{"id":"7311709f-98c2-4b25-992e-703d3841a089","arxiv_id":"2411.12459","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"New all-order two-loop self-energy data for Z=5-50, extrapolated to hydrogen, gives G60(Z=1) = -104.1(3.8), a 2.8 sigma shift from the accepted value.","lead":"Improved calculations of the two-loop electron self-energy in hydrogen-like ions reach lower nuclear charges and higher precision than before. The extrapolated hydrogen value disagrees with the accepted one by 2.8 standard deviations, which would shift the Rydberg constant by about one standard deviation.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The hydrogen value G60(1) = -104.1(3.8) depends on the unverified Pachucki coefficient B61; a change of about 3 (≈6%) in B61 shifts G60(1) by ≈28, erasing the claimed 2.8σ discrepancy, and the quoted uncertainty excludes this systematic.","rationale":"The reader identified the extrapolation's dependence on the analytic logarithmic constraints (B71, B83, B82, B81) as the weakest assumption. That is a real issue, but those coefficients are explicitly varied within their quoted ranges when estimating the ±3.8 uncertainty, so the extrapolation uncertainty budget partially accounts for them. My concern is different and potentially larger: the quantity being extrapolated, G60(Zα), is defined by subtracting the unverified coefficients B62 and B61 (Eq. 4). Any error in those coefficients shifts all extracted G60 values and propagates directly into G60(1). The paper's own scenario that δB61≈-5 could restore consistency shows that the discrepancy is within the plausible systematic range of an independently unverified calculation. Since the quoted 3.8 uncertainty does not include this source, the 2.8σ discrepancy and the Rydberg shift are not yet robust. The all-order numerical calculations for Z=5-50 are a substantial achievement and are likely correct; the concern is confined to the hydrogen extrapolation and the derived constant shifts. Therefore the conditional verdict remains appropriate, and no verdict change is needed. I partially agree with the reader because both concerns point to reliance on the same Zα-expansion apparatus, but the specific weak link I identify is the B62/B61 subtraction rather than the extrapolation's log constraints.","tokens_in":10636,"tokens_out":14488,"duration_ms":129896,"concrete_test":"Using the G60(Zi) values in Table I and the Supplement's extrapolation routine, vary the subtracted B61 and B62 coefficients: replace G60(Zi) by G60(Zi) + δB61·L(Zi) + δB62·L(Zi)^2, with L(Zi)=ln((Ziα)^-2), for δB61 ∈ {±1, ±3, ±5} and δB62 ∈ {±0.1, ±0.3, ±0.5}. Re-fit the cubic polynomial and extrapolate to Z=1. If |G60(1)+104.1| exceeds 3.8 for any |δB61|≤3 or |δB62|≤0.3, the 2.8σ claim is not robust. A positive control: δB61≈+2.9 should recover G60(1)≈-75.8, confirming the sensitivity.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is not the all-order data itself but the extrapolated hydrogen value G60(1) = -104.1(3.8) and the resulting 2.8σ deviation from the accepted value -75.8(9.4). This value is obtained by subtracting the Zα-expansion terms B40, B50, B63, B62, and B61 (Eq. 4) from the all-order F(Zα) data. Coefficients B62 and B61 are taken from Pachucki (2001) and, as the paper states, \"have yet to be independently verified.\" A change δB61 in B61 shifts every extracted G60(Zi) by +δB61·L(Zi) (L(Zi)=ln((Ziα)^-2)); at Z=1, L≈9.82, so δB61≈+2.9 moves G60(1) by ≈ +28, effectively reconciling with -75.8. The B62 coefficient is even more sensitive because it multiplies L^2≈96.5 at Z=1. This systematic dependence is absent from the quoted ±3.8 uncertainty, which covers only extrapolation scatter and the already-uncertain log coefficients B71, B83, B82, B81 of Eq. (6). The authors themselves entertain that \"a 10% adjustment in the B61 coefficient ... could restore the consistency.\" Therefore the claimed inconsistency between all-order and Zα-expansion approaches, and the derived Rydberg shift, may be an artifact of unverified input coefficients rather than a genuine breakdown of the Zα expansion. The numerical all-order results are likely sound, but the headline hydrogen conclusion is not yet robust.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports new all-order-in-Zalpha numerical calculations of the two-loop electron self-energy for the 1S state of hydrogen-like ions, extending the accessible range down to Z=5 and improving numerical accuracy by more than an order of magnitude through an accelerated subtraction scheme. The authors extract the higher-order remainder function G60(Zalpha) defined in Eq. (4), extrapolate it to Z=1 using a polynomial fit supplemented by analytic constraints on logarithmic coefficients, and obtain G60(1)=-104.1(3.8). This value differs by 2.8 standard deviations from the previously accepted value of -75.8(9.4), and the paper argues that the all-order and Zalpha-expansion approaches are no longer consistent, with implications for the 1S-2S hydrogen transition and the Rydberg constant.","tokens_in":11034,"tokens_out":6455,"duration_ms":60692,"significance":"If the hydrogen extrapolation is robust, this is an important result: it would improve the two-loop self-energy contribution to the hydrogen Lamb shift, change the theoretical prediction for the 1S-2S transition, and shift the Rydberg constant by about one standard deviation. The numerical all-order data themselves are a substantial advance, with clear improvements in accuracy and internal cross-checks between different computational schemes. However, as detailed below, the headline hydrogen value and the claimed inconsistency depend sensitively on unverified Zalpha-expansion coefficients, so the central claim is not yet established.","major_comments":[{"comment":"The extracted G60 values in Table I and the extrapolated hydrogen value G60(1)=-104.1(3.8) are obtained after subtracting the logarithmic terms B63 L^3 + B62 L^2 + B61 L, where B62 and B61 are taken from Pachucki (2001) and are described in the text as \"yet to be independently verified.\" Since L(Z=1)=ln((Zalpha)^-2) is about 9.8 and L^2 about 96.5, a change of about 3 in B61 or about 0.3 in B62 shifts G60(1) by roughly 28, which is comparable to the entire discrepancy with the accepted value -75.8(9.4). The uncertainty quoted in Eq. (7) accounts only for the constraints in Eq. (6) and polynomial extrapolation scatter, not for this dependence. The authors' own statement that \"a 10% adjustment in the B61 coefficient ... could restore the consistency\" confirms that the central 2.8-sigma discrepancy is not robust to plausible variation of unverified input coefficients. The paper should propagate the uncertainty of B61 and B62 into G60(1), or provide an explicit sensitivity analysis, before claiming a discrepancy.","section":"Main text, Eq. (4) and Eq. (7)"},{"comment":"The extrapolation to Z=1 is not a direct all-order result. It uses only data at Z=8,...,50, with Z=5,6,7 excluded, and fits a polynomial to G'_60 after subtracting the B72, B71, and B84 terms using analytic values from the Zalpha expansion. The constraints for B71, B83, B82, and B81 in Eq. (6) come from Ref. [16], the same Zalpha-expansion framework whose consistency with the all-order results is the central claim. The text acknowledges that \"the current numerical data cannot constrain the logarithmic coefficients.\" Thus the claimed inconsistency is partly a joint test of the all-order calculation and these specific analytic inputs, not a pure all-order result. The extrapolation uncertainty should include the uncertainty of these inputs without presupposing their validity, or the conclusion should be reframed accordingly.","section":"Main text, Eq. (5)-(6); Supplemental Material, 'Details of the extrapolation procedure'"},{"comment":"The exclusion of the three lowest-Z data points (Z=5,6,7) and the choice of weights w_i^2=delta_i^2+(a Z_i^{n+1})^2, with the parameter a selected by requiring that variations from removing data points \"are small and approximately equal,\" are heuristic. These excluded points are the ones closest to hydrogen and the only direct all-order anchor near the extrapolation region. The reported polynomial-extrapolation uncertainty is estimated as twice the maximal difference among P2, P3, and P4 fits, which does not account for model-selection uncertainty beyond these three polynomials. A more systematic assessment, such as exploring dependence on the excluded points, the weight parameter a, and higher-degree polynomials, would strengthen the reliability of the extrapolated G60(1).","section":"Supplemental Material, Eqs. (22)-(23)"}],"minor_comments":[{"comment":"The phrase \"break the previously assumed consistency\" is stronger than the evidence supports; consider \"challenge\" or \"question\" in light of the caveats discussed in the major comments.","section":"Summary and Conclusion"},{"comment":"The text says formulas and figures of the main text will be referred to as \"Eq. (?M)\" and \"Fig. ?M\"; these placeholders should be resolved before publication.","section":"Supplemental Material, first paragraph"},{"comment":"The ellipsis in Eq. (11) hides the right-hand side of Eq. (9); the notation is understandable from context but could be made explicit.","section":"Main text, Eq. (11)"},{"comment":"The caption does not identify the shaded region or the meaning of the triangle; the text should define these elements clearly.","section":"Main text, Fig. 4 caption"},{"comment":"The numerical values of B62 and B61 are not given; adding them would help readers assess the sensitivity discussed in the major comments.","section":"Main text, after Eq. (4)"}],"recommendation":"major_revision","confidential_remarks":"The all-order numerical work appears to be a genuine technical advance and deserves publication after the central claim is made robust. My main concern is that the headline hydrogen value and the claimed 2.8-sigma inconsistency depend on unverified B61/B62 coefficients and on analytic logarithmic constraints from the very Zalpha-expansion framework being questioned. A sensitivity analysis and softened conclusions are needed before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know two things about this paper before reading it: the all-order numerical results are a genuine step forward, and the headline hydrogen value is more fragile than the abstract suggests. The new subtraction scheme accelerates partial-wave convergence and lets the authors push all-order computations down to Z=5, with accuracy improved by up to a factor of 30. Table I is consistent with previous calculations where they overlap, and the internal cross-checks are good. That part is solid.\n\nThe soft spot is the extrapolation. To get G60(Z=1)=-104.1(3.8), the authors subtract the Zα-expansion terms B40, B50, B63, B62, and B61 from the all-order F(Zα). B62 and B61 come from Pachucki (2001) and have not been independently verified. The stress-test note is right: a change of about +3 in B61 shifts G60(1) by about +28, and the quoted ±3.8 uncertainty does not include this systematic. The paper itself acknowledges that a 10% adjustment in B61 could restore consistency with the Zα expansion. So the claimed 2.8σ discrepancy and the Rydberg shift are not robust to this unverified input.\n\nThere is also a mild circularity. To extrapolate, they use analytic constraints for the log coefficients (B71, B83, B82, B81) from the Zα expansion, subtract the leading log terms using those coefficients, then compare the result with the Zα expansion and claim inconsistency. The all-order data themselves are independent, so that is fine as far as it goes, but the hydrogen value and its uncertainty depend on the very framework whose consistency is being questioned.\n\nThe paper is honest about all of this—they call for independent checks of B62/B61 and B71, and they describe the fitting procedure in detail in the supplement. That counts for something. The method is important and the all-order data will be used by others. But the central physical claim, that the Rydberg constant shifts by one standard deviation, is not supported with the current uncertainty budget.\n\nI'd send this to a serious referee. It needs a careful reading of the supplementary material and the extrapolation. The right outcome might be a revision that either incorporates the B61/B62 systematic into the uncertainty or explicitly reframes the hydrogen result as preliminary.\n\nBest,\n[Name]","headline":"Solid all-order numerics, but the hydrogen extrapolation is hostage to unverified analytic coefficients, so the 2.8σ discrepancy is not yet credible.","tokens_in":11537,"tokens_out":3538,"would_cite":false,"duration_ms":32809,"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":"All-order two-loop self-energy calculations for hydrogen-like ions, extrapolated to hydrogen, give a value that differs from the accepted one by 2.8σ and shifts the Rydberg constant by one standard deviation.","keywords":["two-loop electron self-energy","Lamb shift","hydrogen","Rydberg constant","all-order Zα calculation","QED bound-state corrections","partial-wave expansion","subtraction scheme"],"falsifier":"A direct calculation of the missing part of the B60 coefficient (or an analytic evaluation of B71) that gives a value near the Zα-expansion prediction would shift the extrapolated G60(1) back toward the accepted value; alternatively, a nonperturbative two-loop self-energy calculation for Z = 4 or lower, performed by an independent group, would settle whether the deviation persists.","tokens_in":10437,"feed_emoji":"⚛️","tokens_out":5435,"duration_ms":41343,"temperature":0.7,"pith_summary":"This paper reports a calculation of the two-loop electron self-energy correction to the 1S Lamb shift of hydrogen-like ions performed to all orders in the binding parameter Zα, with numerical accuracy improved by more than an order of magnitude and calculations extended down to Z = 5. Extrapolating these all-order results to hydrogen gives G60(Z=1) = −104.1(3.8), a value twice as precise as the previously accepted one and different from it by 2.8 standard deviations. If this result is correct, the theoretical prediction for the 1S–2S transition frequency in hydrogen shifts by −2.5 kHz, lowering the Rydberg constant by about one standard deviation and breaking the previously assumed consistency between the all-order and Zα-expansion approaches.","feed_headline":"Two-loop self-energy shift lowers Rydberg constant by 1σ","feed_subtitle":"All-order results at Z = 5–50 disagree with the accepted hydrogen value at 2.8σ and break the assumed theory consistency.","key_machinery":"The enabling technique is an accelerated subtraction scheme for the two-loop self-energy diagrams, generalizing the one-loop method of Ref. [18] to the overlapping and nested diagrams. By subtracting and re-adding diagrams with two explicit interactions with the nuclear binding field, and computing the subtracted contributions in closed form in momentum space without partial-wave expansion, the authors improve the partial-wave convergence by one to two orders of magnitude. This permits calculations down to Z = 5 and reduces the numerical uncertainty at, for example, Z = 10 by a factor of 30.","core_discovery":"The central discovery is that the higher-order remainder function G60(Zα) of the two-loop self-energy, extracted from all-order numerical calculations for nuclear charges Z = 5 to 50, lies outside the band predicted by the Zα expansion and extrapolates to −104.1(3.8) at Z = 1. This disagrees with the previously recommended value of −75.8(9.4), obtained by combining Zα-expansion and older all-order results, by 2.8σ. The authors argue that the improved numerical precision challenges the assumed consistency between the nonperturbative and Zα-expansion approaches, so the theoretical hydrogen Lamb shift must be reexamined, with the updated SESE contribution shifting the 1S–2S prediction by −2.5 kHz and decreasing the Rydberg constant by one standard deviation.","pith_inferences":["If the disagreement is real, one of the logarithmic coefficients constrained analytically (e.g., B71) or the partially computed coefficient B60 must be substantially modified; the paper estimates that a missing contribution δB60 ≈ −40 or a 10% adjustment in B61 would reconcile the two approaches.","A fully independent nonperturbative calculation at Z = 4–7, where the present data are sparse and extrapolation carries the largest weight, would provide a sharp test of the claimed hydrogen value.","The same accelerated subtraction scheme may reduce the numerical cost of other two-loop bound-state QED calculations, including the g-factor corrections needed for ongoing experiments."],"forward_implications":["The theoretical prediction for the hydrogen 1S–2S transition frequency shifts by −2.5 kHz, changing the value of the Rydberg constant by about one standard deviation.","The previously recommended SESE value for hydrogen, G60(1) = −75.8(9.4), is replaced by −104.1(3.8), indicating that the consistency between all-order and Zα-expansion approaches no longer holds.","Improved accuracy for the 1S state also refines other nS energy predictions through the relation n^3 E(nS) = ν_n + E(1S).","The method opens the way to improved calculations of other two-loop QED corrections to the Lamb shift and to the bound-electron g factor."],"supporting_citations":[{"why":"Supplies the one-loop subtraction scheme that the present work generalizes to two loops.","marker":"[18]"},{"why":"Previous all-order two-loop self-energy results for Z ≥ 10, superseded and extended here.","marker":"[11]"},{"why":"Earlier all-order values at Z = 30, 40, 50, used for comparison.","marker":"[27]"},{"why":"Provides the Zα-expansion form of G60 and the analytic constraints on logarithmic coefficients used in the extrapolation.","marker":"[16]"},{"why":"Partial analytic result for B60, compared with the extrapolated value at Z = 0.","marker":"[15]"},{"why":"Derivation of the mα^2(Zα)^6 ln(Zα)^−2 coefficient, whose modification is one proposed reconciliation.","marker":"[8]"},{"why":"The previously accepted hydrogen SESE value that the new result disagrees with by 2.8σ.","marker":"[4]"},{"why":"Experimental 1S–2S transition frequency used to derive the updated Rydberg constant.","marker":"[1]"}],"fun_headline_variants":["Two-loop self-energy rewrites hydrogen's Rydberg constant","New two-loop result shifts Rydberg constant by 1σ","Hydrogen's Rydberg constant nudged by new self-energy","2.8σ tension in hydrogen's two-loop self-energy","Rydberg constant drops 1σ from two-loop QED"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extrapolation from the calculated nuclear charges Z ≥ 8 to hydrogen assumes the analytic constraints on the logarithmic coefficients B71, B83, B82, B81 of the Zα expansion; the numerical data alone cannot determine these logarithms, so the hydrogen value rests on the very Zα-expansion input whose consistency the paper questions.","fun_headline_variants_meta":{"raw":{"variants":["Two-loop self-energy rewrites hydrogen's Rydberg constant","New two-loop result shifts Rydberg constant by 1σ","Hydrogen's Rydberg constant nudged by new self-energy","2.8σ tension in hydrogen's two-loop self-energy","Rydberg constant drops 1σ from two-loop QED"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000164,"raw_usage":{"total_tokens":1206,"prompt_tokens":866,"completion_tokens":340,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":482,"completion_tokens_details":{"reasoning_tokens":251}},"tokens_in":482,"tokens_out":340,"duration_ms":3405,"temperature":1.0,"reasoning_tokens":251,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:29:26.315038+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct calculation of the missing part of the B60 coefficient (or an analytic evaluation of B71) that gives a value near the Zα-expansion prediction would shift the extrapolated G60(1) back toward the accepted value; alternatively, a nonperturbative two-loop self-energy calculation for Z = 4 or lower, performed by an independent group, would settle whether the deviation persists.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous all-order two-loop self-energy results for Z ≥ 10, superseded and extended here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier all-order values at Z = 30, 40, 50, used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Zα-expansion form of G60 and the analytic constraints on logarithmic coefficients used in the extrapolation."},{"cited_title":"Pachucki and U","cited_arxiv_id":null,"evidence_quote":"Partial analytic result for B60, compared with the extrapolated value at Z = 0."},{"cited_title":"Matveev, C","cited_arxiv_id":null,"evidence_quote":"Experimental 1S–2S transition frequency used to derive the updated Rydberg constant."}],"review_version":1}