence the abstract's central claim - are unsupported regardless of GPD choice or higher-twist contamination. This is a correctness risk in the core derivation rather than a disagreement with the external factorization consensus, so it should be resolved by an explicit check. The paper itself admits in Section IV that the large NNLO corrections cast doubt on the perturbative expansion; that admission makes the need for a verified bridge even more acute, but the continuation identity is the more fundamental issue. I therefore recommend keeping the conditional verdict, with the condition explicitly requiring validation of Eq. (10) at one loop and eventually at two loops.","tokens_in":11515,"tokens_out":12613,"duration_ms":103146,"concrete_test":"Compute the one-loop exclusive Drell-Yan TFF for pi^- p -> gamma^* n directly from the known pion EMFF hard-scattering kernel of refs. [23,34], using Eq. (6) with xi=-eta and the timelike log in Eq. (7), without invoking Eq. (10); compare with Eq. (10) applied to the one-loop DVMP TFF. Agreement on both the sign of the i pi term and the x-convolution would support the identity; any disagreement would invalidate the NNLO results. If the NLO check passes, repeat the test at two loops with the explicit coefficient functions of ref. [32] once they are made available.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation (10) is the bridge from the imported NNLO DVMP kernel [32] to the central NNLO cross sections. The derivation in Section II.B is a single line: v_exDY = v*_DVMP, conjugation of C_i^l, and then F_exDY = F*_DVMP + sum (-i pi d/d log Q^2)^n F*_DVMP. Two unstated assumptions are load-bearing. First, Eq. (7) continues log(mu^2/(-Q^2)) = log(mu^2/Q^2) + i pi, but Eq. (10) implements a shift by -i pi (the RHS equals F*_DVMP(log Q^2 - i pi)). The paper never explains why conjugation of F_DVMP changes the sign of the i pi term. Second, Eq. (10) only shifts the explicit logarithms; the hard-scattering coefficient also depends on v (Eq. (9)), and the convolution in Eq. (6) changes the skewness from +eta for DVMP to -eta for Drell-Yan. A sign change in the x-integration can move the singularities at x=+-xi across the contour, producing contributions not captured by substituting v* and shifting log Q^2. If either effect is real, the NNLO TFFs do not follow from Eq. (10) even if ref. [32] is correct. No explicit two-loop T(u,x,xi) or numerical code is provided, so this hidden assumption is currently untested.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims the first NNLO QCD calculation for the exclusive pion- and kaon-induced Drell-Yan processes π^- p → γ^* n and K^- p → γ^* Λ in leading-twist collinear factorization. The technical route is crossing: Eq. (10) relates the Drell-Yan transition form factors to complex-conjugated DVMP form factors through a logarithmic shift, and the NNLO DVMP kernel from the companion paper [32] is imported. Numerical predictions for J-PARC kinematics (Q^2 ~ 2–6 GeV^2, τ ~ 0.2) are made with two GPD models (GK and GUMP) and lattice-motivated pion/kaon DAs. The central phenomenological claim is that NNLO corrections are large and positive, often exceeding 100% of the already large NLO corrections, so their inclusion is imperative. The paper also evaluates the transverse single-spin asymmetry and finds it relatively stable order by order.","tokens_in":11900,"tokens_out":9888,"duration_ms":80632,"significance":"If Eq. (10) and the imported two-loop kernel are correct, this is a useful and timely first NNLO study of these exclusive Drell-Yan channels, directly relevant to the proposed J-PARC measurements. The paper benefits from using two independent GPD parametrizations, explicit DA inputs, scale-variation uncertainties, and a polarization observable. The crossing strategy is plausible and the NLO expression in Eq. (8) is explicit. However, the paper is not self-contained: the all-order relation Eq. (10) is not derived, no NNLO coefficient function is displayed, and the numerical results depend on companion work [32] with overlapping authorship. The phenomenological conclusion is also drawn entirely from leading-twist predictions in a moderate-Q^2 region where the paper itself concedes higher-twist effects can be significant.","major_comments":[{"comment":"The all-order relation F_exDY = F*_DVMP + Σ (1/n!)(-iπ d/dlog Q^2)^n F*_DVMP is the linchpin of the NNLO claim, but it is stated without derivation. Starting from Eq. (7) with L = log(μ^2/Q^2), the timelike continuation is L → L + iπ; a Taylor expansion of F*_DVMP(L) gives exp(iπ d/dL) = exp(-iπ d/dlog Q^2) only if the derivative acts at fixed α_s, fixed GPDs and DAs, and if the coefficient functions are real-analytic in v. None of these conditions is stated. In addition, Eq. (6) changes ξ from +η (DVMP) to −η (Drell-Yan), and the x-integration contour can cross the singularities at x = ±ξ; the paper does not justify the contour deformation. A direct NLO verification of Eq. (10) against Eq. (8) would resolve whether the sign and contour are correct. Without that, the NNLO results rest on an unverified assumption.","section":"§II.B, Eq. (10)"},{"comment":"No NNLO coefficient function C_i^2(u,v) is displayed anywhere, and no numerical code is provided. The entire NNLO prediction is imported from ref. [32] through Eq. (10). Since ref. [32] has overlapping authorship with this manuscript, this is not an independent check. A reader cannot verify the continuation or the numerical implementation. At minimum, the paper should display one explicit two-loop coefficient function (or provide an ancillary file) and compare the NLO limit of the continued formula with the direct NLO calculation from Eq. (8).","section":"§III and §IV"},{"comment":"The abstract and summary state that NNLO corrections are 'imperative for reliable theoretical predictions' and often exceed 100% of the NLO corrections. These claims are drawn entirely from leading-twist predictions at Q^2 ~ 2–6 GeV^2. The paper itself notes after Eq. (3) that higher-twist contributions 'can remain numerically significant in the moderate-Q region.' No estimate of higher-twist or target-mass corrections is provided, and the two GPD models differ by roughly an order of magnitude in the plotted cross sections. The central phenomenological conclusion should either be supported by a quantitative estimate of neglected terms or be reformulated as a statement about the leading-twist perturbative series only.","section":"§II.A and §IV"}],"minor_comments":[{"comment":"The sign convention between ξ and η is confusing: Eq. (2) defines ξ negative for Drell-Yan, Eq. (5) states η ≈ −ξ, and Eq. (6) uses ξ = η for DVMP but ξ = −η for Drell-Yan. Please make the notation consistent throughout, especially in Eq. (10).","section":"Eqs. (2), (5), (6)"},{"comment":"The text says singularities at x = ±ξ are regulated by ξ → ξ − iε, but Eq. (8) has denominators (ξ − x − iε) and (ξ + x − iε). Clarify how these two prescriptions are related and how the continuation to v* affects the iε terms.","section":"Eqs. (7)–(8)"},{"comment":"The symbol n_L = 3 in the numerical setup is not defined. If it means the number of active flavors or the loop order, please define it explicitly.","section":"Section III"},{"comment":"Ref. [30] is incomplete in the bibliography as printed (missing journal/volume/page or arXiv identifier). Please correct.","section":"References"},{"comment":"The text repeatedly states that NNLO corrections 'exceed 100% of the already large NLO corrections' but gives no explicit K-factors or a table quantifying LO→NLO→NNLO ratios. A table would make the central claim quantitative and easier to verify from the figures.","section":"Section III"}],"recommendation":"major_revision","confidential_remarks":"The main technical risk is the unverified analytic continuation in Eq. (10) and the complete reliance on the companion paper [32] for the two-loop input. Because of the author overlap with [32], the editor may wish to ensure that [32] is independently available and accepted before publication of this manuscript. The paper is more a phenomenological application of [32] than a self-contained NNLO calculation; this is acceptable if the missing derivation and an explicit NNLO check are added. I do not see grounds for rejection if the authors can supply the derivation and an NLO cross-check."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline is that this paper gives the first NNLO predictions for pion- and kaon-induced exclusive Drell-Yan at J-PARC kinematics, but the central analytic continuation, Eq. (10), is underived and looks to have a sign problem. That makes the quantitative results conditional.\n\nWhat the paper does well: it takes the group's recent NNLO DVMP calculation [32] and uses crossing to produce exDY predictions. The kaon channel and the TSSA analysis are genuinely new and add modest value. The numerics are presented clearly, with two GPD models and scale-variation bands. If [32] is right and Eq. (10) is right, the conclusion that NNLO corrections are large and positive follows.\n\nWhere it is soft: Eq. (10) is the linchpin, introduced in one sentence. Eq. (7) says the timelike log is log(mu^2/Q^2)+i pi, but Eq. (10) shifts F*_DVMP by -i pi d/d log Q^2. That is a mismatch. Unless there is a compensating step not written down, the sign of the analytic continuation is opposite to the one implied by Eq. (7). Also, changing the skewness from +eta to -eta alters the convolution contour and the xi-singularity treatment; the paper only swaps v -> v* and shifts logarithms, which does not obviously account for that. The authors should show the explicit two-loop coefficient functions or provide code. As it stands, the NNLO numbers are not independently checkable.\n\nOther soft spots: the paper itself admits higher-twist effects can be significant at Q^2 ~ 2-6 GeV^2, but the central phenomenological claim is drawn entirely from leading twist. The GPD model spread (GUMP vs GK) is an order of magnitude, so the actual cross sections are not predictions in a strong sense. The 100% NNLO corrections are so large that they undermine the reliability of the perturbative expansion; the authors acknowledge this but offer no cure.\n\nBottom line: the qualitative conclusion that higher orders matter is probably right; the specific NNLO numbers should not be used until Eq. (10) is derived carefully or confirmed by an independent continuation. This deserves a serious referee because the subfield cares about J-PARC planning and the companion DVMP result is important. But my recommendation is major revision, not acceptance.","headline":"First NNLO exDY predictions rest on a one-line analytic continuation with an apparent sign inconsistency; interesting phenomenology, conditional numbers.","tokens_in":12409,"tokens_out":4463,"would_cite":false,"duration_ms":36354,"reading_group":"yes","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 proves an all-order analytic-continuation relation that turns known two-loop deeply virtual meson production kernels into NNLO QCD predictions for pion- and kaon-induced exclusive Drell-Yan processes, and finds the NNLO correctio","keywords":["exclusive Drell-Yan","generalized parton distributions","NNLO QCD corrections","analytic continuation","transition form factors","distribution amplitudes","transverse single-spin asymmetry","leading-twist factorization"],"falsifier":"Measure the lepton-pair angular distribution in pion-beam exclusive Drell-Yan at Q^2 ≈ 5 GeV^2 and tau ≈ 0.2, isolate the longitudinal-photon part, and compare the integrated |t'| ≤ 0.5 GeV^2 cross section with the NNLO prediction; a disagreement beyond the scale-variation band would indicate missing higher-twist contributions or an incorrect input kernel.","tokens_in":11404,"feed_emoji":"⚛️","tokens_out":6728,"duration_ms":51840,"temperature":0.7,"pith_summary":"The paper aims to establish that next-to-next-to-leading-order (NNLO) QCD corrections to pion- and kaon-induced exclusive Drell-Yan processes are large and positive, frequently exceeding the size of the already large NLO corrections, so any reliable extraction of nucleon generalized parton distributions from upcoming pion- and kaon-beam lepton-pair experiments must include them. To obtain these two-loop corrections without a fresh calculation, the paper proves an all-order relation between the exclusive Drell-Yan transition form factors and the complex conjugate of the deeply virtual meson production transition form factors, then feeds in a recently computed NNLO deeply virtual meson production kernel. It also shows that the transverse single-spin asymmetry is comparatively stable under these higher-order corrections, which makes it a cleaner observable. A sympathetic reader would care because this is the first NNLO treatment of these processes, and it changes the predicted cross sections substantially over much of the planned kinematic range.","feed_headline":"Two-loop corrections double NLO effects in pion and kaon Drell-Yan","feed_subtitle":"Omitting these QCD terms would skew what pion and kaon beam experiments can say about nucleon structure.","key_machinery":"The load-bearing device is Eq. (10), the all-order analytic-continuation identity F_exDY = F*_DVMP + sum_{n>=1} (1/n!)(-i pi d/d log Q^2)^n F*_DVMP for the transition form factors. It follows from expressing the hard-scattering kernel in terms of the pion electromagnetic form factor kernel and identifying v_exDY = v*_DVMP, so the timelike logarithm acquires an i pi. This identity converts any known deeply virtual meson production hard-scattering kernel, currently known through NNLO, into the exclusive Drell-Yan kernel at the same order, bypassing a separate two-loop calculation.","core_discovery":"On its own terms, this paper claims that the transition form factors governing pion- and kaon-induced exclusive Drell-Yan are not independent perturbative objects: Eq. (10) expresses them as the complex conjugate of the corresponding deeply virtual meson production form factors plus a series of derivatives with respect to log Q^2 weighted by powers of (-i pi), an all-order consequence of analytically continuing the timelike logarithm. Because of this relation, the two-loop hard-scattering kernels for pi^- p -> gamma* n and K^- p -> gamma* Lambda are obtained by analytic continuation from the known NNLO deeply virtual meson production kernel. Numerically, within leading-twist collinear factor","pith_inferences":["If the leading-twist expansion is as slowly converging as the size of these NNLO terms suggests, resummed or higher-twist-improved predictions may be needed before the same hard-scattering coefficients can be used to extract GPDs; the paper itself notes that no known method resolves this issue.","The same analytic-continuation dictionary likely applies to other timelike GPD observables, such as timelike Compton scattering or crossed channels of other hard exclusive processes, offering a general way to recycle spacelike NNLO kernels.","A testable extension would measure the ratio of kaon- to pion-induced cross sections; the NNLO treatment predicts an SU(3)-breaking pattern through the kaon distribution amplitude and strange-quark GPDs that could be checked independently.","Because the asymmetry is NNLO-stable, even a low-luminosity polarized-target measurement could usefully constrain the GPD E, a point the paper states but does not develop as a dedicated experimental strategy."],"forward_implications":["NNLO corrections must be included before comparing leading-twist predictions with future pion- and kaon-beam lepton-pair data; dropping them substantially understates the predicted cross sections.","The all-order relation means any future extension of deeply virtual meson production to higher orders automatically upgrades exclusive Drell-Yan predictions to the same order without new diagram calculations.","The transverse single-spin asymmetry is stable against NLO and NNLO corrections, so it can serve as a cleaner observable for constraining the helicity-flip generalized parton distribution than the unpolarized cross section.","The two phenomenological GPD parametrizations used in the paper differ by roughly an order of magnitude in cross section even after NNLO corrections, so cross-section data will mainly discriminate among GPD models unless the hard-scattering piece is pinned down."],"fun_headline_variants":["Two-loop QCD corrections vital for pion, kaon Drell-Yan","NNLO QCD doubles NLO effects in meson-induced Drell-Yan","Pion and kaon Drell-Yan demand two-loop QCD","Two-loop terms reshape pion-kaon Drell-Yan predictions"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that leading-twist collinear factorization is quantitatively accurate at the planned moderate photon virtualities, where higher-twist and target-mass effects are neglected, and that the independently calculated two-loop deeply virtual meson production kernel used as input is correct.","fun_headline_variants_meta":{"raw":{"variants":["Two-loop QCD corrections vital for pion, kaon Drell-Yan","NNLO QCD doubles NLO effects in meson-induced Drell-Yan","Pion and kaon Drell-Yan demand two-loop QCD","Two-loop terms reshape pion-kaon Drell-Yan predictions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000661,"raw_usage":{"total_tokens":2842,"prompt_tokens":711,"completion_tokens":2131,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":2062}},"tokens_in":455,"tokens_out":2131,"duration_ms":12917,"temperature":1.0,"reasoning_tokens":2062,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T23:49:37.162034+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the lepton-pair angular distribution in pion-beam exclusive Drell-Yan at Q^2 ≈ 5 GeV^2 and tau ≈ 0.2, isolate the longitudinal-photon part, and compare the integrated |t'| ≤ 0.5 GeV^2 cross section with the NNLO prediction; a disagreement beyond the scale-variation band would indicate missing higher-twist contributions or an incorrect input kernel.","supporting_citations":[],"review_version":1}