{"id":"5e2063a5-1f0d-4c2c-acbd-b2254813f267","arxiv_id":"2504.20960","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"MINER excludes axion-photon couplings down to about 1e-5 and axion-electron couplings down to about 1e-7 for ALP masses between 3 keV and 10 MeV.","lead":"Using three cryogenic sapphire detectors placed four meters from a research reactor, the MINER collaboration set their first exclusion limits on the coupling of axion-like particles to photons and electrons. The result is a proof-of-principle that small, low-threshold reactor experiments can search for low-mass dark matter candidates.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed limits depend on a reactor-core model that approximates the TRIGA UZrH core as pure thorium (Z=90) and whose flux figure shows thorium capture lines; until the true effective-Z and photon spectrum are used, the derived gaγγ and gaee exclusions are not quantitatively reliable.","rationale":"After reading the preprint in good faith, the strongest claim of a new lab-based exclusion is plausible only if the computed ALP rate is reliable. The weakest link is the reactor-core model used to compute that rate. Section 3.1 correctly describes a TRIGA UZrH reactor, but Section 4.4 replaces the core by pure thorium with Z=90 and Fig. 3's caption refers to thorium-based reactor capture lines. This is not merely a matter of Z=90 versus Z=92; the composition is uranium-zirconium-hydride, whose effective Z for Primakoff production is much lower than 90, and the gamma source spectrum must come from the actual TRIGA model. Because the production rate enters N_s linearly and the limit statistic κ=N_s^2/N_b, even an order-unity misestimate of N_s changes the headline couplings by sqrt(N_s) and can distort the claimed exclusion region. The reader flagged the same assumption; my reading strengthens it by noting the thorium flux caption, which makes the mismatch appear in the source spectrum itself. A rerun with the true core model and element-wise yields is a finite, concrete task and could either restore or overturn the limits. I therefore see no reason to change the conditional verdict: the central claim should not be used in global fits until the core-model dependence is quantified.","tokens_in":12526,"tokens_out":12022,"duration_ms":141821,"concrete_test":"Rerun the signal calculation of Section 4.4 with a validated MCNP model of the actual TRIGA UZrH core. Use element-by-element Primakoff production with Σ_i n_i Z_i^2 (and individual form factors) and the true photon absorption denominator σSM per element, instead of a single Z=90 thorium target. Recompute N_s for gaγγ=1e-5 and ma=1 keV and compare to the pure-thorium value. Then regenerate the Fig. 10 limits if N_s shifts by more than 20%; also show the resulting gamma spectrum to verify that the 233Th capture lines are absent and that the TRIGA fuel lines are present.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that MINER excludes gaγγ ~ 1e-5 and gaee ~ 1e-7. That exclusion is computed from N_s via Eqs. (3)-(5), so the ALP yield must be correct. Section 4.4 states the core is \"approximated as pure thorium (Z=90), representing an average effective atomic number,\" but Section 3.1 identifies the core as a TRIGA reactor fueled by low-enriched uranium in a UZrH matrix. Fig. 3 even labels the simulated flux with 233Th capture gamma lines and \"thorium-based reactors,\" which is inconsistent with a UZrH fuel. For Primakoff production Eq. (2) gives dσ ∝ Z^2 F^2(t); the UZrH core with roughly 1 at% U, 35 at% Zr, and 55 at% H has an n_i Z_i^2-weighted effective Z of about 25, not 90. Substituting Z_eff matters through σ_p/(σ_p+σSM) in Eq. (3): if σSM at MeV energies scales roughly as Z, the signal rate scales approximately as Z_eff rather than Z^2, so the pure-thorium model can overpredict N_s by a factor of 3-4, shifting limits by sqrt. If the source spectrum is also from a thorium model, the error is larger. No justification, composition table, or robustness scan is given. Since κ = N_s^2/N_b in Section 4.5 scales the limits through N_s, the headline exclusions are not supported until this reactor model is validated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a search for axion-like particles (ALPs) using three cryogenic sapphire (Al2O3) detectors operated about 4 m from the 1 MWth TRIGA reactor at Texas A&M University, with 59.5 h of reactor-on data and 163.8 h of reactor-off data. ALP production via the Primakoff and Compton-like processes is calculated from an MCNP photon-flux model, and the expected signal from inverse Primakoff scattering, inverse Compton scattering, and in-flight decay is compared with the measured single-scatter rate using the test statistic κ = N_s^2/N_b. The authors derive 90% exclusion limits down to g_aγγ ≈ 10^-5 GeV^-1 and g_aee ≈ 10^-7 GeV^-1 for ALP masses between 3 keV and 10 MeV, and they project improved sensitivity at the HFIR reactor.","tokens_in":12889,"tokens_out":6250,"duration_ms":64629,"significance":"If the limits are correct, the paper provides a new laboratory constraint in a mass range where reactor experiments are competitive with astrophysical bounds, and it demonstrates that gram-scale cryogenic detectors with ~100 eV thresholds can be used for ALP searches. The experiment has genuine strengths: a low-threshold phonon readout, a well-defined single-scatter selection, and a substantial reactor-off background measurement. However, the quantitative limits rest on two assumptions that need scrutiny: the reactor core is modeled as pure thorium with Z=90, and the background is taken from reactor-off data rather than from a clear on/off statistical treatment. Neither is adequately justified, so the headline exclusions are not yet supported at the claimed precision. The paper does not provide machine-checked code or auxiliary data, but the analysis is simple enough to be reproduced once the core model and statistical procedure are clarified.","major_comments":[{"comment":"The approximation of the TRIGA core as pure thorium with Z=90 is a load-bearing input, not a minor modeling choice. Equation (2) gives dσ_P/dcosθ ∝ Z^2 F^2(t), and Eq. (3) propagates this into the signal N_s; the limits in Fig. 10 then scale through κ = N_s^2/N_b in Sec. 4.5. The actual TRIGA fuel is a UZrH matrix (Sec. 3.1), whose Z_i^2-weighted effective atomic number is around 25-30, not 90. Using Z=90 can therefore overestimate the Primakoff production rate by a substantial factor and shift the exclusion curves by roughly the square root of that factor. The inconsistency is visible in Fig. 3, whose caption identifies capture lines from 'thorium-based reactors' even though the experiment uses a 235U-fueled TRIGA. Please repeat the calculation with the actual fuel composition and with a scan over Z, and show how the limits change; without this, the headline values g_aγγ=1e-5 and g_aee=1e-7 are not quantitatively reliable.","section":"Sec. 4.4, Eq. (2), Fig. 3"},{"comment":"The statistical treatment of reactor-on versus reactor-off data is ambiguous and potentially circular. Section 4.4 states that reactor-off single-scatter events 'serve as a lower limit on the potential ALP-induced event rate,' while Sec. 4.5 computes κ = N_s^2/N_b with N_b from reactor-off data to set the limit. If the reactor-on dataset contains additional reactor-correlated backgrounds (neutrons, gamma flash, activation), then the relevant background for a limit from the on dataset is not the reactor-off rate, and using the off rate can make the limits artificially strong. Please specify the likelihood or test statistic used, report the on and off count rates above threshold with their uncertainties, and include a treatment of systematic errors such as flux normalization, detector mass, and live time.","section":"Secs. 4.3-4.5"},{"comment":"The normalization of the signal formulas needs clarification. In Eq. (3), dNγ/dEγ is described as 'the differential photon flux at the detector' with units time^-1 energy^-1, yet the expression also contains an explicit 1/(4π l_d^2) factor. If dNγ/dEγ is already the flux at the detector position, the geometric factor should not appear; if it is the source emission spectrum, the text should say so. The same issue affects Eq. (5), and the quantity A introduced in the text is not used in the displayed formula. Since this normalization directly sets N_s and hence the limits, please correct the definitions or the equations.","section":"Eqs. (3) and (5)"},{"comment":"The statement that the simulated spectra are 'in good agreement with the findings reported in Ref. [30], thereby validating the exclusion limits' is not an independent validation: Ref. [30] shares multiple authors with the present paper and is the source of the production and detection formulas being applied. Agreement with a prior calculation by the same group checks internal consistency but does not validate the reactor model or the new data. Please compare the predicted flux with a measured reactor gamma spectrum or with an independent MCNP model, or state clearly that no such external validation is being claimed.","section":"Sec. 4.4, Fig. 9"}],"minor_comments":[{"comment":"The y-axis units and the exposure scaling are not stated; please add axis labels and clarify whether the shown rates are per keV per kg per day or per keV per detector.","section":"Fig. 8"},{"comment":"The abstract emphasizes the ~100 eV threshold, but the analysis considers only energies above 3 keV; please state explicitly that the ALP analysis threshold is 3 keV and explain why the lower-energy data remain blinded.","section":"Abstract and Sec. 3.2"},{"comment":"The text contains typos and spacing inconsistencies, including 'data collect' in Sec. 4.3, 'byby thehe' in Sec. 5, and irregular 'o ff' spacing throughout; a careful proofread is needed.","section":"Sec. 4.4"},{"comment":"The coupling g_aee is said to have units of GeV^-1 following Ref. [26], but it is often defined as dimensionless; please state the convention explicitly to avoid confusion.","section":"Eq. (1)"},{"comment":"The caption calls the black curve an 'exclusion limit' while the text says it is obtained from reactor-on data; please clarify whether the black curve is a measured limit or a projected sensitivity curve.","section":"Fig. 10"},{"comment":"The smearing argument quotes the bin width and the baseline resolution, but the fitted σ of the 55Fe Kα peak is not given; please provide the resolution values used for the comparison.","section":"Sec. 4.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is essentially an application of the framework of Ref. [30] by overlapping authors to a new dataset, which is acceptable, but the 'validation by agreement with Ref. [30]' should not be presented as an external check. The main technical risk is the Z=90 reactor-core model; I would ask for a reanalysis with the actual TRIGA core composition before considering acceptance. The overlap with Refs. [30] and [37] should also be disclosed more prominently in the text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is the first ALP exclusion limit from MINER's reactor-on/off data, and that alone earns it a serious referee. The detector work is handled carefully: calibration with 55Fe, single-scatter selection, and on/off run bookkeeping are all described in enough detail to see what was done. The paper is also honest that the sub-3 keV region remains blinded. The limit curve is genuinely new, not just a re-plot of the Ref. [30] projection.\n\nThe soft spot the reader flags is real and is not minor. Section 4.4 approximates the TRIGA core as pure thorium (Z=90) with no justification, and the caption of Fig. 3 even references thorium capture lines, which is inconsistent with a UZrH-fueled core. Primakoff production at the vertex scales as Z^2; the suppression factor sigma_p/(sigma_p+sigma_SM) in Eq. (3) may soften the dependence to something closer to linear in effective Z, but either way a UZrH core with an n_i Z_i^2-weighted effective Z around 25, not 90, could change the predicted N_s by a factor of several. Since the limit flows through kappa = N_s^2/N_b, that translates into a sqrt shift in the coupling bound. The headline gaγγ ~ 1e-5 could move by a factor of two or three, and it could move more if the photon flux model itself is also thorium-based rather than based on the actual TRIGA spectrum. The paper needs a composition table, the actual MCNP flux, and a robustness scan over Z_eff before these numbers go into any global fit.\n\nTwo smaller issues. The on/off background treatment is ambiguous: reactor-off data are described both as the background model and as a \"lower limit\" on any ALP signal, and it is not clearly explained how the reactor-on data enter the kappa test statistic. A short CL_s or likelihood description would fix that. The paper should also plot its curve against existing reactor ALP bounds (CONUS, nu-cleus, TEXONO, and similar), not only astrophysical bounds; the text mentions them but does not make the comparison explicit.\n\nWhat is genuinely good: the low ~100 eV threshold is demonstrated, the setup is described in enough detail to be understood, and the HFIR projection is clearly labeled as a projection, not a measurement. The overlap with Ref. [30] is fine, since the real step here is applying that formalism to actual data.\n\nMy verdict: conditional acceptance. This deserves peer review, and the main flaw is fixable with an honest re-analysis of the reactor core model. I would not cite the numerical limits in my own work until that re-analysis is done.","headline":"First MINER ALP exclusion from reactor on/off data is a real incremental result, but the pure-thorium (Z=90) core approximation is load-bearing enough that the quoted couplings should not be trusted until the reactor model is fixed.","tokens_in":13507,"tokens_out":4548,"would_cite":false,"duration_ms":49526,"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 MINER experiment excludes axion-like particles with couplings as small as $g_{a\\gamma\\gamma}=10^{-5}$ and $g_{aee}=10^{-7}\\,\\mathrm{GeV}^{-1}$ for masses 3 keV to 10 MeV.","keywords":["axion-like particles","reactor neutrino experiment","cryogenic sapphire detector","Primakoff process","axion-photon coupling","axion-electron coupling","low-threshold phonon detector","TRIGA reactor"],"falsifier":"Recompute the expected ALP event rate using the actual elemental composition of the TRIGA fuel-moderator assembly (uranium-zirconium-hydride) and regenerate the exclusion curves; if the $g_{a\\gamma\\gamma}$ limit moves up by roughly $(90/Z_{\\rm eff})^2$, the pure-thorium core approximation is the load-bearing assumption and the published limits are over-optimistic.","tokens_in":12350,"feed_emoji":"⚛️","tokens_out":10406,"duration_ms":94084,"temperature":0.7,"pith_summary":"This paper reports the first axion-like particle (ALP) search from the MINER experiment, which places a stack of three cryogenic sapphire crystals about four meters from the 1 MW TRIGA reactor at the Texas A&M Nuclear Science Center. The detectors run with a roughly 100 eV threshold and a 3 keV analysis floor, and the analysis compares 59.5 hours of reactor-on single-scatter data with 163.8 hours of reactor-off background data. Assuming ALP production through the Primakoff and Compton-like processes in the core and detection through inverse Primakoff scattering, inverse Compton scattering, and decay, the experiment excludes $g_{a\\gamma\\gamma}=10^{-5}\\,\\mathrm{GeV}^{-1}$ and $g_{aee}=10^{-7}\\,\\mathrm{GeV}^{-1}$ for ALP masses between 3 keV and 10 MeV at 90% confidence. The result matters because it shows a gram-scale, low-threshold crystal detector can set laboratory bounds on low-mass ALPs that are normally the province of far larger experiments or astrophysical observations.","feed_headline":"435-gram sapphire stack sets reactor axion limits at 10^-5","feed_subtitle":"Cryogenic detectors at the MINER reactor bound axion-photon and axion-electron couplings from 3 keV to 10 MeV.","key_machinery":"The load-bearing object is the Primakoff process, used in both directions: a photon scatters coherently off a high-$Z$ nucleus in the reactor core ($\\gamma + N \\to a + N$), whose cross-section is enhanced by $Z^2$, and the inverse process produces a photon in the sapphire crystal. The signal rate is the convolution of the MCNP-modeled reactor photon flux, the production cross-section, an ALP survival probability that accounts for decay in flight, and the detection cross-section, with the standard-model photon scattering cross-section (from XCOM) reducing the available production rate. A mass-by-mass $\\kappa = N_s^2/N_b$ hypothesis test with $\\kappa=4.61$ then converts the reactor-on versus reactor-off single-scatter rates into a 90% confidence exclusion curve.","core_discovery":"The central claim is that the low-background environment of the MINER sapphire detectors, combined with the high photon flux at 4 meters from a reactor core, makes the experiment a competitive probe of low-mass ALPs. Using single-scatter events and a $\\kappa = N_s^2/N_b$ test statistic with $\\kappa = 4.61$ (90% confidence), the paper derives exclusion limits on the two couplings of Eq. (1). For $g_{a\\gamma\\gamma}$, the limit is flat at $\\sim 10^{-5}\\,\\mathrm{GeV}^{-1}$ up to $m_a \\simeq 30$ keV, then tightens as the two-photon decay channel becomes efficient; for $g_{aee}$, it reaches $\\sim 10^{-7}\\,\\mathrm{GeV}^{-1}$ with a feature near 1 MeV from pair decay. The calculation treats signals as unsmeared per 0.4 keV bin, justified by the measured 35–101 eV detector resolution, and reproduces the reference spectrum of Ref. [30]. The paper further projects that moving the same detector payload to the 85 MW HFIR reactor and expanding it to five crystals totaling 725 g would reach previously unexplored ALP parameter space, including the QCD axion band.","pith_inferences":["If the effective $Z$ of the TRIGA core is substantially lower than the assumed 90, the true Primakoff production rate falls and the published exclusion limits weaken; a composition-specific simulation would quantify this shift.","The same low-threshold phonon-detector technology could be applied to other intense photon sources, such as medical-isotope or spent-fuel facilities, to probe the same ALP couplings at new sites.","The deferred axio-electric (atomic ionization) channel likely dominates detection for sub-keV ALPs, so extending the analysis below 3 keV may strengthen the electron-coupling limit beyond the values reported here."],"forward_implications":["The current 435 g payload excludes couplings down to $g_{a\\gamma\\gamma}=10^{-5}$ and $g_{aee}=10^{-7}$, establishing MINER as a laboratory probe of low-mass ALPs.","Scaling the same detector tower to five crystals at the 85 MW HFIR reactor is projected to enter previously unexplored ALP parameter space, including masses up to 20 keV for the photon coupling.","The sub-3 keV energy range, currently blinded for the CEvNS analysis, remains a ready extension for even lighter ALPs once unblinded.","At masses below the decay threshold the limit is essentially background-limited, so improvements in shielding and event discrimination directly translate into stronger coupling bounds."],"supporting_citations":[{"why":"Provides the scattering and decay signal-rate formalism used throughout the analysis.","marker":"[30]"},{"why":"Supplies the MCNP-modeled reactor photon flux spectrum and the MINER background model.","marker":"[37]"},{"why":"Gives the Primakoff differential cross-section formula used for ALP production in the core.","marker":"[31]"},{"why":"Provides the same Primakoff cross-section in the form used for the reactor yield estimate.","marker":"[32]"},{"why":"Supplies the standard-model photon scattering cross-section that limits Primakoff production in the core.","marker":"[33]"}],"fun_headline_variants":["Sapphire detectors at reactor squeeze axion couplings to 10^-5","MINER sapphires exclude axion-electron coupling down to 10^-7","Cryogenic sapphire stack rules out low-mass axions near reactor","Reactor's sapphire detectors bound axion-photon coupling to 10^-5"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The signal rate calculation approximates the reactor core material as pure thorium with $Z=90$; since the Primakoff cross-section scales as $Z^2$, a lower true effective nuclear charge would reduce the predicted ALP yield and weaken the exclusion limits by a corresponding factor.","fun_headline_variants_meta":{"raw":{"variants":["Sapphire detectors at reactor squeeze axion couplings to 10^-5","MINER sapphires exclude axion-electron coupling down to 10^-7","Cryogenic sapphire stack rules out low-mass axions near reactor","Reactor's sapphire detectors bound axion-photon coupling to 10^-5"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000806,"raw_usage":{"total_tokens":3587,"prompt_tokens":1041,"completion_tokens":2546,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":657,"completion_tokens_details":{"reasoning_tokens":2461}},"tokens_in":657,"tokens_out":2546,"duration_ms":18904,"temperature":1.0,"reasoning_tokens":2461,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:15:11.959316+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the expected ALP event rate using the actual elemental composition of the TRIGA fuel-moderator assembly (uranium-zirconium-hydride) and regenerate the exclusion curves; if the $g_{a\\gamma\\gamma}$ limit moves up by roughly $(90/Z_{\\rm eff})^2$, the pure-thorium core approximation is the load-bearing assumption and the published limits are over-optimistic.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the scattering and decay signal-rate formalism used throughout the analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the MCNP-modeled reactor photon flux spectrum and the MINER background model."},{"cited_title":"Aloni, C","cited_arxiv_id":null,"evidence_quote":"Gives the Primakoff differential cross-section formula used for ALP production in the core."}],"review_version":1}