{"id":"e2e3d82f-78c6-44d3-b8a6-219d7c564883","arxiv_id":"2607.10716","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Migdal ionization in a germanium detector can turn reactor-produced sub-MeV dark matter into observable signals, yielding new 95% C.L. limits on the DM–proton cross section for masses 0.01–2.6 MeV.","lead":"This paper proposes using the Migdal effect in germanium detectors to search for dark matter produced in nuclear reactors through the decay of dark photons. It derives new upper limits on dark matter–proton scattering for sub-MeV masses using TEXONO reactor data, independent of cosmological assumptions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The ON–OFF residual used in Eq. (13) likely still contains reactor-antineutrino events, so the χ² fit can absorb them into the DM template and the reported σ̄χp limits lose their stated coverage.","rationale":"I agree with the reader's weakest_assumption. This is not an ad hominem or a consensus dispute; it is an internal statistical issue in the derivation of the central limit. The paper explicitly identifies only 'reactor independent backgrounds' as removed by ON–OFF subtraction, and it provides no estimate or model for the surviving reactor-antineutrino contribution. This matters because TEXONO's low-threshold germanium detector is designed to observe reactor neutrino-electron scattering, so the residual in the 0.3–10 keVee region can be dominated by a known SM background that is being left in the data. Fitting the DM template to this residual can bias the best-fit r and corrupt the Δχ²-based confidence interval, so the claimed 95% C.L. limits are not a well-defined statement about σ̄χp unless the neutrino background is either absent by construction or included in the fit. The concrete test is feasible: inspect the TEXONO data definition and, if needed, rerun the analysis with a neutrino background term. I considered the source-model uncertainty (four hand-picked transitions) as an alternative concern, but that affects normalization rather than the statistical validity of the inference; the residual-background issue is the most load-bearing because it determines whether the probability statement about σ̄χp is meaningful at all. The reader already identified this, and the recommended CONDITIONAL verdict remains appropriate until the check is performed.","tokens_in":11371,"tokens_out":11158,"duration_ms":128945,"concrete_test":"Read TEXONO [92] to determine whether the published 'residual spectrum' is raw ON−OFF or already SM-background-subtracted. If raw, recompute Eq. (13) with an additional reactor-antineutrino ν-e scattering template (standard flux at L=28 m, ~2.9 GW, Ge target) included in the model with a free normalization, and compare the resulting 95% C.L. σ̄χp for a representative point (m_V=3.2 MeV, m_χ=0.1 MeV) with the published value. A shift greater than ~30% or a best-fit r consistent with the neutrino-only hypothesis would indicate the concern lands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is the TEXONO-derived 95% C.L. limits on σ̄χp from Eq. (13). The fit uses the reactor ON–OFF residual as the data vector R_b. ON–OFF subtraction removes reactor-independent backgrounds only; reactor-antineutrino events (ν-e elastic scattering and, at low energies, CEνNS) are present in ON and absent in OFF, so they remain in R_b. The paper says only that 'reactor independent backgrounds are reduced' (Sec. III) and does not add a neutrino-background term to the χ² in Eq. (13). Since the neutrino spectrum is smooth and can be partially absorbed by the r⁴-multiplied Migdal template, the minimization shifts r_min and the Δχ²=2.71 interval; the resulting σ̄χp limit is then a limit on DM plus unmodeled background, not on DM alone. This directly undermines the claimed coverage of the new limit. If the TEXONO public residual was already produced by subtracting the SM reactor-neutrino expectation, this concern does not land; the definition in Ref. [92] must be checked.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new channel for detecting sub-MeV dark matter produced in nuclear reactors: dark photons emitted in neutron-capture nuclear de-excitation decay invisibly to χχ̄, and the resulting χ-nucleus scattering in a germanium detector is detected via the Migdal ionization signal. Using four selected E1 transitions from 238U(n,γ)239U and 10B(n,γ)11B, the authors compute the reactor-produced χ flux at the TEXONO baseline, model the germanium response including quenching and isolated-atom Migdal probabilities, and fit the predicted template to the TEXONO ON–OFF residual spectrum. They derive 95% C.L. upper limits on the reference DM–proton cross section σ̄χp, ranging roughly from 6.7×10^-35 to 8.0×10^-33 cm² for 0.01 MeV < mχ < 2.6 MeV and m_V between 3.2 and 6.9 MeV, explicitly noting that the limit is independent of the cosmological DM abundance.","tokens_in":11658,"tokens_out":6991,"duration_ms":82047,"significance":"If the analysis holds up, this is a useful complementary probe of light dark matter that does not rely on the halo flux or relic abundance. The paper has several strengths: it uses publicly available TEXONO data, builds the signal template from external inputs (cFAC atomic wavefunctions, ENSDF/Mughabghab nuclear yields), and clearly states the main approximations. The parameter scaling is transparent: the signal depends on ε^4 while the cross section depends on ε^2, so the limit conversion via r_lim^2 is straightforward. The paper also honestly lists the limitations of the source model and Migdal treatment. However, two of those limitations — the neutrino content of the ON–OFF residual and the unquantified systematic uncertainties in the source and atomic models — are load-bearing for the central limit claim.","major_comments":[{"comment":"The observable used in the χ² fit is the TEXONO reactor ON–OFF residual spectrum. The paper states only that 'reactor independent backgrounds are reduced by the subtraction' and then treats the residual as the DM signal. In the TEXONO experiment, the ON–OFF subtraction is designed to isolate reactor-correlated events, which include not only any DM-produced signal but also Standard Model reactor antineutrino events (ν-e elastic scattering and, at low recoil energies, potentially CEνNS). Unless the TEXONO residual has already been corrected for the SM neutrino expectation, Eq. (13) is fitting DM plus an unmodeled neutrino background. Because the neutrino spectrum is smooth and the DM template is multiplied by r^4, the χ² minimization can partially absorb the neutrino events into the DM template, shifting r_min and the Δχ²=2.71 interval; the resulting σ̄χp limit then loses its stated 95% co","section":"Sec. III, Eq. (13)"},{"comment":"The source model is restricted to four selected E1 transitions from 238U and 10B. The paper acknowledges that secondary production (γ e⁻ → V e⁻), additional capture isotopes, and other multipolarities are omitted, but it does not quantify how these omissions affect the final limit. For an upper limit, omitting additional sources is conservative in normalization, but the spectral shape can still change the bin-by-bin fit and therefore the limit in a non-conservative direction for particular mχ values. The authors should either include a bracketing estimate — e.g., adding all ENSDF lines above the m_V threshold with their quoted photon yields — or vary the yields within their Mughabghab/ENSDF uncertainties and show the resulting band on σ̄χp. At minimum, the choice of the four lines should be justified as a complete or conservative subset for the 3.2–6.9 MeV mediator mass range.","section":"Sec. II, Eq. (4)"},{"comment":"The Migdal ionization probabilities are computed in the isolated-atom approximation, and the paper explicitly states that germanium crystal-environment effects are neglected. Since the detection channel relies entirely on these probabilities, a systematic uncertainty in their normalization or spectral shape propagates directly into the cross-section limit. The isolated-atom approximation may shift the ionization probability by a non-negligible factor at the low recoil velocities relevant here; the velocity cutoff v_N/c ≥ 10^-4 is a step in the conservative direction, but the atomic model itself is not bracketed. The authors should compare their cFAC-based Migdal probabilities with solid-state calculations available in the literature (e.g., the tight-binding or dielectric-response approaches used in existing sub-GeV DM analyses) and either adopt a conservative envelope or assign a systema","section":"Sec. III, Eq. (8)"}],"minor_comments":[{"comment":"References [92] and [105] are the same paper (TEXONO, PRD 99, 032009). Duplicate citations should be merged or cross-referenced.","section":"References"},{"comment":"The notation 'P(GW)' is used without explaining that P is the reactor thermal power in GW. Define P and N_n consistently.","section":"Sec. II, Eq. (4)"},{"comment":"The phrase 'cF AC' should be 'cFAC' for consistency with Refs. [101,102].","section":"Sec. III, text around Eq. (12)"},{"comment":"The manuscript states that the 'successive onset of each orbital’s threshold... produces the characteristic changes of slope in the Migdal spectrum shown in Fig. 2.' It would be helpful to show the TEXONO residual data points and the best-fit DM template overlaid, e.g., as an inset or a new panel, so the reader can assess the fit quality and the χ² behavior underlying Fig. 3.","section":"Fig. 2 and Fig. 3"},{"comment":"The use of '∓' in the definition of E_min/E_max is correct but a brief parenthetical stating that the minus sign gives E_min could improve readability.","section":"Sec. II, Eq. (5)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a straightforward application of the authors' earlier dark-photon-from-nuclear-de-excitation framework (Ref. [86]) to the Migdal channel. The novelty is incremental but real, and the paper is unusually explicit about its approximations. The main technical risk is the ON–OFF residual interpretation: the TEXONO residual is almost certainly a reactor-neutrino signal, not a background-subtracted null spectrum. If the authors can show that their limits are unaffected by including the SM neutrino expectation, the paper could become publishable after the requested revisions. I also note that the source model's completeness is not demonstrated, although for an upper limit this is likely conservative. I did not find concerns about citation ethics beyond the natural reuse of the authors' own prior work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new step in this paper is putting two known pieces together — reactor-produced dark photons from nuclear de-excitation and Migdal ionization in germanium — and deriving a cosmology-independent limit on sub-MeV dark matter. The resulting 0.01–2.6 MeV exclusion is new and the calculation is mostly standard: the flux model uses four E1 lines from 238U and 10B, the Migdal probabilities come from cFAC atomic wavefunctions in the isolated-atom approximation, and the rate equations are internally consistent. The authors also state their approximations honestly, including the omitted secondary production and crystal effects.\n\nThe load-bearing problem is the TEXONO residual. The paper uses the reactor ON–OFF residual spectrum as the observable without adding a neutrino background term to the χ² in Eq. (13). In a reactor neutrino experiment, ON–OFF subtraction removes reactor-independent backgrounds; it leaves the reactor antineutrino events, which generate ν–e scattering and CEνNS in the same low-energy region as the Migdal template. Minimizing r will then partially absorb that neutrino spectrum into the DM template, shifting r_lim and breaking the stated coverage. The paper says only that ‘reactor independent backgrounds are reduced’ and does not state whether the published residual already subtracts the SM neutrino expectation. That is checkable in Ref [92], but as written this is a correct and serious concern. If the residual is already neutrino-free, the concern disappears; the authors need to show that.\n\nThe other soft spots are lesser. The source model uses four selected transitions with no quantified nuclear-data systematics, and the paper ships no code, binning, or response tables, so the exact limit is not independently reproducible. Both are fixable. The self-citation to the authors' earlier flux paper is fair, since the Migdal step is the new part.\n\nWho gets value: the dark sector and low-threshold detector community. The idea is promising and the limit, if corrected, would complement CMB/BAO and SENSEI. I would not treat the present limit as reliable until the neutrino subtraction is clarified. Send it to peer review, because the concept deserves referee time, but make the residual handling the first thing the referee checks.","headline":"New reactor-DM Migdal search is a good idea, but the TEXONO residual analysis likely mistakes antineutrinos for signal and needs a corrected treatment.","tokens_in":12187,"tokens_out":5453,"would_cite":false,"duration_ms":60356,"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 argues that the Migdal effect—ionization of an atom when its nucleus suddenly recoils—can make reactor-produced sub-MeV dark matter visible to germanium detectors, and uses a reactor ON–OFF residual spectrum to set limits on dark","keywords":["Migdal effect","reactor-produced dark matter","dark photon","kinetic mixing","sub-MeV dark matter","germanium detector","nuclear de-excitation","ON-OFF residual spectrum"],"falsifier":"Compute the expected reactor antineutrino–electron scattering rate in the same 300 eV_ee and above energy bins used for the dark-matter template and add it to the fit; if the neutrino component rivals the Migdal prediction in the bins that drive the chi-square, the extracted cross-section limit is not a clean dark-matter bound.","tokens_in":11204,"feed_emoji":"⚛️","tokens_out":4560,"duration_ms":47036,"temperature":0.7,"pith_summary":"The paper tries to establish that light dark matter made in nuclear reactors, which standard searches miss because nuclear recoils are quenched below detector threshold, can be seen through the Migdal effect: when a dark-matter particle scatters off a germanium nucleus, the sudden recoil can ionize a bound electron, adding detectable electron-equivalent energy. The authors compute the reactor-produced dark-matter flux from neutron-capture de-excitation lines, fold in the germanium detector response, and fit the predicted Migdal signal against a reactor ON–OFF residual spectrum. They report 95% confidence upper limits on the dark-matter–proton cross section for dark-matter masses from 0.01 MeV to about 2.6 MeV. Because the dark matter is generated in the reactor rather than in the early universe, these bounds do not require the particle to make up any particular fraction of the cosmological dark matter.","feed_headline":"Migdal ionization exposes sub-MeV dark matter from reactors","feed_subtitle":"New limits on dark-matter–nucleus scattering for masses 0.01–2.6 MeV, set without assuming dark matter fills the cosmos.","key_machinery":"The load-bearing object is the Migdal effect: the ionization of a bound atomic electron when the nucleus is suddenly accelerated by a dark-matter collision. The paper pairs it with a dark-photon production mechanism in which neutron-capture nuclear de-excitation emits an on-shell vector mediator that decays into a dark-matter pair, producing box-shaped energy spectra; the predicted rate is then fit to the residual spectrum with a chi-square that scales as the fourth power of the kinetic-mixing parameter, since production and detection each scale as the mixing squared.","core_discovery":"For MeV-scale dark photons that decay invisibly into a dark-matter pair, neutron-capture transitions in reactor fuel and shielding emit on-shell mediators whose two-body decay produces box-shaped dark-matter spectra peaking at MeV energies. In a germanium detector, the quenched elastic nuclear recoil from this flux lands below the 300 eV_ee analysis threshold, but the Migdal ionization signal adds electronic energy and populates the observable window. From the reactor ON–OFF residual spectrum, the paper derives one-sided 95% C.L. upper limits on the reference dark-matter–proton cross section, roughly 6.7×10^{-35} to 8.0×10^{-33} cm², for mediator masses 3.2–6.9 MeV and dark-matter masses 0.0","pith_inferences":["If reactor antineutrino events remain in the ON–OFF residual at the level of the predicted Migdal signal, the reported limits could be biased; a fit that explicitly includes the neutrino background would quantify this.","The same Migdal strategy could be applied to other artificial dark-matter sources, such as spallation neutron facilities, where the source geometry and spectral lines are better controlled.","The isolated-atom Migdal probabilities used here may differ in a crystal environment; detector-specific atomic calculations could shift the predicted rates and limits.","Extending the source model beyond the four retained transitions, or including multipolarities and secondary production, would likely tighten or reshape the excluded region."],"forward_implications":["Sub-MeV dark matter from reactors, previously invisible because quenching pushes elastic recoils below threshold, becomes testable through ionized-electron signals in germanium.","The resulting limits complement cosmological and astrophysical bounds and remain valid even if dark matter is only a small fraction of the universe's mass.","Because the signal scales steeply with the kinetic-mixing parameter, even a null residual spectrum gives a meaningful cross-section bound for MeV-scale mediators.","Recent experimental confirmation of the Migdal effect in neutron scattering makes the signal mechanism a grounded rather than hypothetical channel.","A more complete nuclear de-excitation database would change both the normalization and spectral shape of the predicted flux, directly affecting the limits."],"fun_headline_variants":["Migdal effect tightens limits on reactor dark matter","Migdal ionization: a window into sub-MeV dark matter","Reactor dark matter constrained via Migdal effect","Migdal effect sets new limits on light dark matter","Probing reactor dark matter with Migdal ionization"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The ON–OFF subtraction leaves reactor antineutrino events in the residual spectrum, and the analysis uses that residual as the dark-matter observable without modeling or subtracting them, so comparable neutrino rates could be absorbed into the dark-matter template and shift the limit.","fun_headline_variants_meta":{"raw":{"variants":["Migdal effect tightens limits on reactor dark matter","Migdal ionization: a window into sub-MeV dark matter","Reactor dark matter constrained via Migdal effect","Migdal effect sets new limits on light dark matter","Probing reactor dark matter with Migdal ionization"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00064,"raw_usage":{"total_tokens":2747,"prompt_tokens":671,"completion_tokens":2076,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":415,"completion_tokens_details":{"reasoning_tokens":2000}},"tokens_in":415,"tokens_out":2076,"duration_ms":21555,"temperature":1.0,"reasoning_tokens":2000,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T07:11:31.551858+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the expected reactor antineutrino–electron scattering rate in the same 300 eV_ee and above energy bins used for the dark-matter template and add it to the fit; if the neutrino component rivals the Migdal prediction in the bins that drive the chi-square, the extracted cross-section limit is not a clean dark-matter bound.","supporting_citations":[],"review_version":2}