{"id":"063db235-f28b-45e7-a780-506ae80617c6","arxiv_id":"2608.07010","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The electron-capture branching ratio of 76As into the first excited state of 76Ge is measured as (0.0572 ± 0.0029 ± 0.0074) percent, the first result with a full uncertainty budget.","lead":"This paper reports a new measurement of how often arsenic-76 atoms capture an electron to reach an excited state of germanium-76, with a result of 0.0572 percent. The value is about twice the only earlier estimate, so it matters for calculations used in the search for neutrinoless double beta decay.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SDD live-time correction is internally inconsistent: the stated t_live=144.8 h and dead-time range 73-79% imply K_t-dep=5.22, but Table III lists 5.62; the quoted branching ratio scales linearly with this factor.","rationale":"The reader's weakest assumption focused on the live-time estimate being unbiased after the digitizer reset-signal correction, validated only by a 4.2% deviation. My stress-test agrees that the SDD live-time correction is the most load-bearing part of the argument, but it identifies a sharper and more concrete problem: an internal numerical inconsistency between the reported t_live, the tabulated K_t-dep, and the stated dead-time range. This is not merely a question of systematic uncertainty; it is a consistency failure in the primary correction factor that scales the result. If K_t-dep = 5.62 is what was used, then t_live is 134.6 h, not 144.8 h, and the average dead time exceeds the stated maximum of 79%. If t_live = 144.8 h is correct, the branching ratio should be about 7% lower. The paper's quoted uncertainty budget does not absorb this discrepancy because it is a central-value shift, not an added variance. The most likely explanation is a typographical error or an ambiguity in how K_t-dep is defined (e.g., a count-weighted average rather than the simple time ratio), but the text and table as written are mutually incompatible. This reinforces the reader's CONDITIONAL verdict rather than overturning it: the measurement is plausible and carefully cross-checked, but the key correction factor must be clarified or independently reproduced before the result can be treated as definitive. I therefore recommend no change to the CONDITIONAL verdict, with the concrete check above as the requirement for acceptance.","tokens_in":10811,"tokens_out":21516,"duration_ms":229219,"concrete_test":"Recompute K_t-dep for the full data set from the per-hour live-time model used in the analysis and compare to Table III. Specifically, verify whether the sum of per-hour live times equals 134.6 h (as K=5.62 implies), 144.8 h (as stated in the text), or ~139 h (as row-wise K values imply), and confirm whether the model's dead-time range actually peaks at 79%. Then recompute ν_EC* with the corrected value; if the result shifts by ~7%, the quoted 0.0572% is not a stable central value until the discrepancy is explained.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central value ν_EC* = 0.0572% depends linearly on the rate-dependent correction K_t-dep = k_RCS * t_real/t_live,SDD. For the full data set, the text gives t_real = 741.8 h, t_live,SDD = 144.8 h, and k_RCS = 1.020, which yields K_t-dep = 1.020 * 741.8 / 144.8 = 5.22. Table III, however, lists K_t-dep = 5.62 for the same data set. Using K_t-dep = 5.62 implies t_live,SDD = 1.020 * 741.8 / 5.62 = 134.6 h, a 7% difference. Moreover, the text states that the dead time ranges from 79% (highest rates) to 73% (lowest rates); both t_live = 144.8 h and t_live = 134.6 h correspond to average dead times of 80.5% and 81.8%, respectively, above the stated maximum. The row-wise K_t-dep values in Table III, when combined with k_RCS = 1.020, imply Σt_live ≈ 139 h, again inconsistent with both 144.8 h and the dead-time range. Because ν_EC* is directly proportional to K_t-dep, the quoted result would be 0.0572% with K_t-dep = 5.62 but roughly 0.053% with K_t-dep = 5.22. This unresolved arithmetic inconsistency means the largest rate-dependent correction in the measurement is not self-consistently defined, and the central claim is not stable until it is resolved.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a new measurement of the electron-capture branch of 76As into the first excited state of 76Ge (EC*). 76As is produced by 75As(n,γ) in an activated As2O3 sample, and the EC* branch is tagged by the coincidence of the 562.9 keV γ-ray in a HPGe detector and the Ge Kα X-ray in a silicon drift detector. The number of EC* decays is obtained with a two-dimensional sideband background subtraction, and the total number of 76As decays is obtained from the 559.1 keV γ-ray line. The result is ν_EC* = (0.0572 ± 0.0029(stat.) ± 0.0074(syst.))%, which is about 70% larger than the only previous measurement of ≈0.027%. The paper also proposes several future approaches to measure the unobserved EC into the 76Ge ground state.","tokens_in":14,"tokens_out":3168,"duration_ms":99867,"significance":"If correct, the result provides a valuable experimental benchmark for nuclear-structure calculations relevant to 0νββ matrix elements, and it is the first measurement of this branch with a quantified uncertainty budget. The experimental work is careful in several respects: the 2D sideband method is tested for separability, the result is cross-checked with one-dimensional and two-dimensional fits, partition consistency is demonstrated, and time-dependent rate effects are investigated. The claimed log ft = 6.99 is a concrete physics output. However, the central value rests on a rate-dependent live-time correction whose numerical consistency is not established, and the factor-of-two discrepancy with the previous measurement is not discussed. These issues must be resolved before the measurement can be considered reliable.","major_comments":[{"comment":"The rate-dependent correction factor K_t-dep is internally inconsistent. The text gives t_real = 741.8 h, t_live,SDD = 144.8 h, and k_RCS = 1.020, which yields K_t-dep = 1.020 × 741.8 / 144.8 = 5.22. Table III lists K_t-dep = 5.62 for the same total dataset, implying t_live,SDD = 1.020 × 741.8 / 5.62 ≈ 134.6 h, a 7% difference. Because ν_EC* in Eq. (3) is directly proportional to K_t-dep, the quoted central value is not stable: using 5.22 instead of 5.62 would change the result from 0.0572% to about 0.053%, a shift comparable to the systematic uncertainty. In addition, the stated dead-time range of 79% to 73% is inconsistent with either value: t_live = 144.8 h corresponds to an average dead time of 80.5%, and t_live = 134.6 h to 81.8%. The row-wise K_t-dep values in Table III, combined with k_RCS = 1.020, imply a total live time of about 139 h, again inconsistent with both 144.8 h and the stated range. This arithmetic inconsistency in the largest rate-dependent correction must be resolved before the branching ratio can be quoted.","section":"Branching ratio calculation and Table III"},{"comment":"The live-time correction is validated only by a 4.2% deviation between the rate- and energy-dependent estimator and test experiments with known live times. The dead time is 73–79%, so the correction is large, and the text states that a rate-independent bias would cancel. However, the discrepancy found in Table III (7%) is larger than the 4.2% validation uncertainty and shows that the internal consistency of the correction is not established. The paper should demonstrate, with a first-principles model or an independent measurement, that the correction factor is accurate to the claimed level; otherwise the 2% contribution to the systematic uncertainty is not supported.","section":"Branching ratio calculation"},{"comment":"The new result is about 70% larger than the only previous measurement of ν_EC* ≈ 0.027% [9]. The paper states this difference but does not discuss possible sources of the discrepancy, such as different background treatment, normalization, or efficiency calibrations. A quantitative comparison of the two measurements, including the uncertainties and any known systematic differences, is needed to substantiate the claim that the new measurement supersedes the previous one. Without such a discussion, the physical conclusion rests on a single measurement that is in tension with the earlier value.","section":"Conclusion / comparison with previous measurement"}],"minor_comments":[{"comment":"The text says the rate-dependence check uses data summarized into eight quantiles, while the caption of Figure 8 says ten quantiles; this should be corrected.","section":"Cross checks"},{"comment":"In Table I, t_real is given as 741.8 h, whereas the total row of Table III gives 742 h; the difference is presumably due to rounding, but the two tables should be made consistent.","section":"Branching ratio calculation"},{"comment":"The word 'Inlet' in the caption of Figure 5 should be 'Inset'.","section":"Figure 5"},{"comment":"The notation in Eq. (4) mixes εHPGe(559.1 keV) with the definition of ε_ref; the text should clarify that ε_ref is the full-energy efficiency at 559.1 keV and that the ratio εHPGe(559.1)/εHPGe(562.9) is the one used.","section":"Equation (4)"},{"comment":"The paper states that the EC into the 76Ge ground state has not been observed yet, but the later section on future measurements makes clear that this remains unmeasured; the distinction between EC* and EC0 should be stated more prominently in the abstract or introduction to avoid confusion about what was measured.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The central claim is a new measurement with a carefully described analysis, but the live-time correction inconsistency is a load-bearing issue that must be fixed. The previous measurement [9] is cited with overlapping authorship (ref. [10] includes a coauthor of this manuscript); this is not a circularity problem for the central result, but the manuscript should acknowledge that two input nuclear data quantities come from references with common authorship. The claim of 'first full uncertainty budget' is reasonable if the live-time correction is resolved, but it should be phrased carefully relative to the earlier measurement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper before citing it: it is a serious experimental effort with a real new number, but there is an arithmetic inconsistency in the largest rate-dependent correction that the authors need to fix before the result is treated as final.\n\nThe genuinely new thing is the uncertainty budget. The prior measurement of the 76As EC* branch (ref. [9]) gave ~0.027% with no full systematic treatment. This paper measures (0.0572 ± 0.0029(stat.) ± 0.0074(syst.))%, about twice the old value, and lays out a 2D sideband analysis, rate-dependence checks, partition consistency checks, and a systematic budget dominated by the 12% simulated X-ray efficiency uncertainty. The method is sensible and the internal cross-checks shown in Figs. 7 and 8 are appropriate. On the face of it, this is a within-subfield benchmark measurement for NME calculations in the 76Ge 0νββ program.\n\nThe soft spot is the dead-time correction. The stress-test note is right: for the full dataset, the text gives t_real = 741.8 h, t_live,SDD = 144.8 h, k_RCS = 1.020, which gives K_t-dep = 5.22, but Table III lists K_t-dep = 5.62. The same table's run-wise K_t-dep values plus k_RCS imply a summed live time near 139 h, not 144.8 h. The text also says dead time ranges from 73% to 79%, but both 144.8 h and 134.6 h correspond to average dead times above 80%. The branching ratio is directly proportional to K_t-dep, so the result shifts from 0.0572% to roughly 0.053% depending on which value is used. That is a 7% relative shift, comparable to the quoted statistical uncertainty and not negligible against the 13% systematic. The authors do state a 4.2% validation of their live-time model from test experiments, and a rate-independent bias would cancel, but the internal inconsistency means the correction is not self-consistently defined as written. They need to reconcile the table with the text, or show where the extra 7% comes from.\n\nMinor points: the factor-of-two disagreement with the previous measurement is not quantitatively addressed beyond noting it, and no data or simulation code are released, so the dominant efficiency systematic cannot be independently checked. The citation pattern looks fine; two inputs come from overlapping groups but the central claim is an independent measurement.\n\nBottom line: this deserves a serious referee, but the live-time inconsistency should be resolved before the value enters the literature as definitive. I would not cite the central number yet, though I would cite the paper as the current best-effort measurement once the authors clarify the correction.","headline":"Careful second measurement of a small EC branch in 76As, with a genuinely quantified uncertainty budget, but an internal inconsistency in the dead-time correction means the central value is not yet stable.","tokens_in":12,"tokens_out":700,"would_cite":false,"duration_ms":67615,"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 reports the first measurement with a fully quantified uncertainty budget of the electron-capture branching ratio of 76As into the first excited state of 76Ge, finding (0.0572 ± 0.0029 (stat.) ± 0.0074 (syst.))%.","keywords":["electron capture","branching ratio","76As","76Ge","neutrinoless double beta decay","nuclear matrix element","coincidence counting","silicon drift detector"],"falsifier":"Independently re-measure the branching ratio with a data acquisition system that records the full reset pulse (or with a different X-ray detector), so the dead time is made negligible; if the resulting branching ratio differs from 0.0572% by more than the combined uncertainties, the central claim would be refuted. A cheaper check is to compute the live time from a first-principles simulation of the reset preamplifier and see whether it reproduces the 4.2% test-experiment deviation.","tokens_in":10620,"feed_emoji":"⚛️","tokens_out":5244,"duration_ms":47897,"temperature":0.7,"pith_summary":"The paper measures how often 76As decays by electron capture into the first excited state of 76Ge, a decay branch relevant for nuclear matrix element calculations in the search for neutrinoless double beta decay of 76Ge. Using a silicon drift detector for the characteristic Ge X-ray and a high-purity germanium detector for the 562.9 keV gamma ray, the authors find a branching ratio of 0.0572% with, for the first time, a complete uncertainty budget. The result is about 70% larger than the only previous measurement of about 0.027%, and it sits above the ~0.03% theory prediction. If correct, this strengthens the experimental benchmarks used to test nuclear structure models that feed into neutrinoless double beta decay matrix element calculations.","feed_headline":"Electron-capture branch of 76As measured at 0.0572%","feed_subtitle":"First fully quantified result, ~70% above the previous value, sharpening the 76Ge double-beta-decay benchmarks.","key_machinery":"The measurement is carried by a relative coincidence method: the number of EC* decays is obtained from coincident 562.9 keV gamma-ray (HPGe) and 9.9 keV Ge K-alpha X-ray (silicon drift detector) events, normalized to the 559.1 keV gamma-ray reference line that counts 76As decays. Background in the coincidence region is subtracted by a two-dimensional sideband method that assumes the background energy distribution factorizes into independent gamma-ray and X-ray components, and the X-ray detector's large (73–79%) rate- and energy-dependent dead time, caused by a digitizer dynamic-range configuration error that truncates the preamplifier reset signal, is corrected in one-hour time bins.","core_discovery":"The central claim is that the branching ratio ν_EC* of the electron capture of 76As into the first excited state of 76Ge is (0.0572 ± 0.0029 (stat.) ± 0.0074 (syst.))%, measured via coincident detection of the 562.9 keV gamma ray and the 9.9 keV Ge K-alpha X-ray, with a two-dimensional sideband subtraction of a background that includes both random and true beta-decay coincidences. The paper states this is the first measurement with the full uncertainty budget quantified, and that the result translates to a log ft value of 6.99. The measured branch is about 70% larger than the previous measurement of ≈0.027%.","pith_inferences":["If the result holds up, the factor-of-two discrepancy with the 2014 measurement suggests a systematic effect in one of the two experiments; a third, independent measurement would be needed to decide which one is biased.","The paper's dead-time correction could be replaced by a fully first-principles model of the digitizer/preamplifier interaction, which would likely reduce the dominant systematic uncertainty and make a future EC0 measurement more precise.","The measured value near the upper edge of theory predictions may motivate updated nuclear-structure calculations for the neighboring EC0 branch, since the two branches share phase-space and nuclear-structure inputs."],"forward_implications":["The measured log ft = 6.99 replaces the previous value of 7.34 as the benchmark for the 2− to 2+ first-forbidden non-unique electron-capture transition.","A ~70% larger EC* branch alters the expected 562.9 keV and Ge X-ray yields in 76Ge detectors, a relevant input for background modeling in future LEGEND operations.","With the EC* branch quantified, the remaining unmeasured EC0 branch can be attacked with a 4π scintillator veto detector, as the paper proposes.","The two-dimensional sideband method and the count-rate quantile consistency checks establish a template for coincidence measurements at high dead time."],"supporting_citations":[{"why":"Previous measurement of the EC* branch (≈0.027%) that this work updates and exceeds by about 70%.","marker":"[9]"},{"why":"Supplies the pK capture probability and the log ft conversion for the electron-capture transition.","marker":"[10]"},{"why":"Provides the Ge fluorescence yield ωK and the Se X-ray emission probabilities used in the analysis.","marker":"[14]"},{"why":"Gives the emission probability of the 559.1 keV reference gamma ray used to normalize the number of 76As decays.","marker":"[15]"},{"why":"Supplies the K-alpha emission probability pα used to correct for Ge X-ray line branching.","marker":"[16]"},{"why":"Geant4 simulation used to obtain the coincidence and reference detection efficiencies, including self-absorption and true coincidence summing.","marker":"[17]"}],"fun_headline_variants":["First fully quantified 76As EC branch: 0.0572%","76As electron capture branch 70% above previous measurement","Coincident detection nails 76As EC branch at 0.0572%","New 76As EC branch sharpens 76Ge double-beta decay benchmarks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on the assumption that the rate- and energy-dependent live time of the X-ray detector is estimated without bias after correcting the truncated preamplifier reset signal, even though that estimate was validated only by a 4.2% deviation in test experiments rather than a first-principles model.","fun_headline_variants_meta":{"raw":{"variants":["First fully quantified 76As EC branch: 0.0572%","76As electron capture branch 70% above previous measurement","Coincident detection nails 76As EC branch at 0.0572%","New 76As EC branch sharpens 76Ge double-beta decay benchmarks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000232,"raw_usage":{"total_tokens":1519,"prompt_tokens":1004,"completion_tokens":515,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":620,"completion_tokens_details":{"reasoning_tokens":434}},"tokens_in":620,"tokens_out":515,"duration_ms":4975,"temperature":1.0,"reasoning_tokens":434,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:38:46.944313+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Independently re-measure the branching ratio with a data acquisition system that records the full reset pulse (or with a different X-ray detector), so the dead time is made negligible; if the resulting branching ratio differs from 0.0572% by more than the combined uncertainties, the central claim would be refuted. A cheaper check is to compute the live time from a first-principles simulation of the reset preamplifier and see whether it reproduces the 4.2% test-experiment deviation.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous measurement of the EC* branch (≈0.027%) that this work updates and exceeds by about 70%."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the pK capture probability and the log ft conversion for the electron-capture transition."},{"cited_title":"Rodriguez, A","cited_arxiv_id":null,"evidence_quote":"Provides the Ge fluorescence yield ωK and the Se X-ray emission probabilities used in the analysis."},{"cited_title":"Marnada, H","cited_arxiv_id":null,"evidence_quote":"Gives the emission probability of the 559.1 keV reference gamma ray used to normalize the number of 76As decays."},{"cited_title":"Sch¨ onfeld, and H","cited_arxiv_id":null,"evidence_quote":"Supplies the K-alpha emission probability pα used to correct for Ge X-ray line branching."},{"cited_title":"The im- plementation of the detectors was optimized to match calibration measurements with point-like radiation stan- dards","cited_arxiv_id":null,"evidence_quote":"Geant4 simulation used to obtain the coincidence and reference detection efficiencies, including self-absorption and true coincidence summing."}],"review_version":1}