{"id":"e2ee90cd-1558-4a51-a34a-c86e5ae8c870","arxiv_id":"2611.16415","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A 228Th-D2O photoneutron source was experimentally characterized, yielding a forward emission rate of 1734 per hour and a neutron spectrum concentrated at 0.17-0.22 MeV for low-energy nuclear recoil calibration.","lead":"This paper develops a compact 228Th-D2O photoneutron source that produces sub-MeV neutrons by converting 2.615 MeV gamma rays from 208Tl via deuteron photodisintegration. The source provides a long-lived calibration tool for measuring few-keV nuclear recoil responses in large liquid xenon detectors used for dark matter and neutrino experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"Absolute 3He detection efficiency (7.63%) is not validated by an independent absolute measurement; the systematic uncertainty (3.76%) is derived entirely from simulation.","rationale":"The reader's weakest_assumption correctly identifies that the absolute ³He detection efficiency is not independently validated. This is indeed the most load-bearing concern because the absolute forward emission rate is a core numerical claim (1734 h⁻¹) and directly determines the utility of the source for calibration campaigns. The paper provides a thorough systematic uncertainty evaluation but omits an external validation step, which would be standard practice for such measurements. The reader's CONDITIONAL verdict is therefore appropriate, and our stress test does not change it. The condition for future work could specifically include an absolute efficiency calibration, but the reader's existing condition on coincidence gamma contamination is also valid.","tokens_in":20271,"tokens_out":674,"duration_ms":23873,"concrete_test":"Calibrate the entire detection setup (³He counter + PE moderator) using a well-characterised neutron source with known absolute emission rate, e.g., a traceable ²⁵²Cf source or an ²⁴¹Am-Be source with certified output, placed at the same position and geometry as the ²²⁸Th-D₂O source. Compare the measured net count rate with the Geant4-predicted rate to either confirm the 7.63% efficiency within ≤3.76% or quantify a systematic offset.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of a forward neutron emission rate of 1734 ± 76 ± 65 h⁻¹ (Eq. 4) rests on the Geant4-simulated absolute detection efficiency of 7.63% for the moderated ³He counter (Section 3). While the relative position dependence is validated (Fig.5), the absolute efficiency is not cross-checked against any calibrated neutron source. The systematic uncertainty (Table 1) arises from parameter variations within simulation, most notably a 3.27% contribution from a conservatively assumed 1 cm air gap in D₂O. However, the simulation of neutron transport through D₂O, polyethylene moderator, and capture in ³He may harbor unrecognised systematic errors—e.g., in the nuclear data (cross sections, thermalisation) or geometry details—that are not captured by the limited parameter scan. An independent absolute measurement would both reduce and validate this uncertainty. The paper acknowledges the need for final validation with PandaX-20T data, but the absolute rate itself is fundamental to the claim of sufficiency for calibration campaigns.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper presents the development and experimental characterization of a compact 228Th–D2O photoneutron source intended for low-energy nuclear recoil (NR) calibration of liquid xenon detectors. The source exploits the D(γ,n)p reaction from 2.615 MeV gamma rays in the 228Th decay chain. Through Geant4 simulations the D2O geometry and source position were optimized, and experimental measurements were performed using a moderated 3He proportional counter to determine the forward neutron emission rate (1734 ± 76 stat. ± 65 sys. h−1) and a Bonner sphere spectrometer (BSS) with MAXED unfolding to measure the neutron energy spectrum (peak at 0.17–0.22 MeV). A Geant4 simulation for the PandaX-20T detector suggests that about 1000 selected NR events could be accumulated in a 220 h exposure. The source offers a longer half-life (1.91 years) compared to 88Y–Be sources and provides a sub-MeV neutron spectrum suitable for few-keV NR calibration.","tokens_in":19938,"tokens_out":2977,"duration_ms":26887,"significance":"If the measurements are reliable, the 228Th–D2O source fills a practical need: a long-lived, compact photoneutron source with a clean low-energy spectrum for calibrating the few-keV NR response of large liquid xenon detectors. This is valuable for dark matter and CEνNS experiments where accurate low-energy response characterization is critical. The work includes careful D2O/H2O background subtraction, systematic uncertainty quantification, and comparison with Geant4 simulations. The potential dual-use for high-energy gamma calibration (2.615 MeV line close to the 0νββ Q-value) adds further interest. However, the reliance on a single simulated absolute detection efficiency (7.63%) without independent absolute calibration and the absence of an uncertainty band on the unfolded spectrum limit the strength of the central claims.","major_comments":[{"comment":"The forward neutron emission rate (1734 ± 76 stat. ± 65 sys. h−1) rests entirely on the Geant4-simulated absolute detection efficiency of 7.63% for the moderated 3He counter. The systematic uncertainty (3.76%) is derived from simulation parameter variations, but the absolute efficiency has not been validated against an independent calibrated neutron source. The relative position dependence is validated (Fig. 5), but this does not verify the absolute scale. The authors state that final validation with PandaX-20T data is needed, but this limitation should be stated more prominently in the abstract and conclusions, and the potential impact of an uncalibrated absolute efficiency on the emission rate should be discussed quantitatively.","section":"Section 3, Eq. (4)","authors_acknowledge":"A final validation based on the complete PandaX-20T detector simulation or data is still needed. 6.","authors_acknowledge_p":0.65},{"comment":"The unfolded neutron spectrum is shown without any uncertainty band. The authors acknowledge in the text that 'A quantitative assessment of the statistical significance of any discrepancy is not attempted here because the figure does not show an uncertainty band for the unfolded spectrum' (sentence [b213]). For a paper that uses this spectrum as input to the PandaX-20T calibration simulation, the absence of spectral uncertainties is a significant omission. The authors should provide an uncertainty band (e.g., using the covariance matrix from the MAXED unfolding) or explain how the shape uncertainty affects the simulation results. Without this, the reliability of the shape-dependent predictions (e.g., the 4.60 events h−1 NR rate) is unclear.","section":"Section 4, Fig. 7","authors_acknowledge":"A quantitative assessment of the statistical significance of any discrepancy is not attempted here because the figure does not show an uncertainty band for the unfolded spectrum.","authors_acknowledge_p":0.93},{"comment":"The simulation reports neutron-induced NR and gamma-induced ER rates separately but does not evaluate the probability that a 2.615 MeV gamma ray deposits energy in the same event window as a selected NR. The authors correctly note that 'This coincidence-induced ER contamination should be quantified in a future event-level simulation before claiming a specific NR sample purity' (sentence [b280]). This omission directly affects the claim that 'about 1000 selected nuclear-recoil events' can be obtained. The paper should either provide a rough estimate of the coincidence contamination or explicitly state that the 1000-event figure refers to raw NR candidates before purity selection. The current presentation could mislead readers.","section":"Section 5, final paragraph","authors_acknowledge":"This coincidence-induced ER contamination should be quantified in a future event-level simulation before claiming a specific NR sample purity.","authors_acknowledge_p":0.94}],"minor_comments":[{"comment":"The caption states 'Because the simulated cross sections are modified with respect to the physical ones, only the relative angular shape is meaningful.' This is a crucial statement that should also appear in the main text, not only in the figure caption. Additionally, the reader needs to know what modifications were made and why, or at least a reference for the modified cross sections.","section":"Section 2.1, Fig. 2 caption","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The schematic in Fig. 3 has components 1,2,3,4,5,6,7,8 but the caption lists them as 1, Fused Silica Bottle; 2, Actuating Rod; etc. The numbering in the figure should match the caption list exactly (currently the caption starts with '1. Fused Silica Bottel' – typo 'Fused' is misspelled).","section":"Section2.2, Fig. 3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"In Fig. 7, the vertical axis label reads '/;dE φ E d A)'. This appears to be a formatting artifact from the PDF conversion. It should be clearly given as 'E dφ/dE (arb. units)' or similar. Also, the inset panel (B) should have the same axis label or explain th normalization.","section":"Section 4, Fig. 7 axes","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The reference to BambooMC [42] is appropriate, but the paper does not state whether the simulation uses the same Geant4 version and physics list as the source optimization. A brief statement of consistency would be helpful.","section":"Section 5, first paragraph","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper uses the phrase 'forward end face' repeatedly without defining a coordinate system. A coordinate axis or a sketch showing the forward direction relative to the detector would improve clarity.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"minor_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"This paper reports the first dedicated experimental characterization of a 228Th–D2O photoneutron source for low-energy nuclear recoil calibration in LXe detectors. The core ountputs—a forward emisssion rate of 1734 h⁻¹ and a neutron spectrum peakd at 0.17–0.22 MeV—come from careful measurements with proper D2O/H2O backgound subtraction and Bonner sphere unfolding. The geometry optimization and comparison to simulation are weell done. The systematic uncertaintiy breakdown (Table 1) is transparent and reasonble for a simulation-based efficiency. The PandaX‑20T Monte Caro study demonstrates the source's fesibility for calibration campaignnes.\n\nThe load-bearing claim is the forward emission rate, which rests entirely on the Geant4‑simulated absolute detection efficiency of 7.63% for the moderated ³He counter (Section 3). The pap er acknowledges the need for final validation with PandaX‑20T data (last sentence of Section 5), but the absolute rate itself is not cross‑checked against any independent calibrated neutron source. The systematic uncertainty (3.76%) is derived from parameter variations inside the same MC framework—there is no independent absolute measurement to validate that uncertainty envelope. This is the pap er's softest spot. The paper also notes that cooincidence γ‑induced ER contamination is not yet evaluted in the PandaX simuulation (Sect. 5), and that the angular distribution in Fig. 2 is onlly relative because the cross sections were modifed (Sect. 2.1). Those are acknowledged limitations, not flaws.\n\nOverall, this is a competent, honest piece of instrument development. The measurments are sound but the absolute nor malization rests on simulation in a way that the authors don't overstate, but reviewers should push for a clearer discussion of how that efficiency could be validated—or at least an explcitit statement that the rate is model‑dependent until cross‑checked. The paper deserves a serious peer review: it provide s a useful engineering tool for the LXe community, and the data are worth publishing if the caveat on absolute efficiency is properly flagged.\n\nIn summry: yes to peer review, but the referee should ask the authors to either perform an independent efficiency cross‑check with a known neutron source (if possible) or strengthen the text to make the simulation‑dependence of the central result unmistakable.","headline":"Solid engineering paper with a central result that depends on unvalidated simulated efficiency; worth publishing after a clearer discussion of the absolute calibration caveat.","tokens_in":20486,"tokens_out":1838,"duration_ms":16578},"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek/deepseek-v4-flash","headline":"A compact 228Th–D₂O photoneutron source provides a stable low-energy neutron field for calibrating few-keV nuclear recoils in large liquid xenon detectors, with a measured forward emission rate of 1734 ± 76 (stat.) ± 65 (sys.) h⁻¹ and a…","keywords":["Photoneutron source","Low-energy neutron calibration","Liquid xenon detector","Bonner sphere spectrometer","Nuclear recoil","Thorium-228","Heavy water photodisintegration","PandaX-20T"],"falsifier":"An independent absolute measurement of the neutron emission rate using a technique such as the manganese‑bath activation method would directly test the simulated detection efficiency. Alternatively, deploying the source in PandaX‑20T and comparing the observed nuclear‑recoil event rate with the simulated prediction of ~4.60 events h⁻¹ would validate the combined source and detector model at the level of the full calibration chain.","tokens_in":21130,"feed_emoji":"🔬","tokens_out":4582,"duration_ms":36579,"temperature":0.7,"pith_summary":"This paper develops and characterizes a compact photoneutron source using the 2.615 MeV gamma ray from ²²⁸Th decay to drive the D(γ,n)p reaction in heavy water. The authors optimize the geometry so that the source is placed 1 cm from the forward face of the D₂O volume, maximizing forward neutron emission. They measure the emission rate with a ³He proportional counter and the energy spectrum with a Bonner sphere spectrometer, obtaining a forward rate of 1734 ± 91 h⁻¹ and a spectrum concentrated at 0.17–0.22 MeV, entirely below 1 MeV. Simulating deployment in the PandaX‑20T detector, they predict about 1000 selected nuclear‑recoil events in a 220‑h exposure. If correct, this source offers a longer‑lived alternative to ⁸⁸Y‑Be sources for low‑energy calibration of dark‑matter and neutrino detectors, with the added benefit of a simultaneous 2.615 MeV gamma‑ray line for high‑energy calibration.","feed_headline":"Th-228-D2O source yields 1734 neutrons/h for Xe calibration","feed_subtitle":"Long-lived (1.9 y) source provides stable field at 0.17–0.22 MeV, enabling ~1000 nuclear recoil events in PandaX-20T within 220 h.","key_machinery":"The D(γ,n)p photodisintegration reaction, driven by the 2.615 MeV gamma ray from ²⁰⁸Tl in the ²²⁸Th decay chain. This two‑body reaction’s kinematics, with only 0.391 MeV above the deuteron binding energy, naturally restricts neutron energies to the sub‑MeV range. The measured spectrum is shaped by the reaction kinematics, elastic scattering on deuterium and oxygen in D₂O, and geometry‑dependent escape probabilities, as confirmed by Geant4 track‑by‑track analysis.","core_discovery":"The central claim is that a ²²⁸Th–D₂O photoneutron source, with the ²²⁸Th disk positioned 1 cm from the forward end face of a D₂O cylinder, produces a neutron field whose forward emission rate is 1734 ± 76 (stat.) ± 65 (sys.) h⁻¹ and whose energy spectrum is peaked at 0.17–0.22 MeV, with all neutrons below 1 MeV. The authors assert that this spectrum is consistent with Geant4 simulations and that the source is suitable for few‑keV nuclear‑recoil calibration of large liquid‑xenon detectors. They further claim that the source’s half‑life (1.91 y) is substantially longer than that of commonly used ⁸⁸Y‑Be sources (106.6 d), and that its spectrum lacks the ~0.95 MeV component present in ⁸⁸Y‑Be, making it cleaner for low‑energy calibrations.","pith_inferences":["The source could be adapted for other detector technologies (e.g., solid‑state, organic scintillators) that require a low‑energy neutron field for nuclear‑recoil calibration, possibly with different shielding designs.","The strong dependence of the escaping spectrum on D₂O geometry suggests that varying the source position or D₂O volume could tailor the neutron energy distribution for specific calibration needs, e.g., shifting the peak energy.","An independent absolute measurement of the neutron emission rate (e.g., by manganese‑bath activation) would strengthen confidence in the Geant4‑derived detection efficiency, which currently carries the dominant systematic uncertainty.","The paper’s note that the simulated cross sections were modified for Fig.2 (only the relative angular shape is meaningful) hints that the absolute angular distribution of emitted neutrons is not fully validated, which could affect the simulated event rate in detectors installed at non‑forward positions."],"forward_implications":["Large liquid‑xenon detectors (e.g., PandaX‑20T, XENONnT, LZ) can use this source for extended low‑energy nuclear‑recoil calibration campaigns without frequent source replacement due to its 1.91‑year half‑life.","The simultaneous 2.615 MeV gamma ray provides a calibration line near the 0νββ Q‑value of ¹³⁶Xe, enabling cross‑calibration of the high‑energy electron‑recoil response in a separate deployment mode.","The measured emission rate and spectrum serve as a benchmark for validating Geant4 simulations of neutron transport and detection in liquid‑xenon detectors.","The optimized shielding design (tungsten annulus + forward tungsten cylinder) demonstrates a path to achieve a usable NR/ER ratio while staying within data‑acquisition bandwidth.","An exposure of ~220 h in PandaX‑20T is predicted to yield ~1000 selected nuclear‑recoil events, sufficient for characterizing the detector’s few‑keV response."],"supporting_citations":[{"why":"Provides the absolute cross section of deuteron photodisintegration at 2.615 MeV and the kinematics of the reaction; the central nuclear data on which neutron yield calculations depend.","marker":"[26]"},{"why":"Describes the ⁸⁸Y‑Be photoneutron source that the present source aims to supplant, providing the baseline for comparison in spectral purity and half‑life.","marker":"[21]"},{"why":"Reports the use of a ⁸⁸Y‑Be photoneutron source in the XENONnT low‑energy nuclear‑recoil calibration, establishing the application context.","marker":"[22]"},{"why":"Supplies the half‑life and decay data of ²²⁸Th and its daughters, including the ²⁰⁸Tl branching ratio needed to compute the 2.615 MeV emission rate.","marker":"[25]"},{"why":"Geant4 simulation toolkit used for all Monte‑Carlo calculations of neutron production, transport, and detection efficiency.","marker":"[29]"},{"why":"Describes the PandaX‑20T detector design and geometry, which is the target detector for the calibration simulation.","marker":"[31]"},{"why":"MAXED unfolding algorithm used to recover the neutron spectrum from Bonner‑sphere data; central to spectral measurement.","marker":"[36]"}],"fun_headline_variants":["1734 neutrons/h from compact Th-D2O source at 0.2 MeV","1.9-year photoneutron source sharpens Xe recoil calibrations","Clean 0.17–0.22 MeV neutron field for PandaX-20T","Th-228 D2O source: long-lived alternative to Y-Be for Xe detectors"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The Geant4 simulation of the 3He counter’s absolute detection efficiency (7.63%) is taken as correct, but it is not validated by an independent absolute measurement; the simulation is normalized to data only for relative position dependence, not for absolute normalization.","fun_headline_variants_meta":{"raw":{"variants":["1734 neutrons/h from compact Th-D2O source at 0.2 MeV","1.9-year photoneutron source sharpens Xe recoil calibrations","Clean 0.17–0.22 MeV neutron field for PandaX-20T","Th-228 D2O source: long-lived alternative to Y-Be for Xe detectors"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":0.000229,"raw_usage":{"cost":0.000228928,"is_byok":false,"cost_details":{"upstream_inference_cost":0.000228928,"upstream_inference_prompt_cost":0.000115808,"upstream_inference_completions_cost":0.00011312},"total_tokens":1436,"prompt_tokens":1032,"completion_tokens":404,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":256,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":310}},"tokens_in":1032,"tokens_out":404,"cost_source":"openrouter_usage","duration_ms":4598,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":310,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-13T15:17:33.614067+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":"An independent absolute measurement of the neutron emission rate using a technique such as the manganese‑bath activation method would directly test the simulated detection efficiency. Alternatively, deploying the source in PandaX‑20T and comparing the observed nuclear‑recoil event rate with the simulated prediction of ~4.60 events h⁻¹ would validate the combined source and detector model at the level of the full calibration chain.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the absolute cross section of deuteron photodisintegration at 2.615 MeV and the kinematics of the reaction; the central nuclear data on which neutron yield calculations depend."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the ⁸⁸Y‑Be photoneutron source that the present source aims to supplant, providing the baseline for comparison in spectral purity and half‑life."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the use of a ⁸⁸Y‑Be photoneutron source in the XENONnT low‑energy nuclear‑recoil calibration, establishing the application context."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the half‑life and decay data of ²²⁸Th and its daughters, including the ²⁰⁸Tl branching ratio needed to compute the 2.615 MeV emission rate."},{"cited_title":"GEANT4–a simulation toolkit","cited_arxiv_id":null,"evidence_quote":"Geant4 simulation toolkit used for all Monte‑Carlo calculations of neutron production, transport, and detection efficiency."},{"cited_title":"PandaX-xT: a Multi-ten-tonne Liquid Xenon Observatory at the China Jinping Underground Laboratory","cited_arxiv_id":"2402.03596","evidence_quote":"Describes the PandaX‑20T detector design and geometry, which is the target detector for the calibration simulation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"MAXED unfolding algorithm used to recover the neutron spectrum from Bonner‑sphere data; central to spectral measurement."}],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}