{"id":"c82ea968-b621-4a2d-a702-12254b90c081","arxiv_id":"2607.15188","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The muon-line radiative-recoil correction to the muonium Lamb shift at order α^6(m/M)^2 m is found to be π² times the common prefactor, replacing a previously quoted (139/32−2 ln2) coefficient.","lead":"This paper computes a tiny QED correction to muonium's energy levels: the radiative-recoil term of order α^6(m/M)^2 m. The result, a simple π² coefficient, replaces an incorrect value in the current theory compilation and is directly relevant to upcoming precision muonium experiments.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central coefficient in Eq. (14) rests on an unverified restriction to two-photon exchange; the paper’s own footnote promises qualifications that never appear.","rationale":"The reader’s weakest-assumption identification is correct and is also the load-bearing concern from my reading. The paper explicitly says the naive expectation of many-photon contributions is realized in the previous order Z^2α(Zα)^4(m/M)^2m, so the claim that it does not happen at the next order is not self-evident. The argument is delegated to [17,18] and the promised qualifications in footnote 1 are absent. This matters because Eq. (14) is a clean π² coefficient; if any extra-photon diagram contributed, the result would be incomplete. I do not see an internal algebraic inconsistency in the cancellations: the 1/γ terms cancel, the logarithmic divergences cancel among the diagrams, and the surviving π² is a simple outcome. The authors are established in this niche and the cited references likely contain the relevant argument, but the manuscript itself does not supply it, and the missing footnote is an explicit warning. Therefore CONDITIONAL remains the appropriate verdict. I would not escalate to REJECT because the concern is a missing derivation, not a demonstrated error, and I would not downgrade to ACCEPT because the central premise is not verifiable from the text alone.","tokens_in":6390,"tokens_out":9707,"duration_ms":80473,"concrete_test":"Analytically check the power counting for diagrams with one additional exchanged photon: take the radiatively corrected muon factor H_μν from Eq. (5), insert an extra photon into the two-photon kernel, and examine the leading small-momentum (k ≲ mZα) behavior, comparing the power of Zα with the Fig. 1 contribution. In particular, isolate the non-softened 1/k pieces of H_μν (anomalous magnetic moment and electric form-factor slope) and determine whether they generate a three-or-more-photon contribution at the same nominal order as Eq. (14). If the extra photon is not suppressed by at least one power of Zα, the central coefficient is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim Eq. (14) depends on the p.2 assertion that only the two-photon-exchange diagrams of Fig. 1 contribute at order Z^2α(Zα)^5(m/M)^2m, because 'addition of any extra exchanged photon always produces an extra power of Zα.' This is the least secure step: the same statement is acknowledged to fail at the neighboring order Z^2α(Zα)^4(m/M)^2m, and the paper does not derive the softening for the present order—it cites [17,18] and footnote 1 promises 'qualifications of this statement below' that never appear in the text. Since Eq. (14)'s coefficient is exactly 1 after the π² cancellation, any unsuppressed extra-photon diagram would change the result; the numerical gap from the compilation value (139/32−2 ln2) makes this sensitivity concrete. The algebraic cancellations (1/γ terms, logarithmic divergences) are internally coherent, but the diagram-selection premise is a genuine gap rather than a demonstrated error.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a new calculation of the radiative-recoil correction to the muonium Lamb shift of order Z^2 α (Zα)^5 (m/M)^2 m. Starting from the two-photon-exchange scattering-approximation formula Eq. (1), the authors evaluate the diagrams of Fig. 1 (muon-line self-energy, vertex, and spanning-photon insertions, plus crossed exchanges) in the Yennie gauge. They find that the linearly divergent 1/γ terms cancel and that the logarithmic infrared divergences cancel among the three contributions, leaving JΣ + 2JΛ + JΞ = π^2 after omitting the lower-order µ/γ part. This gives the central result Eq. (14), ΔE_μ = (Z^2 α)(Zα)^5 n^{-3} (m_r^3/M^2) δ_{l0}, replacing the previously quoted coefficient (139/32 − 2 ln 2) from the compilation [11]. Adding the electron-line result Eq. (15) yields the combined α^6 coefficient (−95/32 + 8 ln 2).","tokens_in":6541,"tokens_out":2586,"duration_ms":22403,"significance":"If correct, this is a genuine advance: it removes a suspected error in the current compilation of muonium energy levels and provides a needed theory input for the ongoing 1S−2S and 2S−2P experiments. The paper is admirably transparent about the previous value and about the internal consistency checks: the cancellation of the 1/γ and logarithmic divergences is explicitly tracked through Eqs. (10)–(13), and the reduced-mass prefactor is handled exactly. The final result is parameter-free and falsifiable by comparison with future precision data. However, the two most load-bearing steps — the values of JΛ and JΞ in Eq. (12) and the restriction to the Fig. 1 two-photon-exchange diagrams — are not demonstrated in the manuscript. The internal algebra is coherent, but the central claim cannot be fully verified from the text as written.","major_comments":[{"comment":"The results JΛ = π^2/4 + 3S0 − 12S2 + (32/3)(µ/γ)(ln(1/γ) − 1/3) and JΞ = π^2/2 − 3S0 + 12S2 − (16/3)(µ/γ) are stated without derivation. These values are load-bearing: the final coefficient 1 in Eq. (14) is exactly the π^2/π^2 cancellation generated by these terms, and the claimed cancellation of S0 and S2 depends on their coefficients. The manuscript should show the integral representations for JΛ and JΞ and at least an outline of how the angular/momentum integrations produce the quoted results. Without this, Eq. (12) is an unverifiable assertion rather than a derived result.","section":"§3, Eq. (12)"},{"comment":"The claim that only the two-photon-exchange diagrams of Fig. 1 contribute at order Z^2 α(Zα)^5(m/M)^2m is justified by the statement that 'the infrared behavior of any radiatively corrected Feynman diagram ... is softer than the behavior of the respective skeleton diagram' and by a reference to [17,18]. The manuscript itself notes that the analogous statement fails at the neighboring order Z^2 α(Zα)^4(m/M)^2m, and footnote 1 promises 'qualifications of this statement below' that do not appear anywhere in the text. Because Eq. (14) has no leftover numerical coefficient, any unsuppressed extra-photon diagram at the same nominal order would change the result by an O(1) factor. A direct power-counting argument for this specific order, or an explicit statement of the promised qualifications, is required to make the diagram selection secure.","section":"p.2, diagram-selection argument"},{"comment":"The paper discards the linearly divergent terms proportional to µ/γ in Eq. (13), saying they 'produce well known contributions of the previous order' and citing [20]. This is plausible because the previous-order contributions are known, but the manuscript does not show that the retained π^2 term is unambiguously separated from the lower-order terms after the scattering-approximation cutoff is removed. Since the whole calculation is performed in the scattering approximation, the identification of the finite piece at order µ^2 relative to the leading µ/γ divergence should be justified more explicitly; otherwise the coefficient in Eq. (14) may depend on the regularization of the linear divergence.","section":"Eqs. (7)–(13), treatment of µ/γ terms"}],"minor_comments":[{"comment":"The promise of 'qualifications of this statement below' is never fulfilled. Either supply the qualifications or delete the footnote.","section":"Footnote 1 / p.2"},{"comment":"The notation 'γ<µ→0' is hard to parse. It should be written as a limit, e.g., 'in the limit γ → 0 with γ ≪ µ', and the definitions of the cutoff γ and of the functions S0, S2 should be stated explicitly.","section":"Eq. (11)"},{"comment":"The notation 'Z^2α(Zα)^5' is redundant (and, for muonium, Z=1). Consider writing the order as α^6(m/M)^2 or, if the general-Z form is intended, define it once. Also, Eq. (16) writes α^6, which is consistent with Z=1, but the transition from Z^2α(Zα)^5 to α^6 should be noted.","section":"Eq. (14) and abstract"},{"comment":"The DOI '10.1103/f1z4-xzq2' looks like a placeholder and should be replaced with the correct identifier.","section":"Reference [12]"},{"comment":"The phrase 'Apparently the authors of [11] noticed...' is speculative. It would be cleaner to say that the quoted result has the same functional form as the nonrecoil α(Zα)^5 contribution and to state plainly that the substitution rule is invalid at this order.","section":"p.4, wording around Eq. (9)"},{"comment":"Some displayed equations use nonstandard spacing in the exponent 'Z 2α' and in the factor '1 /γ'; these should be corrected in the final manuscript.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper is short and relies heavily on the authors' earlier work [13,14,19,22]. The self-citations are appropriate for a calculation of this type and do not appear circular. The main concern is completeness: the decisive algebraic results (JΛ, JΞ) and the diagram-selection argument are not shown. The referee report asks for these to be supplied; if the authors can provide them, the paper may be suitable for publication. I would not reject at this stage, since there is no demonstrated internal error, but the manuscript in its current form does not meet the verification standard expected for a result that changes a published coefficient."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short take: a genuine new coefficient for the muonium Lamb shift, and it corrects a mistake in the current compilation. The paper is thin in the wrong place — the two decisive integrals, JΛ and JΞ, are quoted without derivation — but the cancellations it does display are coherent, and the result is plausible.\n\nThe new thing is Eq. (14): the Z^2 α(Zα)^5 (m/M)^2 m radiative-recoil contribution to the 1S-2S and 2S-2P intervals is exactly (Z^2 α)(Zα)^5 n^{-3} m_r^3/M^2 δ_{l0}, i.e. coefficient 1 after the π^2 from the momentum integrals cancels the 1/π^2 in the prefactor. Combined with the known electron-line result this gives (−95/32 + 8 ln2) α^6 n^{-3} m_r^3/M^2, replacing the (139/32 − 2 ln2) entry in [11] that the authors argue came from an invalid substitution α→Z^2α and m→M. That correction to the record is concrete and useful for the current round of muonium experiments.\n\nWhat is good: the internal bookkeeping is clean. The 1/γ linear divergences cancel, the logarithmic IR divergences hidden in S0 and S2 cancel between diagrams, and the leftover µ/γ term reproduces the known lower-order contribution. The attribution of the compilation error is persuasive, and the connection to the electron-line radiative-recoil result [21] makes the combined value natural.\n\nSoft spots: a referee cannot check the central result without redoing the calculation, because JΛ and JΞ are simply stated. That may be acceptable in this niche, but it makes the paper a claim rather than a demonstrated computation. The more serious conceptual point is the restriction to exactly two photon exchanges. The paper argues on p.2 that any extra exchanged photon costs a power of Zα because radiatively corrected diagrams have softer low-momentum behavior, and it cites [17,18] for details. The neighboring order Z^2 α(Zα)^4 (m/M)^2 m explicitly does not have this property, so the burden is on the cited reviews to carry the argument. That is probably fine, but the footnote promising 'qualifications of this statement below' is never fulfilled — the text just moves on. Either the qualification should be given or the footnote deleted.\n\nFor whom: this is for the very small audience that maintains muonium theory compilations and for precision QED people who care about few-hundred-Hz corrections. It deserves a serious referee, and I would ask for derivations of JΛ and JΞ and for the promised footnote to be sorted out. Not a desk reject.","headline":"Genuine new coefficient for the muonium Lamb shift that corrects the recent compilation, but the decisive integrals are quoted rather than shown and a promised footnote never materialises — worth refereeing carefully.","tokens_in":7106,"tokens_out":4147,"would_cite":true,"duration_ms":33110,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81V45","81T18"],"pacs":["12.20.-m","31.30.jf","36.10.Dr"],"model":"deepseek-v4-flash","headline":"The paper shows the Z²α(Zα)^5(m/M)²m radiative-recoil correction to the muonium Lamb shift has coefficient 1, not the value in a recent compilation, and gives the combined electron-plus-muon term.","keywords":["muonium","Lamb shift","radiative-recoil correction","QED bound states","two-photon exchange","mass-ratio expansion","fine structure","scattering approximation"],"falsifier":"Calculate the three-photon-exchange diagrams (or an equivalent NRQED/NRQCD matching) at order Z²α(Zα)^5(m/M)²m; if they contribute a non-vanishing δ_{l0} term, Eq. (14) is incomplete. A lighter check: evaluate the same three diagram classes in Feynman gauge instead of the Yennie gauge; the final coefficient must be unchanged, so any residual gauge dependence would signal an error in the infrared subtractions.","tokens_in":6196,"feed_emoji":"⚛️","tokens_out":7467,"duration_ms":60506,"temperature":0.7,"pith_summary":"The paper computes a radiative-recoil contribution to the muonium Lamb shift of order Z²α(Zα)^5(m/M)²m, the next uncalculated term in the mass-ratio expansion. It finds that when the two-photon-exchange diagrams with radiative insertions on the muon line are evaluated in the scattering approximation, the momentum integral supplies a factor π² that cancels the 1/π² in the prefactor, leaving the energy shift with coefficient exactly 1. The paper also shows that a value quoted in a recent compilation came from an invalid scaling of a nonrecoil result, and it combines the new muon-line term with the known electron-line term to give a total α⁶(m/M)² correction of (-95/32 + 8ln2). A sympathetic reader cares because new muonium 1S–2S and 2S–2P measurements have reached the precision where such a term matters.","feed_headline":"π² cancels: new muonium Lamb-shift term has coefficient 1","feed_subtitle":"The correction replaces a mistaken compiled value and sharpens theory for new muonium spectroscopy.","key_machinery":"The central object is the two-photon-exchange scattering-approximation integral, Eq. (1), which expresses the hard spin-independent energy shift as a loop integral over the product of a light-lepton factor L_μν and a heavy-lepton factor H_μν. The heavy factor is radiatively corrected by one-loop self-energy, vertex, and spanning-photon insertions; choosing the Yennie gauge for radiative photons makes the individual integrals tractable. After rescaling the loop momentum by the heavy mass, the muon factor carries an explicit Z²α/π, and the electron factor supplies a factor µ = m/M. Linearly infrared-divergent terms of order 1/γ and logarithmic divergences cancel among the three diagram classes","core_discovery":"On its own terms, the paper's finding is Eq. (14): the radiative-recoil contribution of order Z²α(Zα)^5(m/M)²m to the muonium Lamb shift is ΔE_μ = (Z²α)(Zα)^5 n^{-3} (m_r³/M²) δ_{l0}, with no logarithmic or rational coefficient. The π² factor coming from the vertex and spanning-photon integrals cancels the 1/π² factor in the scattering-approximation formula (7). After adding the electron-line contribution of the same order, the combined spin-independent correction is ΔE_t = α⁶ n^{-3} (m_r³/M²)(-95/32 + 8ln2). The paper further argues that the previously tabulated coefficient (139/32 − 2ln2) arose from substituting α→Z²α and m→M in the nonrecoil result and multiplying by (m/M)², a recipe that","pith_inferences":["The scattering-approximation integral is exact in the mass ratio, so the same diagrams evaluated at m = M would give the analogous radiative-recoil correction in positronium; a straightforward adaptation of Eq. (14) would yield a definite α⁶ coefficient there.","A gauge-independence check of the π² cancellation would be valuable: repeating the calculation in Feynman gauge or any other gauge should leave the final coefficient invariant after summing all three diagram classes; if a residual gauge dependence appears, one of the infrared subtractions is incomplete.","The paper's central assumption—that extra exchanged photons are suppressed—can be tested by evaluating the three-photon-exchange diagrams at the same nominal order; if they contribute, the quoted coefficient would be incomplete, just as at the neighboring order where such diagrams are known to matter."],"forward_implications":["The muonium Lamb shift, the 2S–2P interval, and the 1S–2S transition each gain a definite correction of order α⁶(m/M)²; for S-states it is (-95/32 + 8ln2) α⁶ m_r³/(n³M²) once both lepton lines are included.","Recent compilations that quote 139/32 − 2ln2 for the Z²α(Zα)^5(m/M)²m term need to be revised; the error is traceable to scaling a nonrecoil correction rather than computing the recoil diagrams.","The coefficient is simple and free of logarithms after the π² cancellation, so the contribution can be combined directly with other α⁶ terms in a full theory prediction.","Because the term is proportional to δ_{l0}, it affects S-states only, leaving the fine-structure splitting unchanged at this order.","The cancellation of linear and logarithmic infrared divergences among the three diagram classes provides a consistency check that the computed coefficient is complete within the two-photon sector."],"fun_headline_variants":["π² cancels: muonium Lamb-shift term gets clean coefficient 1","Radiative recoil term simplifies: π² cancels to coefficient 1","New muonium correction fixes wrong compiled value","One more muonium correction: π² cancels, coefficient 1","Muonium theory: radiative-recoil term becomes simple 1"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The result stands only if the two-photon-exchange diagrams are the entire story at this order; the paper assumes that any additional exchanged photon adds an extra power of Zα because radiatively corrected diagrams have softer low-momentum behavior, an assumption that is known to fail for the neighboring order Z²α(Zα)^4(m/M)²m and is referenced rather than derived here.","fun_headline_variants_meta":{"raw":{"variants":["π² cancels: muonium Lamb-shift term gets clean coefficient 1","Radiative recoil term simplifies: π² cancels to coefficient 1","New muonium correction fixes wrong compiled value","One more muonium correction: π² cancels, coefficient 1","Muonium theory: radiative-recoil term becomes simple 1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000204,"raw_usage":{"total_tokens":1178,"prompt_tokens":648,"completion_tokens":530,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":392,"completion_tokens_details":{"reasoning_tokens":437}},"tokens_in":392,"tokens_out":530,"duration_ms":4437,"temperature":1.0,"reasoning_tokens":437,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T23:53:03.679043+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate the three-photon-exchange diagrams (or an equivalent NRQED/NRQCD matching) at order Z²α(Zα)^5(m/M)²m; if they contribute a non-vanishing δ_{l0} term, Eq. (14) is incomplete. A lighter check: evaluate the same three diagram classes in Feynman gauge instead of the Yennie gauge; the final coefficient must be unchanged, so any residual gauge dependence would signal an error in the infrared subtractions.","supporting_citations":[],"review_version":1}