{"id":"e07c6b77-c01d-42b6-81d2-57fb507eca7e","arxiv_id":"2511.15912","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Proton energy response of Nab's silicon detectors was stable over a year, with a dead layer of about 55 nm, and the timing model predicts proton time-of-flight bias uncertainty below 0.3 ns.","lead":"Scientists tested the silicon detectors for the Nab neutron-decay experiment with low-energy protons and radioactive sources, measuring how much energy protons lose in the detector's entrance window and how stable the signal is over a year. These results are used to predict that detector-related timing errors for proton time-of-flight stay below 0.3 ns, meeting the experiment's target precision.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proton timing-bias claim rests on an electron-calibrated pulse-shape model; measured 30 keV proton waveforms exist but were not compared to NESSE predictions. A direct proton pulse-shape/rise-time comparison is needed.","rationale":"The paper's energy-response work—dead-layer thickness, temperature stability, cross-talk—is careful and independently supported by SIMS-based modeling. The central timing claim, however, depends on NESSE predicting proton pulse shapes from an electron-only calibration. This is exactly the reader's weakest assumption, and it is load-bearing: if proton pulse shapes differ in a way not captured by the electron fit, the timing offsets and hence the proton time-of-flight bias could exceed the 0.3 ns requirement. The concern is not that the model is wrong, but that it is unvalidated for the particle type that matters most for the headline claim. The paper even has the data needed to close the gap: 30 keV proton waveforms were recorded on the same pixels and bias voltages used in the rise-time analysis. A direct comparison of measured and simulated proton rise-time distributions would settle whether the extrapolation is valid. Because the concern is specific, addressable, and does not undermine the energy-response results, the existing CONDITIONAL verdict remains appropriate rather than a full rejection.","tokens_in":26484,"tokens_out":5022,"duration_ms":57425,"concrete_test":"Extract 10-90% rise-time distributions and average waveforms from the existing 30 keV proton datasets (Section III D, at -300 V and at least one lower bias) on pixels 76, 87, 97, 106, and 114. Generate matched NESSE pulses using the fitted impurity densities and the 0.979 drift-velocity scaling factor, add the same baseline noise used for the electron comparison, and compare the simulated and measured proton rise-time distributions. If the distributions agree within statistical and noise uncertainties, the electron-to-proton transfer is supported; if they disagree, or if refitting impurity densities with proton rise-times shifts the fitted values beyond uncertainties, the <0.3 ns timing-bias claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the transfer of NESSE from electron calibration to proton timing. In Section VI A 2, the simulator is tuned to 109Cd (87 keV) and 113Sn (364 keV) electron rise-time distributions, adjusting impurity densities and a single electron-drift-velocity scaling factor (0.979±0.002). In Section VI B 2, the same simulator generates 30 keV proton pulses, and the resulting timing offsets t_d from Eq. (7) are used to claim proton timing systematic uncertainties below 0.3 ns. No measured proton pulse shape is compared with NESSE, even though the Manitoba datasets include 30 keV proton waveforms at 12 bias voltages (Section III D) on the same radial pixel set used for the impurity-density fit. If proton pulses differ—because the charge cloud is created very close to the entrance window, with high local ionization density and dead-layer/weighting-field effects that may be negligible for electron events but are not validated for protons—the timing offset could shift by more than the 0.3 ns budget. The quoted 0.2 ns uncertainty at -300 V is a spread over simulated events, not an estimate of model error from the electron-to-proton extrapolation. The energy/dead-layer characterization is credible, but it does not test the pulse-shape model for protons.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper characterizes a Nab silicon detector using a dedicated low-energy proton source and radioactive electron sources. The energy calibration combines 109Cd/113Sn peaks with Geant4 modeling of source-foil energy losses and temperature corrections; proton energy spectra at 25/30/35 keV yield a hard/soft dead layer of about 55/60 nm and are stable at the calibration-uncertainty level over a year of cooling cycles. Cross-talk between neighboring pixels is measured to be below ~1%. Pulse-shape analysis fits NESSE simulations to electron rise-time distributions to extract a radial impurity-density profile and a drift-velocity scaling factor, then applies the same simulator to 30 keV proton pulses to estimate timing offsets. The paper claims the resulting proton timing systematic uncertainties are below 0.3 ns, sufficient for Nab.","tokens_in":26885,"tokens_out":6165,"duration_ms":71016,"significance":"If the timing claim holds, this is a valuable detector characterization that directly supports the Nab experiment's sub-nanosecond time-of-flight requirement. The energy and dead-layer results are carefully executed: uncertainties from source-foil thickness fits, temperature corrections, and peak extraction are propagated, the SIMS-based dead-layer model is a genuinely parameter-free cross-check, and the one-year proton-peak stability study addresses an important operational concern. The pulse-shape simulation infrastructure is well motivated and the impurity-density characterization is useful in its own right. However, the proton-timing conclusion is only as strong as the electron-to-proton model transfer, and that transfer is not directly validated in the manuscript.","major_comments":[{"comment":"The paper's central timing claim rests on NESSE being transferred from electron calibration to proton pulses. The simulator is tuned to 109Cd and 113Sn electron rise-time distributions (Sec. VI A 2), and Fig. 21 then uses the same simulator to report timing offsets for 30 keV protons. No measured proton pulse shape or rise-time distribution is compared with NESSE anywhere in the manuscript, even though 30 keV proton waveforms were recorded at 12 bias voltages (Sec. III D), including on radial pixels. The quoted 0.2 ns uncertainty at -300 V is the spread over simulated events, not an estimate of model error from the electron-to-proton extrapolation. A direct proton-pulse comparison, or an explicit quantified model-uncertainty term, is required before the <0.3 ns conclusion is established.","section":"Sec. VI B 2 / Sec. VI A 2, Fig. 21"},{"comment":"The manuscript describes timing-offset standard deviations as satisfying the Nab 0.3 ns systematic requirement. But the Nab requirement applies to the uncertainty of the correction for the time-of-flight bias, i.e., the difference between electron and proton timing offsets. The per-event spatial spread of simulated t_d values does not by itself bound the systematic error of the mean correction; common-mode model errors shift the mean offset without contributing to that spread. The statement that the timing-bias uncertainty can be reduced to <=0.2 ns is also explicitly conditional on characterizing all pixels and the full electron-energy range, which is not demonstrated in this paper.","section":"Sec. VI B / Sec. VII"},{"comment":"The beta-decay electron that defines t=0 in Nab has a continuous spectrum up to about 1 MeV, while the timing analysis simulates only 87 keV and 364 keV electrons. The paper itself notes an approximately 2 ns average timing-offset difference between these two energies at -300 V (Figs. 20 and 21). Without an energy-dependent electron timing model validated against data, the 'below 0.3 ns' claim applies only to the discrete simulated cases, not to Nab's full electron acceptance. The final paragraph of Sec. VII defers the full electron-energy assessment, so the abstract's blanket statement overstates what is currently established.","section":"Sec. VI B 2 / Sec. VII"}],"minor_comments":[{"comment":"The abstract quotes a 0.25 keV calibration uncertainty while the body text (Sec. V A and elsewhere) quotes ~0.2 keV. Please harmonize the value and clarify whether the difference is intentional.","section":"Abstract vs. Sec. V"},{"comment":"The fit function in Eq. (7) uses an exponential time constant of 1250 and a sigmoid width parameter f, but the units are not stated. Since these parameters enter the timing-offset extraction, please define them explicitly (e.g., ns).","section":"Eq. (7)"},{"comment":"The x-axis is labeled 'Detector pixel ring' but the mapping from pixel numbers (76, 87, 97, 106, 114) to ring numbers is not given in the caption. Adding the mapping would improve reproducibility.","section":"Fig. 18 caption"},{"comment":"The proton-bias-scan dataset is listed as 30 keV protons at 12 bias voltages, but the pixels used are not stated. Clarify whether these include the same five radial pixels (76, 87, 97, 106, 114) used for the impurity-density fit, since this is relevant to the direct proton validation suggested above.","section":"Sec. III D"},{"comment":"Typographical issues: 'AD8011 pre-amplifer' (Sec. III A 2), 'T rapezoidal' (Sec. IV heading), and 'Pehlet al.' (Sec. V B) should be corrected.","section":"Various"}],"recommendation":"major_revision","confidential_remarks":"This is a solid instrumentation paper with a credible energy/dead-layer characterization. The main issue is that the headline proton-timing systematic claim is not directly supported by measured proton pulse shapes; the data needed for that validation appear to already exist in the Manitoba datasets. I would ask for a direct proton-waveform comparison with NESSE and a clearer separation of statistical spread from model uncertainty before accepting the <0.3 ns conclusion."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague —\n\nThe energy and stability work is solid; the sub-0.3 ns proton timing systematic is not yet established. Dead-layer (55±2 nm hard/60±2 nm soft), the SIMS-model agreement, one-year peak stability, and the sub-1% cross-talk are all careful and convincing results from a real calibration chain. The radial impurity-density profile (2e9 to 26e9 cm^-3) and the 0.979±0.002 drift-velocity scaling factor are honest fits to electron rise-time data.\n\nThe soft spot is the electron-to-proton transfer in the timing analysis. NESSE is fitted to 109Cd and 113Sn electron rise times, then applied to 30 keV proton pulses to quote timing systematics below 0.3 ns. The paper notes that dead-layer, grid, self-repulsion, and recapture are negligible for electron event rise times, but proton charge clouds sit right at the entrance window, so that statement is untested for protons. Measured 30 keV proton waveforms exist at 12 bias voltages; not comparing them to NESSE is a missed check. The quoted 0.2 ns at -300 V is simulation spread, not model error; a dead-layer or weighting-field effect could push the proton timing offset by more than the 0.3 ns budget without this validation.\n\nThis is an addressable weakness, not a fatal one. The fix is to show the proton pulse-shape comparison (rise times and, ideally, timing offsets) or at least to present the timing bias as a model-based estimate with sensitivity to the calibration parameters and dead-layer assumptions quantified.\n\nWorth a serious referee. The energy characterization is a solid contribution for Nab; the timing section needs focused revision before the <0.3 ns conclusion is usable. I'd cite the dead-layer and stability results as they stand, but I'd hold off on citing the timing number.","headline":"Energy and stability characterization for Nab detectors is solid; the sub-0.3 ns proton timing systematic rests on an electron-calibrated simulation not validated against measured proton waveforms.","tokens_in":27418,"tokens_out":4502,"would_cite":true,"duration_ms":45371,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Wk"],"model":"deepseek-v4-flash","headline":"Proton timing bias in Nab silicon detectors is shown to stay below the experiment's 0.3 ns requirement.","keywords":["silicon detectors","neutron beta decay","Nab experiment","dead layer","impurity density profile","pulse shape simulation","timing bias","time-of-flight"],"falsifier":"Measure the rise-time distribution of 30 keV protons with the same detector and readout using a pulsed proton beam whose time structure provides the true start time, then compare the measured proton timing offsets to NESSE predictions; if the simulated and measured offsets differ by more than 0.3 ns relative to electron events, the central claim is falsified.","tokens_in":1223,"feed_emoji":"⚛️","tokens_out":4159,"duration_ms":81397,"temperature":0.7,"pith_summary":"This paper characterizes the response of the Nab experiment's large-area segmented silicon detectors to the low-energy protons produced in neutron beta decay. It establishes that the detector dead layer is about 55 nm thick, that detected proton peaks are stable within energy-calibration uncertainty across a year of repeated cooling cycles, and that the dominant pulse-shape contribution to proton time-of-flight bias is below 0.3 ns, meeting the experiment's requirement. The timing conclusion rests on a pulse-shape simulator calibrated against electron source data and then used to predict proton pulse shapes and timing offsets. A sympathetic reader would take this as evidence that the detectors can support the experiment's 0.1% physics goal.","feed_headline":"Proton pulse-timing systematic uncertainty drops below 0.3 ns","feed_subtitle":"Electron-calibrated detector simulations put the Nab time-of-flight systematic inside its required budget.","key_machinery":"The load-bearing tool is NESSE, a pulse-shape simulator named in the paper, which computes induced currents via the Shockley-Ramo theorem using drift trajectories set by the electric field, impurity density, and temperature. The critical step is calibrating this simulator against measured electron rise-time distributions—by simultaneously fitting pixel-by-pixel impurity densities and a global electron drift-velocity scaling factor (0.979±0.002)—and then applying the same calibrated simulator to 30 keV proton pulses to predict timing offsets. This converts a detector characterization into a timing-bias prediction.","core_discovery":"The paper claims that the pulse-shape systematic in proton time-of-flight measurements for the Nab experiment can be predicted and bounded below 0.3 ns. It achieves this by measuring electron rise times from two radioactive sources on five pixels, fitting a radial impurity-density profile from (2±2)×10^9 cm^-3 at center to (26±2)×10^9 cm^-3 at edge, and applying a drift-velocity scaling factor of 0.979±0.002 to match the electron data. The calibrated simulator is then used to generate 30 keV proton pulses and extract proton timing offsets relative to electrons. The accompanying measurements show a dead layer of (55±2) nm hard or (60±2) nm soft, and no significant proton-peak drift over one y","pith_inferences":["If the electron-calibrated drift model extrapolates faithfully to protons, the same calibration procedure could predict timing biases for other particle types without dedicated proton beam time; a direct measurement of proton pulse rise times with a pulsed beam would test this extrapolation.","The radial impurity gradient suggests that Nab's offline analysis will need per-pixel timing corrections, and other float-zone silicon detectors facing similar rise-time variation could adopt the same calibration method.","The stability result was obtained at pressures somewhat higher than Nab's projected ultrahigh vacuum; the lack of observed surface deposits is encouraging but the in-situ UHV case remains to be demonstrated.","The overall approach—deriving a timing model from rise-time distributions of calibration sources—could transfer to other precision beta-decay experiments using segmented silicon detectors."],"forward_implications":["The 0.3 ns timing-bias requirement for Nab proton time-of-flight is achievable, with timing-offset uncertainties around 0.2 ns at -300 V bias.","Optimal detector running conditions are near -300 V bias and about 120 K, where pulse rise times are short and leakage-current noise is not limiting.","The measured radial impurity gradient implies pixel-dependent pulse shapes, and the paper provides the per-pixel impurity map needed to correct or simulate them.","No significant proton-peak drift across a year of liquid-nitrogen cooling cycles indicates that cryo-pumping surface contamination is not degrading the entrance window under the tested conditions.","Nearest-neighbor cross-talk below about 1% keeps the proton trigger threshold low enough for 25–35 keV protons."],"fun_headline_variants":["Nab proton timing systematic cut to 0.3 ns","Nab's proton pulse timing error below 0.3 ns","Electron-calibrated Nab detector yields sub-0.3 ns proton timing","Detector calibration yields <0.3 ns proton timing for Nab"],"cache_read_input_tokens":28672,"weakest_assumption_plain":"The timing-bias estimate assumes that a pulse-shape simulator, tuned only against electron rise-time data from two calibration sources, predicts proton pulse shapes accurately enough that the residual proton-electron timing-offset difference stays below 0.3 ns.","fun_headline_variants_meta":{"raw":{"variants":["Nab proton timing systematic cut to 0.3 ns","Nab's proton pulse timing error below 0.3 ns","Electron-calibrated Nab detector yields sub-0.3 ns proton timing","Detector calibration yields <0.3 ns proton timing for Nab"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001412,"raw_usage":{"total_tokens":5591,"prompt_tokens":844,"completion_tokens":4747,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":588,"completion_tokens_details":{"reasoning_tokens":4670}},"tokens_in":588,"tokens_out":4747,"duration_ms":32737,"temperature":1.0,"reasoning_tokens":4670,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T21:14:22.783714+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the rise-time distribution of 30 keV protons with the same detector and readout using a pulsed proton beam whose time structure provides the true start time, then compare the measured proton timing offsets to NESSE predictions; if the simulated and measured offsets differ by more than 0.3 ns relative to electron events, the central claim is falsified.","supporting_citations":[],"review_version":1}