{"id":"f78c5647-b2cc-4348-8f13-efa1a8b0fa19","arxiv_id":"2411.14355","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"PandaX-4T measured the 124Xe two-neutrino double electron capture half-life as (1.03±0.15_stat±0.08_sys)×10^22 years and a K-shell capture fraction of (65±5)% from 1.73 tonne-years of exposure.","lead":"The PandaX-4T liquid xenon experiment measured the half-life of the ultra-rare process where a xenon-124 atom captures two of its own electrons, finding (1.03±0.15±0.08)×10^22 years. The result independently confirms the XENONnT and LZ measurements and provides a calibration point for the nuclear theory behind neutrinoless double beta decay searches.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central result is not yet robust: the paper does not demonstrate that the floated 125I time model (Eq.","rationale":"The reader's weakest-assumption analysis pinpointed the 125I time-dependent background model, and my reading of the paper lands on the same point. The central calculation is internally consistent: 549 fitted events with the quoted exposure, isotopic abundance, and efficiency reproduce the reported half-life through Eq. 3.9, and the goodness-of-fit p-value is reasonable. The agreement with XENONnT and LZ provides genuine independent support. Nevertheless, the paper's own robustness text in Sec. 4 is explicit that the time-model uncertainty scan was performed 'particularly for 214Pb' and for linear spectral slopes, not for the 125I model. Because the 125I energy peaks are within roughly one resolution sigma of the signal peaks, and because tau_eff is floated with a 90% fractional uncertainty, the separation of 125I from signal relies on the exact form of Eq. 3.4 in a way that is not tested by any reported variation. This is a concrete, falsifiable gap rather than a demonstrated error. The proposed time-split test would settle it using only data already in the analysis. Since the reader already selected CONDITIONAL on essentially this concern, no verdict adjustment is needed.","tokens_in":14069,"tokens_out":6797,"duration_ms":68986,"concrete_test":"Re-fit the combined Run0+Run1 data with all events within 5 x tau_eff (about 15 days for the best-fit tau_eff = 2.9 d) of the AmBe/PuC neutron-calibration injections removed, so that the rapid 125I component is effectively gone and only the constant 2nuDEC signal plus the slow 125I component remain; compare the resulting half-life to (1.03 +/- 0.15) x 10^22 yr. If the shift exceeds the 0.15 x 10^22 yr statistical uncertainty, the 125I time model is load-bearing; if the late-only half-life stays within 1 sigma of the reported value, the concern is retired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The condition on which the central claim rests is that the two-dimensional (E,t) model for 125I -- multi-Gaussian energy template at 67.3/40.4/36.5 keV with fixed ratios and the three-exponential time model of Eq. 3.4 with floated rapid/slow normalizations and free tau_eff = (2.9 +/- 2.7) d -- separates 125I from the constant 124Xe 2nuDEC signal. The energy separation alone is marginal: KK (64.62 keV) and 125I-K (67.3 keV) are 2.7 keV apart against a resolution sigma of about 3.4 keV; KL (37.05 keV) and 125I-L (40.4 keV) are 3.4 keV apart against about 2.6 keV, and 125I-M (36.5 keV) nearly sits on KL. The only robust handle is time, but tau_eff is poorly constrained and the rapid 125I component could partially mimic a constant signal if its effective decay is slow. The paper's stated robustness check for background time models (Table 4, row 'Background models <0.01%'; Sec. 4) explicitly says 'particularly for 214Pb' and varies the 214Pb evolution and the linear spectral slopes; it does not vary the 125I time model, the slow/rapid split, or the fixed 125I K/L/M ratio. Thus the central half-life's immunity to 125I modeling is asserted rather than demonstrated. This is the load-bearing soft spot. Agreement with XENONnT and LZ is reassuring independent evidence, but it cannot by itself prove that the internal 125I-versus-signal separation is unbiased.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The PandaX-4T Collaboration reports a measurement of the two-neutrino double electron capture half-life of 124Xe using 1.73 tonne·yr of combined Run0 and Run1 data. An unbinned two-dimensional (energy, time) maximum likelihood fit with Gaussian-constrained nuisance parameters yields 549 signal events and T1/2 = (1.03 ± 0.15_stat ± 0.08_sys) × 10^22 yr, consistent with the XENONnT and LZ results. The KK capture fraction is measured as (65 ± 5)%. The paper presents the first time-dependent background model in the O(10 keV) region for PandaX-4T and tabulates eight contributions to the systematic uncertainty.","tokens_in":14283,"tokens_out":8113,"duration_ms":74982,"significance":"If robust, this result provides an important independent confirmation of 124Xe 2νDEC and a benchmark for nuclear matrix element calculations. The analysis is methodical: the likelihood function (Eq. 3.5) is well specified, nuisance parameters are tied to measured rates, the half-life formula (Eq. 3.9) is standard, and the fit quality (p = 0.19) is acceptable. The agreement with XENONnT and LZ is a strong cross-check. However, the separation between the 2νDEC signal and the 125I background rests almost entirely on the time model, and the paper does not yet demonstrate that this separation is robust to the plausible variations of that model. The central value is therefore defensible but not yet fully established.","major_comments":[{"comment":"The half-life extraction relies on separating 125I from 124Xe 2νDEC almost entirely on the time axis, because the energy separation is marginal at the quoted resolutions: 125I-K (67.3 keV) is about 0.8σ from 124Xe-KK (64.62 keV), 125I-L (40.4 keV) is about 1.3σ from KL (37.05 keV), and 125I-M (36.5 keV) nearly coincides with KL. However, τ_eff is floated freely and fitted as (2.9 ± 2.7) d, and the slow 125I normalizations are floated independently in each Run0 subset and Run1, with the Run1 slow component fitted as 28 ± 42 events. The \"Background models\" systematic in Table 4 is documented only for 214Pb time evolution and linear spectral slopes; no variation of the 125I K/L/M amplitude ratios, of the slow/rapid split, or of the functional form of Eq. (3.4) is reported. A slowly decaying rapid component or an unconstrained slow component can partially mimic a constant 2νDEC signal, so the claimed <0.01% background-model systematic may not cover the most dangerous model uncertainty. I request a dedicated robustness study that frees the 125I K/L/M ratios, adds a constant 125I-like component in Run1, replaces Eq. (3.4) with single- or double-exponential alternatives, and reports the resulting shifts in N2νDEC and in the final half-life.","section":"Sec. 3.4.2, Eq. (3.4); Table 4"},{"comment":"The slow 125I component has five independent floated normalizations (Run0a–Run0d and Run1), with no coupling to the activation history or to the rapid component, even though Eq. (3.4) describes a single dead-zone reservoir with a common physical lifetime. This freedom, combined with the free rapid normalizations and τ_eff, gives the 125I model substantial ability to absorb events that are kinematically similar to the signal. The goodness-of-fit p = 0.19 cannot resolve this degeneracy, because it tests the overall model against the data rather than the identifiability of the signal/125I decomposition. The authors should demonstrate identifiability directly, for example with toy Monte Carlo fits in which the true 125I time structure and K/L/M ratios are varied, and report the bias and pull of the fitted signal count.","section":"Table 2; Sec. 3.5"}],"minor_comments":[{"comment":"The final nucleus in Eq. (1.3) is typeset as \"124Te\" with an odd arrow; it should be 124Te with a standard arrow.","section":"Eq. (1.3)"},{"comment":"The figure caption contains truncated labels such as \"R es.\"; the full label \"Residual\" would improve readability.","section":"Sec. 4, Fig. 4 caption"},{"comment":"The text says \"The 125I peaks around 40 keV and 70 keV\"; the upper 125I peak is at 67.3 keV, so using the exact value would avoid inconsistency with Sec. 3.4.2.","section":"Sec. 4"},{"comment":"The sentence quoting energy resolutions of (5.3 ± 0.1)% and (6.9 ± 0.1)% should explicitly state that these values refer to the KK (64.62 keV) and KL (37.05 keV) peaks, respectively, to prevent ambiguity.","section":"Sec. 3.3"},{"comment":"The product \"m × t = 1.73 ton·yr\" in Eq. (3.9) is notationally compressed; a sentence clarifying that the exposure appears as a single product and specifying the units of each factor would help readers reproduce the numerical value.","section":"Eq. (3.9)"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the paper is technically careful and the measured half-life agrees with two independent experiments, so I see no reason to doubt the collaboration's competence or the result's plausibility. The missing element is a focused study of the 125I parameterization, which is a finite and doable addition rather than a fundamental flaw. I would not reject on the current evidence, but the paper should not be accepted until the 125I time-model robustness is quantified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a careful, well-documented measurement that independently confirms the 124Xe 2νDEC half-life. I'd bet the central value is right, but the one component that carries the analysis—the 125I time-dependent background—is also the component with the weakest validation argument. Send it to review, but make the 125I model a condition.\n\nWhat's new: not the first measurement—XENONnT and LZ have already published values, and the paper cites both—but it is a new independent detector, with 1.73 tonne·yr exposure and a 14.6% statistical precision on the half-life. The KK capture fraction (65±5%) is a second useful observable. The paper does a lot right: the unbinned 2D likelihood is standard and properly built, the nuisance parameters are Gaussian-constrained, the eight systematic terms add to 7.5%, and the goodness of fit (p=0.19) is acceptable. The internal arithmetic checks out: 549 fitted signal events reproduces the quoted half-life via Eq. 3.9. Agreement with XENONnT and LZ within 1σ is genuinely reassuring.\n\nNow the soft spot, and it is real and load-bearing. The 125I peaks at 67.3, 40.4, and 36.5 keV sit within about one energy-resolution sigma of the KK (64.62 keV) and KL (37.05 keV) signal peaks; separation therefore leans heavily on the time evolution. But the rapid 125I removal lifetime is floated with an enormous uncertainty (2.9±2.7 d), the rapid and slow normalizations float freely, and the slow 125I component in Run1 floats to 28±42 events—consistent with zero but with a large tail. The paper's robustness check for background models (Table 4, row 'Background models <0.01%') explicitly says 'particularly for 214Pb' and varies the 214Pb evolution and the linear spectral slopes. It does not vary the 125I time model, the K/L/M ratio, or the rapid/slow split. So the claim that background model uncertainties are negligible is asserted, not demonstrated, for the one background that matters most. This is not fatal, because the external agreement with two prior experiments lowers the risk, but it deserves a dedicated validation—a control region enriched in 125I, a 125I-free window, or a sensitivity check that deliberately perturbs the 125I model.\n\nWho this is for: nuclear theorists benchmarking NME calculations, and experimentalists in the 0νββ/0νDEC community. It deserves a serious referee. Recommendation: send to peer review, and make the 125I model validation a required revision.","headline":"Independent PandaX-4T measurement of the 124Xe 2νDEC half-life that agrees with XENONnT and LZ; the analysis is careful, but the lynchpin 125I time model is under-validated and needs a dedicated check before acceptance.","tokens_in":15462,"tokens_out":3017,"would_cite":true,"duration_ms":25505,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["23.40.-s"],"model":"deepseek-v4-flash","headline":"A 1.73 tonne-year liquid xenon exposure yields a 124Xe double electron capture half-life of (1.03 ± 0.15 ± 0.08) × 10^22 years.","keywords":["two-neutrino double electron capture","124Xe","half-life measurement","liquid xenon time projection chamber","time-dependent background model","unbinned maximum likelihood","nuclear matrix element","KK capture fraction"],"falsifier":"A future measurement with energy resolution better than about 2 keV at 65 keV, which resolves the 64.62 keV $KK$ peak from the 67.3 keV $^{125}$I peak without relying on the temporal model, would directly confirm or refute the half-life; alternatively, a dedicated $^{125}$Xe activation run that measures $\\tau_{\\rm eff}$ independently could test the temporal decomposition used here.","tokens_in":13677,"feed_emoji":"⚛️","tokens_out":17141,"duration_ms":119751,"temperature":0.7,"pith_summary":"This paper reports a measurement of the half-life of two-neutrino double electron capture ($2\\nu$DEC) in $^{124}$Xe, using 1.73 tonne\\,yr of data from the PandaX-4T liquid xenon detector. An unbinned maximum likelihood fit to energy and time distributions in the 25--75 keV window finds 549 signal events and yields a half-life of $(1.03 \\pm 0.15_{\\rm stat} \\pm 0.08_{\\rm sys}) \\times 10^{22}$ yr. The paper also measures the fraction of decays capturing both electrons from the $K$ shell to be $(65 \\pm 5)\\%$, consistent with nuclear structure calculations. This matters because a reliable $2\\nu$DEC half-life gives nuclear theorists a benchmark for nuclear matrix element calculations that also apply to neutrinoless double electron capture, the search for which probes whether neutrinos are Majorana particles.","feed_headline":"124Xe double electron capture half-life measured at 10^22 yr","feed_subtitle":"A 1.73 tonne-year run finds 549 signal events, sharpening the nuclear theory benchmark for electron capture decay.","key_machinery":"The load-bearing object is the two-dimensional (energy, time) unbinned maximum likelihood fit, with separate probability densities for signal and each background. Signal peaks are fixed at the calculated energies and capture fractions of $2\\nu$DEC in $^{124}$Xe; the critical background, $^{125}$I, is modeled by the time evolution $$N_{125I}(t) = -\\kappa_1 $e^{{-(t-t_0)/\\tau_{125Xe}}$} + \\kappa_2 $e^{{-(t-t_0)/\\tau_{\\rm eff}}$} + \\kappa_3 $e^{{-(t-t_0)/\\tau_{125I}}$},$$ separating a rapidly purified component ($\\tau_{\\rm eff} = 2.9 \\pm 2.7$ d) from a slow component diffusing out of circulation-inaccessible volumes. The fit also carries constrained nuisance parameters for energy resolution and linearity, the efficiency curve, and the slopes and rates of other backgrounds, and the half-life is converted from the fitted signal count using the exposure, isotopic abundance, and efficiency.","core_discovery":"The paper claims that the two-neutrino double electron capture half-life of $^{124}$Xe is $(1.03 \\pm 0.15_{\\rm stat} \\pm 0.08_{\\rm sys}) \\times 10^{22}$ yr, obtained from 549 signal events in a combined analysis of the commissioning and first science runs of PandaX-4T. In the analysis, the signal is modeled as five mono-energetic peaks ($KK$, $KL$, $KM$, $KN$, $KO$) from atomic de-excitation following double electron capture, with energies and relative fractions taken from nuclear structure calculations. The dominant difficulty is the cosmogenic background $^{125}$I, whose electron-capture peaks at 67.3, 40.4, and 36.5 keV sit close to the 64.62 keV $KK$ and 37.05 keV $KL$ signal peaks; the paper builds a time-dependent model for $^{125}$I with a rapidly removed component and a slowly diffusing component, and folds this into a two-dimensional (energy, time) unbinned likelihood together with all other backgrounds. The fit returns a $KK$ capture fraction of $(65 \\pm 5)\\%$, aligned with nuclear model predictions within 1.8$\\sigma$, and the half-life result is consistent with recent measurements in other liquid xenon detectors.","pith_inferences":["A dedicated $^{125}$Xe injection run could measure the effective removal lifetime $\\tau_{\\rm eff}$ directly, turning the floated parameter that carries much of the background separation into a calibrated input.","The same energy-time likelihood structure should transfer to other rare decay searches in liquid xenon, such as low-energy solar neutrino or axion-like particle searches, where activation backgrounds are similar.","If future runs triple the exposure, the statistical uncertainty could drop below 10%, potentially revealing deviations from the current central value and tightening comparisons with nuclear structure models."],"forward_implications":["The measured half-life gives nuclear theorists a concrete benchmark for $2\\nu$DEC and $0\\nu$DEC nuclear matrix element calculations.","The $KK$ capture fraction of $(65 \\pm 5)\\%$ provides a direct test of shell-model predictions for electron capture from different atomic shells.","The successful time-dependent background model shows that large liquid xenon TPCs can handle O(10 keV) rare decay searches even with neutron-activation backgrounds.","Agreement with independent measurements in other detectors confirms the $2\\nu$DEC signal and validates the PandaX-4T energy reconstruction and background handling."],"supporting_citations":[{"why":"Supplies the energies and capture fractions of the five 2νDEC subshell modes (KK through KO) used in the signal model.","marker":"[30]"},{"why":"Provides the event reconstruction, data quality cuts, and fiducial volumes inherited from the dark matter analysis.","marker":"[23]"},{"why":"Supplies the measured 214Pb and 212Pb activities from high-energy spectrum fits that constrain these dominant backgrounds.","marker":"[32]"},{"why":"Provides the in-situ 136Xe 2νββ half-life and abundance used to estimate that background component.","marker":"[36]"},{"why":"Gives an alternative set of 2νDEC energy depositions used to evaluate the signal-model systematic uncertainty.","marker":"[12]"},{"why":"Provides the effective-theory and nuclear-structure-model half-life predictions that the measured value is compared against.","marker":"[47]"},{"why":"Previous measurement of the same 124Xe 2νDEC half-life in a different liquid xenon detector, used as an agreement check.","marker":"[13]"},{"why":"Most recent previous measurement of the same half-life, also used as an agreement check.","marker":"[14]"}],"fun_headline_variants":["PandaX-4T pins 124Xe double electron capture at 10^22 yr","Most precise 124Xe 2νDEC half-life: 1.03×10^22 yr","124Xe double electron capture half-life sharpened by PandaX","549 events nail 124Xe 2νDEC half-life and shell capture ratio"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The assumption that carries the result is that the time-dependent model of the $^{125}$I background, with its rapid removal and slow diffusion components, correctly separates it from the $2\\nu$DEC signal; if the $^{125}$I spectral or temporal shape is wrong, the fitted signal count and half-life would be biased.","fun_headline_variants_meta":{"raw":{"variants":["PandaX-4T pins 124Xe double electron capture at 10^22 yr","Most precise 124Xe 2νDEC half-life: 1.03×10^22 yr","124Xe double electron capture half-life sharpened by PandaX","549 events nail 124Xe 2νDEC half-life and shell capture ratio"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000152,"raw_usage":{"total_tokens":1253,"prompt_tokens":1045,"completion_tokens":208,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":661,"completion_tokens_details":{"reasoning_tokens":113}},"tokens_in":661,"tokens_out":208,"duration_ms":2964,"temperature":1.0,"reasoning_tokens":113,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:17:23.367653+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A future measurement with energy resolution better than about 2 keV at 65 keV, which resolves the 64.62 keV $KK$ peak from the 67.3 keV $^{125}$I peak without relying on the temporal model, would directly confirm or refute the half-life; alternatively, a dedicated $^{125}$Xe activation run that measures $\\tau_{\\rm eff}$ independently could test the temporal decomposition used here.","supporting_citations":[],"review_version":1}