{"id":"85b971c6-85fc-4585-bb2f-e892352949c6","arxiv_id":"2506.16955","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"Using a delayed-coincidence search in 103.7 kg·yr of GERDA Phase II data, the collaboration set a 90% CL upper limit of <0.216 nuc/(kg·yr) on 77Ge production, corresponding to <0.38 nuc/(kg·yr) for 77Ge and 77mGe together.","lead":"The GERDA collaboration searched for a rare isotope, 77Ge, that cosmic rays can create inside their germanium detectors and that could mimic the signal they hunt for in neutrinoless double-beta decay. They found no such events, set a new upper limit on its production, and used the result to show the next-generation experiment LEGEND-1000 will keep this background under control.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Abstract's 'equal production rates' conversion is inconsistent with the body's ε_g=59.5% model; the stated <0.38 limit follows from the model, not from equal rates.","rationale":"The reader's verdict is CONDITIONAL, and I agree that the paper should not be accepted without addressing a clear issue. The most load-bearing concern is not the Monte Carlo efficiency systematic (which the reader identified as the weakest assumption), but the direct contradiction between the abstract and the body regarding how the total production rate is derived. This is an internal inconsistency in the presentation of the central result, not a matter of simulation accuracy. The efficiency systematics, while real and understated by the <0.5% statistical-only uncertainty, would only shift an upper limit by a moderate factor and would not change the qualitative conclusion (no signal observed, production rate limited to a level comparable to the prediction). The abstract/body discrepancy, in contrast, makes the headline number falsely attributed to an assumption that was not used; this requires correction before the paper can be considered accurate. I therefore maintain the CONDITIONAL verdict, targeting the abstract wording rather than the analysis itself. The concrete test would settle whether the inconsistency is as stated by recomputing both conversions; if the abstract's number were instead reproduced under equal rates (which it is not), the concern would be void.","tokens_in":13900,"tokens_out":7753,"duration_ms":82680,"concrete_test":"Recompute the total 77(m)Ge production-rate upper limit from the stated 90% CL limit on ground-state decays (N_77Ge=22.4) using (a) ε_g=0.50 (equal production rates, as the abstract claims) and (b) ε_g=0.595 (the body's model). Verify that the abstract's <0.38 corresponds only to (b), and check whether the abstract can be revised to explicitly reference the model-based ground-state fraction, or whether the number should be changed to <0.43 to match the stated assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central number in the abstract, a total 77(m)Ge production rate of <0.38 nuc/(kg·yr), is presented as derived 'assuming equal production rates'. However, Section 3 derives this value by converting the 77Ge ground-state production limit (<0.216 nuc/(kg·yr)) using the neutron-capture model fraction ε_g = ε_d + (1-ε_d)·ε_IT = (59.5±8.1)% (with ε_d = 50±10% and ε_IT = 19±2%). A genuine 'equal production rates' assumption would set the ground-state fraction to 50%, yielding a total limit of <0.216/0.5 = <0.43 nuc/(kg·yr), not <0.38. The quoted 0.38 arises from ε_g=0.595, contradicting the abstract's stated assumption. This is not a cosmetic issue: it directly misstates the basis of the paper's headline result and would mislead a reader relying on the abstract. The body text is internally consistent, so the fix is straightforward (either correct the abstract wording or recompute the number), but the inconsistency must be resolved before publication for the central claim to be communicated accurately.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a search for in-situ production of 77Ge in the full GERDA Phase II data set (103.7 kg·yr) using a delayed coincidence method that exploits the 114 µs isomeric state of 77As. No candidate events are found (N_obs = 0), with an expected random-coincidence background of N_rc = 0.04, leading to a Feldman-Cousins upper limit of <2.4 counts and a 77Ge production rate of <0.216 nuc/(kg·yr) at 90% CL. Using a neutron-capture model with a ground-state population fraction ε_g = (59.5±8.1)%, the total 77(m)Ge production rate is derived as <0.38 nuc/(kg·yr). A Bayesian update of a previous Monte Carlo prediction gives a posterior central value of 0.18 nuc/(kg·yr). The paper further estimates that the 77(m)Ge background contribution in LEGEND-1000 would be sub-dominant, below 10^-6 cts/(keV·kg·yr).","tokens_in":14060,"tokens_out":7088,"duration_ms":66592,"significance":"If the result holds, it is the strongest experimental constraint on in-situ 77(m)Ge production at LNGS and an important validation of cosmogenic background models for LEGEND-1000. The analysis uses the full GERDA Phase II exposure, a novel digital signal processing approach for pile-up reconstruction, and a carefully described delayed-coincidence selection with cross-checks against calibration data. The null observation, the internal consistency of the limit derivation, and the explicit treatment of nuisance parameters are strengths. The main concerns are a misleading statement in the abstract and the absence of a systematic uncertainty for the simulation-based efficiencies; both are addressable.","major_comments":[{"comment":"The abstract states that the total 77(m)Ge production rate limit of <0.38 nuc/(kg·yr) is derived \"assuming equal production rates\". However, Section 3 derives this number by converting the 77Ge ground-state limit using ε_g = (ε_d + (1−ε_d)·ε_IT) = (59.5±8.1)%, with ε_d = (50±10)% and ε_IT = (19±2)%. A genuine equal-production assumption (ε_g = 0.5) would give a total limit of <0.216/0.5 = <0.43 nuc/(kg·yr), not <0.38. The quoted 0.38 follows from the neutron-capture model, not from equal production rates. This misstates the basis of the headline result and should be corrected in the abstract, for example by stating the adopted model assumption explicitly.","section":"Abstract vs. Section 3"},{"comment":"The uncertainty quoted for the selection efficiencies (<0.5%) is the statistical precision of the generated Monte Carlo samples, not a systematic uncertainty of the MaGe/GERDA detector simulation. The total efficiency directly enters the conversion from counts to production rate; an unquantified systematic bias in the simulated energy/multiplicity, time-difference, or pile-up acceptance would propagate directly into the quoted limit. The paper should either quantify systematic uncertainties of the simulation-based efficiencies (e.g., by varying simulation parameters or comparing with calibration data) or explicitly discuss the expected size of such effects and justify that they are negligible relative to the statistical precision of the null measurement.","section":"Section 2.3, Table 1"}],"minor_comments":[{"comment":"The sentence \"an upper limit on the production rate of was set at <0.216 nuc/(kg·yr)\" is missing the isotope; it should refer to 77Ge.","section":"Abstract"},{"comment":"The axis label \"dT [s]\" appears in both figures, but the plotted values (0–20 and 0–800) are in microseconds, not seconds. Please correct the units to µs.","section":"Figures 5 and 6"},{"comment":"The text states that the FWTM contains 96% of the peak area, while the caption of Figure 6 states 97%. These values should be reconciled.","section":"Section 2.3 and Figure 6 caption"},{"comment":"The x-axis label contains a typo: \"produdction\" should be \"production\".","section":"Figure 7"},{"comment":"The unit \"nuc\" (nuclei) is not defined at first use; please define it explicitly.","section":"Introduction"},{"comment":"The statement that the previous prediction is \"further corroborated\" should be qualified, since the 90% CL upper limit of 0.38 nuc/(kg·yr) is about a factor of two above the predicted central value of 0.21 nuc/(kg·yr); the result is consistent with but does not strongly confirm the prediction.","section":"Section 3"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a clean null result with a careful statistical treatment, and the central limit derivation is internally consistent. The main issues are a misleading statement in the abstract about the production-rate conversion and the lack of a systematic uncertainty on the simulation-based efficiencies. Both are addressable within the manuscript's scope, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, the headline: this is a competent, genuinely useful background measurement from GERDA Phase II, with one real presentation flaw in the abstract that should be fixed before publication. The search is new for GERDA—first delayed-coincidence search for 77Ge via 77mAs using the full 103.7 kg·yr exposure—and the authors built a dedicated pile-up DSP to recover events the standard analysis discards. Zero candidates, expected random background 0.04, and a Feldman-Cousins limit of <0.216 nuc/(kg·yr) for 77Ge, about a factor of 20 better than the old thesis limit. That is the real result.\n\nThe conversion to a total 77(m)Ge rate uses a neutron-capture model with ε_g = (59.5±8.1)% to get <0.38 nuc/(kg·yr). The abstract, however, says this total was obtained 'assuming equal production rates.' Equal rates would give 0.216/0.5 = 0.43, not 0.38. The body is internally consistent; the abstract is worded wrong. That's a correction, not a fatal flaw.\n\nThe efficiency chain (ε_em=34.5%, ε_dT=99.3%, ε_pile-up=94.9%) is transparent and cross-checked against calibration data, but the quoted <0.5% uncertainty is only the Monte Carlo statistical precision; systematic accuracy of the MaGe/Geant4 efficiencies is not quantified. A null result at this level means the conclusion is unlikely to change, but a referee should ask for a sentence on that. The LEGEND-1000 extrapolation leans on a Zenodo poster rather than a paper, and the final background-index estimate in the appendix is a back-of-envelope calculation. Minor, because the claim is only that the cosmogenic background stays sub-dominant, which holds even with a factor of two wiggle.\n\nCitation pattern is fine. The previous Monte Carlo prediction is the same collaboration, but the limit here comes from data alone; the Bayesian update is openly stated and not the main claim. The one model input that matters, ε_d = 50±10%, is treated honestly as a nuisance with uncertainty.\n\nWho is this for? Anyone planning LEGEND-1000 background budgets, and anyone doing delayed-coincidence or pile-up work in HPGe. It deserves a serious referee and publication after the abstract fix. If I were editor, I'd send it out with a note to correct the wording and to state explicitly that the efficiency uncertainty is statistical only.","headline":"Solid GERDA background measurement with a wrong abstract phrase; worth reviewing after a small fix.","tokens_in":15333,"tokens_out":3385,"would_cite":true,"duration_ms":29943,"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":"The full GERDA Phase II data set shows zero 77Ge delayed-coincidence candidates, bounding the in-situ production rate below 0.216 nuclei per kilogram-year at 90% CL.","keywords":["neutrinoless double-beta decay","germanium-77","cosmogenic background","delayed coincidence","isomeric state","pile-up signal processing","GERDA","LEGEND-1000"],"falsifier":"Reanalysing the GERDA Phase II waveforms with a time-difference threshold below 4.5 µs and a delayed-energy acceptance extended into the continuum: if this finds a cluster of candidates whose time differences follow the 114 µs exponential of the $^{77\\mathrm{m}}$As isomer, the paper's $N_\\mathrm{obs}=0$ result and its quoted production-rate limit would be contradicted; if no such cluster appears, the null is strengthened.","tokens_in":13651,"feed_emoji":"⚛️","tokens_out":18451,"duration_ms":162104,"temperature":0.7,"pith_summary":"This paper searches for the in-situ production of $^{77}$Ge inside the GERDA experiment, where neutron capture on the double-$\\beta$-decay isotope $^{76}$Ge can create $^{77}$Ge and its isomer $^{77\\mathrm{m}}$Ge. Their $\\beta$ decays have $Q$-values above the $Q_{\\beta\\beta}$ value of 2039 keV, so they are a potential background that could mimic a signal in the neutrinoless double-$\\beta$ decay search. Using the full 103.7 kg·yr Phase II data set, the analysis looks for a delayed coincidence: the $\\beta$ decay of $^{77}$Ge populates a 475 keV isomeric state in $^{77}$As with a 114 µs half-life, whose subsequent gamma emission appears as a second pulse piled on the first. No such candidate is found, yielding an upper limit of $<0.216$ nuc/(kg·yr) on the $^{77}$Ge production rate and $<0.38$ nuc/(kg·yr) for the combined $^{77}$Ge and $^{77\\mathrm{m}}$Ge rate at 90% CL. The limit agrees with a Monte Carlo prediction of $(0.21\\pm0.07)$ nuc/(kg·yr) and supports the expectation that this cosmogenic background stays sub-dominant in the planned LEGEND-1000 experiment.","feed_headline":"GERDA sees zero 77Ge decays; production limit below 0.216 per kg·yr","feed_subtitle":"The null search confirms the simulated cosmogenic rate and keeps LEGEND-1000's background sub-dominant.","key_machinery":"The load-bearing mechanism is the delayed-coincidence signature of the $^{77}$Ge decay chain. Beta decay of $^{77}$Ge ($T_{1/2}=11.21$ h) populates the $9/2^+$ isomeric state of $^{77}$As at 475 keV ($T_{1/2}=114$ µs) in about 33% of decays, and the delayed electromagnetic de-excitation emits gammas of 211, 264, or 475 keV. Because the 114 µs half-life is much longer than the roughly 1.5 µs charge-collection time of the germanium detectors, the delayed pulse arrives after the prompt pulse in the same detector and appears as pile-up on the waveform's exponential tail. New digital signal processing applies trapezoidal filters of length matched to the trigger separation, extracts the heights of both pulses, reconstructs the time difference, and is sensitive down to 4.5 µs separation; Monte Carlo simulation then fixes the combined selection efficiency at 32.4%. A unified small-signal counting procedure converts the observed zero candidates into the upper limit, with the branching-ratio uncertainties treated as nuisance parameters.","core_discovery":"The central result is a null observation with a sharp bound. In the full GERDA Phase II exposure of 103.7 kg·yr, after applying the delayed-coincidence selection built from the $^{77}$Ge decay chain, the observed number of candidates is $N_\\mathrm{obs}=0$ against an expected random-coincidence background of $N_\\mathrm{rc}=0.04$. A unified small-signal upper-limit calculation gives an upper limit of $<2.4$ delayed-coincidence counts at 90% CL, which translates into a $^{77}$Ge production rate of $<0.216$ nuc/(kg·yr) and, assuming equal production of the ground and isomeric states and a ground-state population probability of $(59.5\\pm8.1)\\%$, a combined $^{77}$Ge and $^{77\\mathrm{m}}$Ge production rate of $<0.38$ nuc/(kg·yr). The paper reads this as corroboration of the earlier Monte Carlo estimate of $(0.21\\pm0.07)$ nuc/(kg·yr), and a Bayesian update centers the rate at $0.18$ nuc/(kg·yr) with a 1$\\sigma$ credibility interval of $[0.106, 0.251]$ nuc/(kg·yr). It further claims that tagging the $^{77\\mathrm{m}}$As isomer can reject about 62% of the $^{77}$Ge background events that survive other cuts, and that this keeps the in-situ cosmogenic background in LEGEND-1000 below $10^{-6}$ cts/(keV·kg·yr), at most 10% of the experiment's background budget.","pith_inferences":["An editorial inference: because the expected random-coincidence rate is only 0.04 counts, applying the $^{77\\mathrm{m}}$As tag as an active veto in LEGEND-200 costs almost no signal efficiency, so the same delayed-coincidence logic could be run online as a live background tag rather than offline.","An editorial inference: the factor-of-two headroom between the simulation and the limit is small next to the simulation's 35% systematic uncertainty, so the agreement is weaker than it might appear; a larger exposure is needed to say whether the true rate sits near the prior or near the limit.","An editorial inference: the eight observed continuum candidates with delayed energies between 200 and 500 keV do not follow the 114 µs lifetime, but a future analysis that includes the continuum rather than only the three gamma peaks could gain additional sensitivity."],"forward_implications":["The combined $^{77}$Ge and $^{77\\mathrm{m}}$Ge in-situ production rate at the GERDA site is bounded below $0.38$ nuc/(kg·yr) at 90% CL, about a factor of two above the Monte Carlo prediction and roughly an order of magnitude tighter than the earlier GERDA-based bound of $<4.1$ nuc/(kg·yr).","A Bayesian update of the simulation prior by the new likelihood centers the production rate at $0.18$ nuc/(kg·yr) with a 1$\\sigma$ credibility interval of $[0.106, 0.251]$ nuc/(kg·yr), a scaling factor of $0.85^{+0.35}_{-0.34}$ relative to the original estimate.","Tagging the $^{77\\mathrm{m}}$As delayed de-excitation rejects about 62% of the $^{77}$Ge background events that survive the other analysis cuts, lowering the $^{77}$Ge/$^{77\\mathrm{m}}$Ge background-index contribution to $(1.50\\pm0.07\\pm0.67)\\times10^{-6}$ cts/(keV·kg·yr).","For LEGEND-1000, the in-situ cosmogenic background contribution is estimated below $10^{-6}$ cts/(keV·kg·yr), no more than 10% of the target background index of $<10^{-5}$ cts/(keV·kg·yr).","After one year of LEGEND-200 running at its final detector mass, the sensitivity of this delayed-coincidence search doubles, so the limit would tighten by a factor of two or a signal could appear."],"supporting_citations":[{"why":"Final GERDA Phase II results—supply the 103.7 kg·yr exposure, the standard event selection, and the neutrinoless double-beta decay context.","marker":"[2]"},{"why":"A Monte Carlo prediction of the 77(m)Ge production rate of (0.21 ± 0.07) nuc/(kg·yr) and the background index that this analysis tests.","marker":"[4]"},{"why":"The LEGEND-1000 design report—defines the <10^-5 cts/(keV·kg·yr) background goal used in the extrapolation.","marker":"[5]"},{"why":"A comparable delayed-coincidence search in another germanium experiment, whose null result motivates and contextualises the analysis.","marker":"[6]"},{"why":"Decay-scheme data—fix the half-lives, Q-values, and the (33 ± 1)% branching ratio into the 77mAs isomer.","marker":"[7]"},{"why":"The Monte Carlo simulation of the GERDA detector response—sets the energy, multiplicity, and pile-up selection efficiencies.","marker":"[8]"},{"why":"The unified statistical procedure that converts N_obs = 0 and N_rc = 0.04 into the <2.4 count upper limit.","marker":"[12]"},{"why":"The production-rate estimate and neutron-moderator design used for the LEGEND-1000 background projection.","marker":"[17]"}],"fun_headline_variants":["GERDA finds zero 77Ge decays; limit <0.216 nuc/(kg·yr)","No 77Ge production seen in GERDA: rate <0.216 per kg·yr","GERDA null search confirms simulated 77Ge cosmogenic rate","Zero 77Ge events in GERDA bound background for LEGEND-1000","GERDA's 77Ge null result tightens limit to 0.216 nuc/(kg·yr)"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The limit assumes the Monte Carlo simulation of the germanium detectors correctly predicts how often a real $^{77}$Ge decay chain would survive the selection, because the quoted $<0.5\\%$ efficiency uncertainty covers only the statistical spread of the generated samples and not the full systematic accuracy of the detector simulation; if the simulated efficiency is too high, the production-rate limit is too strong.","fun_headline_variants_meta":{"raw":{"variants":["GERDA finds zero 77Ge decays; limit <0.216 nuc/(kg·yr)","No 77Ge production seen in GERDA: rate <0.216 per kg·yr","GERDA null search confirms simulated 77Ge cosmogenic rate","Zero 77Ge events in GERDA bound background for LEGEND-1000","GERDA's 77Ge null result tightens limit to 0.216 nuc/(kg·yr)"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001384,"raw_usage":{"total_tokens":5757,"prompt_tokens":1255,"completion_tokens":4502,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":871,"completion_tokens_details":{"reasoning_tokens":4387}},"tokens_in":871,"tokens_out":4502,"duration_ms":33295,"temperature":1.0,"reasoning_tokens":4387,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:15:36.637892+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reanalysing the GERDA Phase II waveforms with a time-difference threshold below 4.5 µs and a delayed-energy acceptance extended into the continuum: if this finds a cluster of candidates whose time differences follow the 114 µs exponential of the $^{77\\mathrm{m}}$As isomer, the paper's $N_\\mathrm{obs}=0$ result and its quoted production-rate limit would be contradicted; if no such cluster appears, the null is strengthened.","supporting_citations":[{"cited_title":"Agostiniet al.(GerdaCollab.), Final Results ofGerda on the Search for Neutrinoless Double-βDecay, Phys","cited_arxiv_id":null,"evidence_quote":"Final GERDA Phase II results—supply the 103.7 kg·yr exposure, the standard event selection, and the neutrinoless double-beta decay context."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"A comparable delayed-coincidence search in another germanium experiment, whose null result motivates and contextualises the analysis."},{"cited_title":"Ver- sion available athttp://www.nndc.bnl.gov/ensarchivals/ Full Evaluation Balraj Singh ENSDF 30-Sep-2020","cited_arxiv_id":null,"evidence_quote":"Decay-scheme data—fix the half-lives, Q-values, and the (33 ± 1)% branching ratio into the 77mAs isomer."}],"review_version":1}