{"id":"a40c9bf5-3f1a-4831-b873-635afbb9ed23","arxiv_id":"2507.09208","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The JUNO calibration house has been installed and tested, maintaining oxygen below 10 ppm and radon below 15 mBq/m3.","lead":"JUNO's calibration house is a sealed chamber that links calibration equipment to the neutrino detector's central vessel while keeping oxygen and radon out of the liquid scintillator. Onsite tests reported here achieved oxygen below 10 ppm and radon below 15 mBq/m3, meeting the experiment's background requirements.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Radon budget passes only through the no-convection transfer fraction; the numeric margin is a factor of about 2.5 rather than the factor of about 100 that the text implies, so modest chimney mixing would violate the 0.005 Bq detector budget.","rationale":"The reader correctly identified the no-convection diffusion assumption in Section 4.2.2 as the weakest point. My stress-test sharpens this into a concrete arithmetic problem: the paper compares the measured house concentration times the 1% transfer coefficient to a 0.2 mBq/m^3 requirement, but that requirement is itself a house-concentration limit only under the assumption that all house radon enters the detector. When the check is written with consistent units, C_house x epsilon x V_house < 0.005 Bq, the measured 8 mBq/m^3 passes by a factor of about 2.5, not by the factor of about 100 that the text implies. Because the margin is thin, the central radon claim depends critically on the transfer fraction remaining below about 2.5%. If convection or mixing occurs in the 4 m chimney, the detector budget can be exceeded even though the house concentration itself is below the stated 15 mBq/m^3. This does not mean the paper is wrong; the no-convection assumption is stated explicitly and may be valid. But the acceptance should be conditioned on verifying that the transfer fraction is actually as low as assumed, because the margin is much smaller than the presentation suggests. The concrete test of measuring the ratio at the chimney bottom, or a bounded CFD study, would settle the question. I do not see a reason to reject the paper outright; the design and measurements are otherwise well documented, and the leak rate and oxygen claims are independently supported by onsite tests. However, the radon budget claim should be conditional on this verification.","tokens_in":10294,"tokens_out":14562,"duration_ms":169025,"concrete_test":"Measure the 222Rn concentration in the liquid scintillator near the bottom of the chimney during a representative calibration-house nitrogen-flush and source-deployment cycle, using a small dedicated sampling probe or an existing tube, and compare it with the simultaneously measured radon concentration in the calibration house. If the observed ratio exceeds about 2.5% (equivalently, if the inferred influx to the CD exceeds 0.005 Bq), the stated radon budget fails. A complementary bounded CFD simulation with buoyancy and cable-motion forcing could also bound the transfer fraction; a value above 2.5% would invalidate the acceptance.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central radon check in Section 4.2.3 compares the measured house concentration (8 +/- 4 mBq/m^3 in the flushed setup) times the 1% surface-absorption coefficient of Section 4.2.1, giving 0.08 +/- 0.04 mBq/m^3, against a 'requirement' of 0.2 mBq/m^3. This comparison is dimensionally muddled: the 0.2 mBq/m^3 value was obtained by dividing the 0.005 Bq detector budget by the 25 m^3 house volume, so it is a house concentration only if all radon in the house enters the CD. The correct inequality is C_house x epsilon x V_house < 0.005 Bq. Using epsilon = 1% gives 8e-3 Bq/m^3 x 0.01 x 25 m^3 = 0.002 Bq, which passes by a factor of only 2.5. Thus the claimed margin is much thinner than the paper's presentation suggests, and it rests entirely on the transfer fraction epsilon. Model 2 (Section 4.2.2) gives epsilon = 0.35% by assuming 20 years of pure diffusion of radon and its decay products through 4 m of quiescent liquid scintillator in the chimney. If convection, thermal gradients, or mechanical stirring during calibration deployments transports radon through the chimney, epsilon can exceed 2.5% and the 0.005 Bq detector budget is violated. The no-convection assumption is therefore load-bearing, and the actual safety margin is far smaller than the factor of 100 implied by comparing 8 mBq/m^3 with 0.2 mBq/m^3.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes the design, installation, and commissioning tests of the calibration house for the JUNO central detector, focusing on mechanical interfaces, glove boxes, cable routing, a nitrogen flush system, and control of oxygen and radon concentrations. The authors report that the oxygen concentration can be maintained below 5 ppm, the radon concentration in the flushed configuration is [2,15] mBq/m3 at 90% C.L., and the total onsite leak rate is below 2.9e-5 mbar L/s, meeting the design requirement of 4e-5 mbar L/s. To connect the measured house radon concentration to the detector background budget, the paper introduces two diffusion models that yield conversion factors of about 1% and 0.35%, and then compares the measured concentration times the 1% factor (0.08±0.04 mBq/m3) with a derived requirement of 0.2 mBq/m3.","tokens_in":10661,"tokens_out":8100,"duration_ms":82496,"significance":"The manuscript is a useful engineering and commissioning report for an auxiliary system of JUNO. Its strengths are the direct measurements: oxygen below 5 ppm with a calibrated sensor, radon measured with a dedicated detector and reported with a proper confidence interval, and leak checks performed with calibrated instruments meeting the stated limits. The glove-box functionality tests and the cable-distribution validation are also concrete and reproducible. If the radon-budget analysis is clarified and the sensitivity of the conversion factor is properly quantified, the paper will serve as a solid reference for the JUNO collaboration and for similar low-background calibration facilities.","major_comments":[{"comment":"The radon concentration requirement is stated as 20 mBq/m3 in Section 2, but Section 4.2 derives a requirement of 0.2 mBq/m3 (0.005 Bq divided by the 25 m3 house volume). These two numbers differ by a factor of 100, and the manuscript does not explain which is the actual design requirement or how they are related. If the 0.2 mBq/m3 value is intended as the requirement on the effective concentration after the transfer fraction, the text must say so explicitly; otherwise the reader cannot tell whether the measured 8±4 mBq/m3 meets the requirement directly or only after applying the conversion model.","section":"Section 2 and Section 4.2"},{"comment":"The comparison of 0.08±0.04 mBq/m3 with 0.2 mBq/m3 is dimensionally consistent only if the 0.2 mBq/m3 is understood as the allowable product C_house × ε (with ε the transfer fraction). The resulting margin is a factor of about 2.5, not the factor of 100 that a casual comparison of 8 mBq/m3 with 20 mBq/m3 would suggest. The authors should present this margin explicitly and discuss how the conclusion changes if ε is larger than the nominal 1% value.","section":"Section 4.2.3"},{"comment":"The radon budget rests on the assumption that the 4 m of liquid scintillator in the chimney is quiescent with no convection during data taking and calibration deployments. Calibration operations involve moving cables, source deployment, and possible thermal gradients, all of which could mix the LS and increase the transfer fraction beyond the 1% or 0.35% assumed. The paper should either justify the no-convection assumption for the actual operating conditions or provide a conservative upper bound on ε that accounts for mixing. In addition, the erfc solution in Eq. (4.1) is derived for a stable species and does not explicitly account for the radioactive decay of 222Rn; the authors should clarify how decay is treated and whether the model over- or under-estimates the radon reaching the detector.","section":"Section 4.2.2"}],"minor_comments":[{"comment":"The abstract states the radon concentration can be kept below 15 mBq/m3; this is the upper end of the 90% C.L. interval [2,15] mBq/m3, not the central value. Consider reporting the central value and interval for precision.","section":"Abstract"},{"comment":"The phrase 'equivalent to 0.2 mBq/m3' should state that this is obtained by dividing the 0.005 Bq budget by the internal volume of the calibration house (25 m3), so that the reader understands the implicit assumption behind the equivalence.","section":"Section 4.2"},{"comment":"The surface-area model assumes that every radon atom colliding with the LS surface in the chimney is absorbed and enters the CD, while collisions with the house walls lead to reflection. This is a strong assumption about the sticking coefficient; a brief justification or a reference to a similar treatment would strengthen the model.","section":"Section 4.2.1"},{"comment":"The diffusion coefficient from reference [21] is given without context. State its measurement conditions and whether it applies to radon in the JUNO liquid scintillator at the relevant temperature.","section":"Section 4.2.2"},{"comment":"The measured radon concentration after flushing is reported as 8±4 mBq/m3, with a relative uncertainty of 50%. It would be informative to also report the 90% C.L. upper limit (15 mBq/m3) in the budget comparison, since the central value alone may underestimate the risk.","section":"Section 4.2.3"},{"comment":"There are a number of typographical and grammatical errors, for example 'large' in Section 3.1 and 'the CLS SS cable jump out' in Table 1. A careful proofreading pass would improve readability.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript contains a clear inconsistency between the 20 mBq/m3 requirement in Section 2 and the 0.2 mBq/m3 derived in Section 4.2. This needs to be resolved before the paper can be accepted, because it directly affects whether the headline claim about radon control is supported. The diffusion-model dependence of the radon budget is also a concern that should be addressed with additional justification or sensitivity studies."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe JUNO calibration house paper is a solid, straightforward technical report. What's new is the first detailed description of the house itself: the interface design, the five nitrogen flush layouts compared empirically, and the onsite oxygen/radon/leak measurements. The flush layout comparison using oxygen as a tracer is a nice piece of practical engineering, and the leak check strategy (SF6 sniffing for feedthroughs, double O-ring pressure drop for flanges) is sensible. The measured oxygen <5 ppm and radon [2,15] mBq/m3 support the headline claims.\n\nThe soft spot is the radon budget arithmetic in Section 4.2.3. The paper compares 0.08 mBq/m3 (8 mBq/m3 times 1%) against a 'requirement' of 0.2 mBq/m3, but that requirement was derived by dividing the 0.005 Bq detector budget by the 25 m3 house volume—which is only valid if all radon in the house enters the CD. The correct check is C_house * epsilon * V_house < 0.005 Bq. With C = 8 mBq/m3, epsilon = 1%, V = 25 m3, you get 0.002 Bq, which passes but only by a factor of 2.5, not by the large margin the presentation implies. The paper's own arithmetic shows the same factor of 2.5, so the stress-test's claim of a hidden factor of 100 is overstated, but the dimensional muddle should be fixed in revision. More importantly, the margin is thin enough that the no-convection assumption becomes load-bearing. If mixing or convection in the chimney raises the effective transfer fraction above ~2.5%, the 0.005 Bq budget is broken. The authors do state the assumption explicitly, and both models (1% geometric, 0.35% diffusive) are presented as simple estimates, so I don't think this is fatal—but a sensitivity study around epsilon would strengthen the claim considerably.\n\nOverall: the paper is honest, clearly written, and the measurements look real. For a JINST-style technical note, it's a solid accept after minor revision. I'd want the radon budget section rewritten to get the units straight and to acknowledge the margin is only a factor of ~2.5. The calibration house itself appears to work as designed.\n\nRecommendation: send to peer review; accept with minor revisions.","headline":"A solid technical report on the JUNO calibration house; the radon budget has a units bug and a thin safety margin, but the measurements and design work are real.","tokens_in":11211,"tokens_out":5197,"would_cite":false,"duration_ms":49636,"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":"This paper establishes that the JUNO calibration house, a sealed stainless-steel chamber connecting calibration equipment to the central detector, maintains oxygen below 10 ppm and radon below 15 mBq/m³, with a total leak rate below…","keywords":["calibration house","JUNO","radon concentration control","oxygen concentration control","nitrogen flush","leak rate","liquid scintillator","background suppression"],"falsifier":"Measure the radon activity in the liquid scintillator near the bottom of the chimney during steady-state operation and compare it with the radon concentration in the calibration house; if the ratio exceeds the 0.35% prediction by more than the measurement errors, the diffusion-only model is wrong and the house requirement would need to be tightened.","tokens_in":10121,"feed_emoji":"☢️","tokens_out":5276,"duration_ms":53934,"temperature":0.7,"pith_summary":"This paper reports on the calibration house, a sealed stainless-steel chamber that connects calibration equipment to the central detector of the JUNO neutrino experiment. The paper shows that the house meets its design requirements: oxygen concentration can be kept below 10 ppm, radon concentration below 15 mBq/m³, and total leak rate below 2.9×10⁻⁵ mbar·L/s. Onsite installation confirmed that the mechanical and electrical interfaces work with the calibration sub-systems. The authors also present two radon diffusion models that convert the measured house radon concentration into an expected background contribution to the detector, concluding the contribution is within the 0.2 mBq/m³ budget.","feed_headline":"JUNO calibration house holds oxygen below 10 ppm and radon under 15","feed_subtitle":"Sealed chamber with nitrogen flushing keeps the neutrino detector's liquid scintillator clean and background low.","key_machinery":"The load-bearing objects are the nitrogen flush system and the radon diffusion models. The flush system, with a specific inlet/outlet layout (layout 5) that avoids flow shortcuts, reduces oxygen from 20% to below 5 ppm within about an hour. The radon diffusion model 2 uses the complementary error function, $R = \\operatorname{erfc}(h/\\sqrt{4a\\tau})$, with $h = 4$ m of liquid scintillator in the chimney, $a = 1.49\\times10^{-9}$ m²/s, and $\\tau = 20$ years, giving $R \\approx 0.35\\%$ as the fraction of house radon that reaches the detector. This factor, applied to the measured radon concentration, translates the house level into a detector background.","core_discovery":"The central claim is that the calibration house, as built, meets JUNO's requirements for background and leak tightness. The evidence includes nitrogen flush tests showing oxygen falls below 5 ppm, radon measurements with a customized detector showing concentrations as low as 8±4 mBq/m³ with nitrogen flush, a glove box that passed all operational tests, and an onsite leak check of 21 flanges plus feedthroughs giving a total leak rate below 2.9×10⁻⁵ mbar·L/s. The paper further argues, using a diffusive transport model through 4 m of quiescent liquid scintillator in the chimney, that the radon entering the central detector is about 0.35% of the house concentration, so the measured house level corresponds to 0.08±0.04 mBq/m³, below the 0.2 mBq/m³ budget.","pith_inferences":["The 0.35% diffusion factor is the largest leverage on the background budget; if even mild convection exists in the chimney or during detector filling, the factor could be closer to the model-1 estimate of 1%, still within the budget but with a thinner margin.","The nitrogen-flush layout-selection method, using multiple oxygen sensors to find a layout that avoids flow shortcuts, could be applied to other large detector volumes with complex internal geometry.","The measurement that dust adds 22±15 mBq/m³ suggests that maintaining cleanliness after installation is as important as selecting low-radioactivity materials; periodic re-cleaning or radon monitoring might be warranted.","If the radon diffusion coefficient in liquid scintillator were measured under flow or temperature-gradient conditions, the 20-year integration time assumption could be replaced with a more realistic operating scenario."],"forward_implications":["JUNO can deploy the four calibration sub-systems without compromising the liquid scintillator's oxygen and radon budgets.","The double O-ring nitrogen pressure-drop leak check method is suitable for large thin-wall chambers that cannot be vacuum-tested.","The radon contribution from the calibration house is small: 0.08±0.04 mBq/m³, about 40% of the 0.2 mBq/m³ budget.","Cleanliness of the chamber matters: dust contributed 22±15 mBq/m³ before cleaning, comparable to the contribution from the calibration sub-systems themselves.","Glove box operations can be performed without detectable oxygen ingress during typical 30-minute interventions."],"supporting_citations":[{"why":"Private communication that sets the required overall leak rate of less than 4×10⁻⁵ mbar·L/s at 1500 Pa gauge pressure.","marker":"[15]"},{"why":"JUNO radioactivity control strategy that provides the radon and oxygen limits and the 0.2 mBq/m³ budget used for the calibration house.","marker":"[16]"},{"why":"Thermodynamics textbook that supplies the diffusion model used in equation 4.1.","marker":"[20]"},{"why":"Study of radon diffusion behavior in liquid scintillator that provides the diffusion coefficient a = 1.49×10⁻⁹ m²/s.","marker":"[21]"},{"why":"PandaX-4T background control paper referenced for the customized radon detector used in the measurements.","marker":"[22]"},{"why":"Development of high-sensitivity radon emanation measurement systems that gives the detection method and the 5% systematic uncertainty.","marker":"[23]"},{"why":"Feldman-Cousins unified approach used to compute the 90% confidence level limits on radon concentrations.","marker":"[24]"}],"fun_headline_variants":["JUNO calibration house hits radon and oxygen targets with N2 flush","Calibration house meets JUNO's leak and radon requirements","JUNO house: oxygen 5 ppm, radon 8 mBq/m³ after nitrogen flush","JUNO calibration house cuts detector radon to 0.08 mBq/m³"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The radon entering the detector is assumed to travel only by diffusion through 4 meters of still liquid scintillator in the chimney, with no convection and with radon and its decay products diffusing at the same rate; if the liquid moves or mixes, the true radon entry could be higher than the 0.35% factor used.","fun_headline_variants_meta":{"raw":{"variants":["JUNO calibration house hits radon and oxygen targets with N2 flush","Calibration house meets JUNO's leak and radon requirements","JUNO house: oxygen 5 ppm, radon 8 mBq/m³ after nitrogen flush","JUNO calibration house cuts detector radon to 0.08 mBq/m³"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000208,"raw_usage":{"total_tokens":1361,"prompt_tokens":857,"completion_tokens":504,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":412}},"tokens_in":473,"tokens_out":504,"duration_ms":5227,"temperature":1.0,"reasoning_tokens":412,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:00:54.453121+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the radon activity in the liquid scintillator near the bottom of the chimney during steady-state operation and compare it with the radon concentration in the calibration house; if the ratio exceeds the 0.35% prediction by more than the measurement errors, the diffusion-only model is wrong and the house requirement would need to be tightened.","supporting_citations":[{"cited_title":"Zhao,New Concept Physics Course: Thermodynamics (Second Edition), Higher Education Press (2005)","cited_arxiv_id":null,"evidence_quote":"Thermodynamics textbook that supplies the diffusion model used in equation 4.1."},{"cited_title":"Research of radon diffusion behavior in liquid scintillator","cited_arxiv_id":"2301.06982","evidence_quote":"Study of radon diffusion behavior in liquid scintillator that provides the diffusion coefficient a = 1.49×10⁻⁹ m²/s."},{"cited_title":"Development of High-Sensitivity Radon Emanation Measurement Systems with Surface Treatment Optimization","cited_arxiv_id":"2503.00739","evidence_quote":"Development of high-sensitivity radon emanation measurement systems that gives the detection method and the 5% systematic uncertainty."},{"cited_title":"Feldman and R.D","cited_arxiv_id":null,"evidence_quote":"Feldman-Cousins unified approach used to compute the 90% confidence level limits on radon concentrations."}],"review_version":1}