{"id":"1e65eb0d-0593-494f-9a49-a55639599180","arxiv_id":"2608.02740","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Rare muon decays into a positron and two dark matter particles could let MEG-II and Mu3e probe lepton-flavor-violating dark matter interactions at TeV scales, with a radiative decay extending the reach to sub-MeV dark matter.","lead":"This paper studies whether dark matter could appear in rare muon decays that also change muons into electrons. It predicts that the MEG-II and Mu3e experiments could be sensitive to such interactions at the TeV scale, and that the same interactions could have produced dark matter in the early universe.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unquantified muon-lifetime/EW-fit constraint may exclude the low-Λ portion of the claimed reach and the freeze-in overlap; the central claim needs this bound folded in.","rationale":"The reader’s weakest assumption identifies exactly the gap I consider most load-bearing: the paper uses a shape-only analysis that discards total-rate information while an existing, quantifiable constraint from the muon lifetime/electroweak fit is mentioned but never evaluated. My estimate confirms the concern is numerically plausible: the claimed TeV-scale Λ corresponds to LFV branching ratios of 10⁻⁴–10⁻³, which are large compared with the precision of the muon lifetime and would perturb the electroweak fit at the level of several to tens of MeV in M_W. The same low-Λ region is where the light-DM freeze-in curves lie, so the overlap claim is directly at risk. I do not think this requires a rejection, because the paper’s framework is coherent and the gap is addressable by adding the missing constraint; the current CONDITIONAL verdict is appropriate. I agree with the reader’s assessment rather than raising a new concern. The one concrete check that would settle the matter is a quantitative electroweak-fit recast of the μ→eχχ width, overlaid on the paper’s figures.","tokens_in":34339,"tokens_out":19500,"duration_ms":179570,"concrete_test":"Derive the 95% CL upper bound on BR(μ→eχχ) from a global electroweak fit by adding one parameter δG_F for an additional contribution to the muon decay width and fitting to τ_μ, M_W, M_Z, α, m_t, m_H (e.g., with HEPfit or an analytic M_W constraint). Convert to Λ_eff via Λ⁴ = 1/(2 G_F² BR) and overlay this bound on Figs. 4, 6, and 7. If the bound lies above the freeze-in curves for m_χ ~ 0.1–1 MeV, or above the lower edge of the projected MEG-II/Mu3e reach, the central claims must be revised to the surviving region.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The analysis in Sec. 3.1 deliberately normalizes the total decay count to the SM expectation and builds a shape-only χ², discarding total-rate information. But any additional width from μ→eχχ shortens the measured muon lifetime, as the authors acknowledge only in passing in Sec. 5.2 (“constraints on the total muon lifetime as part of the EW fit”) without quantifying it. From Eq. (2.6), Γ_NP/Γ_SM = 1/(2 G_F² Λ⁴); for the vector benchmark Λ_eff = 1 TeV gives BR_NP ≈ 3.7×10⁻³ and Λ_eff = 2 TeV gives ≈ 2.3×10⁻⁴. The measured τ_μ determines G_F to ~10⁻⁶ precision, and while a shift in G_F can be partly absorbed by redefining the Fermi constant, the resulting inconsistency with M_W and other electroweak observables is not negligible. If the EW-fit bound is BR_NP ≲ 10⁻⁴–10⁻³, it excludes the lower part of the projected MEG-II/Mu3e sensitivity bands (Figs. 4 and 7). More importantly, the freeze-in curves in Fig. 6 for m_χ ~ 0.1–1 MeV require Λ around 1–2 TeV, which would sit in the region already disfavored by the lifetime/EW-fit bound—and possibly by the paper’s own rough SN1987A estimate (Λ_μe_V ≳ 2.8 TeV, Eq. 5.6). The paper should quantify this bound and show it in the sensitivity plots before claiming that the LFV operators are unconstrained except by SN cooling.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates the sensitivity of MEG-II and Mu3e to lepton-flavor-violating muon decays into a positron and a dark matter pair, μ+→e+χχ̅, described by dimension-six four-fermion operators. The authors compute the differential decay rates for all Lorentz structures, perform a shape-only analysis of the Michel spectrum (Sec. 3), and show that scales Λ_μe_X of order TeV can be probed for m_χ above about 1 MeV, with the pure right-handed vector case giving the strongest reach. For lighter DM, the radiative decay μ+→e+χχ̅γ at MEG-II is proposed as a complementary channel that restores sensitivity down to the massless limit and breaks the left-handed degeneracy (Sec. 4). The paper also computes freeze-in production of χ through the same LFV operators and finds an overlap between the experimental reach and the parameter region reproducing the observed relic abundance for reheating temperatures below about 10 MeV (Sec. 6). Appendices collect results for scalar and tensor interactions and a flavor-scheme example. The central claims are the TeV-scale laboratory reach and the freeze-in connection.","tokens_in":34741,"tokens_out":21380,"duration_ms":181376,"significance":"If the claimed sensitivities hold, this work identifies a new, phenomenologically motivated use of high-intensity muon facilities: probing LFV interactions with a dark sector at effective scales comparable to, or exceeding, those from direct searches. The paper's systematic enumeration of Lorentz structures, full m_χ-dependent differential rates, and explicit treatment of detector effects are strengths, as is the proposal of the radiative channel to cover the calibration-dominated region. The freeze-in connection is an interesting addition that links rare-decay searches to early-universe DM production. However, the significance is substantially tempered by two issues that are not adequately treated in the manuscript: the unquantified total muon-lifetime/EW-fit constraint on the same operators, and the paper's own SN1987A estimate (Eq. 5.6) which is stronger than the projected reach and would exclude the claimed freeze-in overlap. These must be resolved before the main conclusions can be accepted.","major_comments":[{"comment":"The paper mentions 'constraints on the total muon lifetime as part of the EW fit' without quantifying them. Because the shape-only analysis of Sec. 3.1 explicitly discards total-rate information by fixing the total number of muons to the SM expectation, the existing muon-lifetime bound is essential for assessing the reach. For the vector benchmark, Eq. (2.6) gives Γ_NP/Γ_SM = 1/(2 G_F² Λ⁴), so BR_NP ≈ 3.7×10⁻³ at Λ = 1 TeV and ≈ 2.3×10⁻⁴ at Λ = 2 TeV. The muon lifetime determines G_F with relative precision around 10⁻⁶, and a shift in G_F is constrained by the electroweak fit (e.g., via M_W). A conservative bound of BR_NP ≲ 10⁻⁴–10⁻³ would exclude Λ below roughly 1.3–2.4 TeV, covering most of the projected MEG-II/Mu3e reach in Figs. 4, 5, and 7 and removing the claimed freeze-in overlap. The authors should compute this bound explicitly and show it in the sensitivity plots.","section":"Sec. 5.2"},{"comment":"The authors' own SN1987A estimate gives Λ_μe_V ≳ 2.8 TeV for the vector benchmark. This lower bound is stronger than the projected sensitivities shown in Figs. 4, 5, and 7, and it lies above the freeze-in curves in Figs. 6 and 7 once Eq. (6.8) is corrected as discussed in the next major comment. If this estimate is taken at face value, the central claims that MEG-II and Mu3e 'can probe' this parameter space and that freeze-in through the same operators is testable are internally inconsistent with the complementary bound presented in the same paper. The authors must either provide a more reliable SN calculation that relaxes the bound or reframe the conclusions to state that the experiments probe scales below an existing exclusion. Showing the SN bound in Figs. 6 and 7 is mandatory.","section":"Sec. 5.2, Eq. (5.6)"},{"comment":"The displayed analytical freeze-in formulas contain (10 MeV/T_R)^{5/2} e^{-10 MeV/T_R}. The Boltzmann suppression of the thermal muon abundance should be e^{-m_μ/T_R} with the corresponding power of (m_μ/T_R), since x_R = m_μ/T_R and F_{1→3}(x_R) ≃ (1/3)√(2/π) x_R^{5/2} e^{-x_R} as stated in the text. As written, the exponential suppression is underestimated by a factor e^{-(m_μ−10 MeV)/T_R}, which is about e^{-95.7 MeV/10 MeV} ≈ 10⁻⁴ at T_R = 10 MeV. This is a quantitative error in a central equation. The numerical coefficient (4.73 TeV) should be re-derived with the correct factor, and the claim that the analytical expressions 'provide an excellent approximation to the full numerical solution' should be rechecked after the correction.","section":"Sec. 6, Eq. (6.8)"}],"minor_comments":[{"comment":"The sum in the χ² definition is written with a 'P_i' that should be a summation symbol ∑_i.","section":"Sec. 3.1, Eq. (3.7)"},{"comment":"The sentence 'we will fix keep fixed the Michel spectrum normalization' contains a typo; it should read 'we will keep fixed'.","section":"Sec. 3.3"},{"comment":"The phrase 'since the decay channel considered here is open only for m_χ < m_μ/2' appears in the discussion of the scattering contribution; this should be clarified to avoid confusion.","section":"Sec. 6, after Eq. (6.2)"},{"comment":"The y-axis labels '102 103' are ambiguous; please use superscripts or explicit units (GeV) in the axis title.","section":"Fig. 4 and Fig. 7"},{"comment":"The statement that the trapping regions at low Λ are 'essentially excluded by direct experimental searches' is too strong for the muon operator; the NA64 limit Λ_μμ ≳ O(10 GeV) only excludes a small part of the trapping interval, which extends up to several TeV.","section":"Sec. 5.2, Tab. 1"},{"comment":"The phrase 'comparable to astrophysical constraints from supernova cooling' is imprecise given Eq. (5.6), which gives a bound stronger than the projected reach; the comparison should be quantified.","section":"Introduction and Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper relies heavily on Ref. [83] (which includes one of the authors) for the flavor-diagonal SN bounds and on Ref. [7] for the MEG-II analysis framework. The internal inconsistency between the projected reach and the SN estimate of Eq. (5.6) is concerning and should be a central point of the revision. If the lifetime and SN bounds indeed exclude the claimed parameter space, the main conclusions of the paper would need to be substantially reframed. The work is within the scope of the journal and the technical machinery is otherwise sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new content here is the complete Lorentz-structure scan of μ→eχχ, the radiative μ→eχχγ channel as a handle for light DM, and the freeze-in connection with T_R below ~10 MeV. The rate calculations, benchmark definitions, and the shape-based statistical procedure are clean and presented in enough detail that a reader can reproduce them. The authors also flag the calibration region near the Michel endpoint and the left-handed degeneracy honestly. That is real, credit-worthy work.\n\nThe soft spot is the total muon lifetime. The paper mentions in Sec. 5.2 that LFV operators are constrained by the EW fit, but it never quantifies that bound or shows it in the sensitivity plots. For the vector operator, Γ_NP/Γ_SM ≈ 1/(2 G_F^2 Λ^4), so Λ = 1 TeV gives BR_NP ≈ 3.7×10⁻³ and Λ = 2 TeV gives ≈ 2.3×10⁻⁴. The measured muon lifetime and global EW fit plausibly constrain such an additional width at the 10⁻⁴–10⁻³ level. If that is right, the shape-only projections in Figs. 4 and 7 are over-optimistic for the low-Λ side of the bands, and the freeze-in curves for mχ ~ 0.1–1 MeV at T_R = 6–10 MeV—which sit at Λ ~ 2–4 TeV—may partially lie in an already-disfavored region. The paper's own rough SN1987A bound (Λ_μeV ≳ 2.8 TeV) is also not drawn in the freeze-in figures, even though it is estimated in Sec. 5.2. This is a specific, addressable gap rather than a fatal flaw: the radiative channel for mχ < 1 MeV and the high-mass reach above ~2 TeV can still survive, but the claimed sensitivity and the cosmological overlap need to be re-plotted with the lifetime/EW and SN bounds included.\n\nMinor: the radiative background estimate relies on a simple random-coincidence model; that is fine as a first pass, but the 4% systematic is a guess. The scalar and tensor results in the appendix are a useful completeness check.\n\nThis paper is for particle phenomenologists working on dark matter, lepton flavor, or the MEG-II/Mu3e physics case. It deserves a serious referee: the central method is sound, and the missing bound can be added by the authors in revision. I would not desk-reject it.","headline":"Solid systematic EFT study of μ→eχχ at MEG-II/Mu3e, but it underplays an existing total-muon-lifetime/EW constraint that could cut into its TeV-scale reach and freeze-in region.","tokens_in":35241,"tokens_out":4430,"would_cite":true,"duration_ms":45325,"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":"Rare muon decays may expose dark matter at TeV scales.","keywords":["lepton flavor violation","dark matter","muon decay","Michel spectrum","freeze-in","effective field theory","MEG-II","Mu3e"],"falsifier":"A measurement of the total muon lifetime at the level of $10^{-6}$ that sets $\\mathrm{BR}(\\mu^+\\to e^+\\chi\\bar\\chi)\\lesssim 10^{-6}$ would falsify the claimed O(TeV) reach for vector operators, since at $\\Lambda\\sim 1$\\text{--}$2$ TeV the vector branching ratio is roughly $10^{-3}$ to $10^{-4}$.","tokens_in":34161,"feed_emoji":"⚛️","tokens_out":6267,"duration_ms":52318,"temperature":0.7,"pith_summary":"This paper argues that high-intensity muon experiments can detect dark matter through the lepton-flavor-violating decay $\\mu^+\\to e^+\\chi\\bar\\chi$, even though charged lepton flavor is conserved in the Standard Model except for tiny neutrino-mass effects. The key move is to ignore the total event rate and compare only the shape of the positron energy and angular spectrum, the Michel spectrum, with the Standard Model prediction. The authors show that this shape-only analysis reaches new-physics scales of order TeV for dark matter masses above roughly 1 MeV, and that the radiative decay $\\mu^+\\to e^+\\chi\\bar\\chi\\gamma$ at MEG-II extends sensitivity down to massless dark matter. If correct, the same interactions could also generate the observed dark matter abundance by freeze-in when the reheating temperature is below about 10 MeV, making these experiments probes of both particle physics and cosmology.","feed_headline":"Muon-decay shapes can reveal dark matter at TeV scales","feed_subtitle":"MEG-II and Mu3e could probe flavor-violating dark matter couplings and, for low reheating, explain its abundance.","key_machinery":"The machinery is an effective field theory of dimension-six four-fermion operators $(\\bar e \\Gamma_X \\mu)(\\bar\\chi \\Gamma_X \\chi)$ with $\\Gamma_X\\in\\{1,\\gamma^5,\\gamma^\\mu,\\gamma^\\mu\\gamma^5,\\sigma^{\\mu\\nu}\\}$, normalized by an effective scale $\\Lambda^{\\mu e}_X$. The analysis builds a shape-only $\\chi^2$ from binned positron energy and $\\cos\\theta_e$ distributions, fixing the total muon count to the Standard Model expectation from a calibration phase, which isolates spectral distortion from any overall rate change. For light dark matter, the radiative decay adds the photon energy and opening angles as extra kinematic variables, and the freeze-in calculation uses the Boltzmann equation with decay and annihilation collision terms to connect the same operator to the relic abundance.","core_discovery":"The central claim is that the decay $\\mu^+\\to e^+\\chi\\bar\\chi$, mediated by dimension-six four-fermion operators of all Lorentz structures, can be distinguished from the Standard Model Michel decay by shape alone, provided the dark matter mass exceeds about 1 MeV and the experiment records enough muons. For pure right-handed or vector/axial couplings, the positron angular distribution differs from the Standard Model and yields effective scales $\\Lambda^{\\mu e}_X$ of order TeV; for pure left-handed couplings the signal becomes degenerate with the Standard Model as $m_\\chi\\to 0$, but the radiative decay $\\mu^+\\to e^+\\chi\\bar\\chi\\gamma$ removes that degeneracy and restores sensitivity down to the massless limit. In addition, for reheating temperatures between the big bang nucleosynthesis bound and about 10 MeV, muon decay and $\\mu\\bar e$ annihilation through the same operators can produce the observed dark matter relic abundance via freeze-in, and the parameter region overlaps with the projected reach of MEG-II and Mu3e.","pith_inferences":["A precision measurement of the total muon lifetime at the $10^{-6}$ level would independently test the claimed reach: for vector operators the lepton-flavor-violating branching ratio is roughly $1/(2G_F^2\\Lambda^4)$, so a TeV-scale $\\Lambda$ corresponds to a branching ratio of order $10^{-3}$ to $10^{-4}$, far above the current lifetime precision.","Because the shape-only analysis uses the Michel spectrum as its own background, the same procedure could be recycled for other invisible final states, such as $\\mu\\to e$ plus a light scalar, and for constraining the Lorentz structure of any future signal by comparing energy and angular bins.","A detected signal would make the freeze-in link testable: for reheating temperatures above about 10 MeV the same operator would overproduce dark matter, so a positive observation at MEG-II or Mu3e would either pin the reheating temperature below that scale or require additional suppression of the electron-flavor diagonal coupling."],"forward_implications":["MEG-II and Mu3e can constrain $\\Lambda^{\\mu e}_X$ to $\\mathcal O(1\\text{--}2)$ TeV for dark matter masses above about 1 MeV, a reach comparable to supernova cooling bounds.","The radiative decay $\\mu\\to e\\chi\\bar\\chi\\gamma$ at MEG-II remains sensitive down to massless dark matter and is competitive with, or stronger than, the non-radiative channel in several regions.","For a purely left-handed vector interaction the non-radiative search loses sensitivity as $m_\\chi\\to 0$, but the radiative channel removes this degeneracy.","For scalar interactions the non-radiative channel gives significantly stronger bounds than the radiative one because the scalar radiative signal is suppressed by the small invariant mass of the invisible pair.","If the reheating temperature lies between roughly 4 and 10 MeV, freeze-in through the same lepton-flavor-violating operator can produce the observed dark matter abundance in regions accessible to MEG-II and Mu3e."],"supporting_citations":[{"why":"Supplies the MEG-II setup and current limits that define the detector response and dataset.","marker":"[2, 3]"},{"why":"Describes the Mu3e experiment and its projected sensitivity, which motivate the Mu3e projections.","marker":"[4–7]"},{"why":"Provides the radiative-decay search strategy and background estimate for MEG-II that the radiative analysis adapts.","marker":"[7]"},{"why":"Supplies the detector response and efficiency parametrization used to convolute the theoretical rates.","marker":"[19]"},{"why":"Gives the measured muon polarization P = -0.85 adopted in both experiments.","marker":"[24]"},{"why":"Supplies Mu3e detector performance and energy resolution used for the Mu3e analysis.","marker":"[28]"},{"why":"Provides supernova cooling bounds on flavor-diagonal and off-diagonal operators that are compared with the projected reach.","marker":"[83]"},{"why":"Defines the freeze-in mechanism used to compute the dark matter relic abundance.","marker":"[106]"},{"why":"Gives the BBN lower bound on the reheating temperature used to delimit the viable freeze-in region.","marker":"[108]"}],"fun_headline_variants":["Muon decays reveal dark matter at TeV scales","MEG-II and Mu3e can probe TeV-scale dark matter","Flavor-violating decay unveils dark matter","Muon experiments see dark matter at TeV","Probing dark matter via muon decay shapes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole reach rests on the assumption that the total muon lifetime does not already provide a stronger bound than the spectral shape, because the analysis deliberately ignores the overall event rate in favor of shape only.","fun_headline_variants_meta":{"raw":{"variants":["Muon decays reveal dark matter at TeV scales","MEG-II and Mu3e can probe TeV-scale dark matter","Flavor-violating decay unveils dark matter","Muon experiments see dark matter at TeV","Probing dark matter via muon decay shapes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000181,"raw_usage":{"total_tokens":1315,"prompt_tokens":959,"completion_tokens":356,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":279}},"tokens_in":575,"tokens_out":356,"duration_ms":3783,"temperature":1.0,"reasoning_tokens":279,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:02:27.620192+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A measurement of the total muon lifetime at the level of $10^{-6}$ that sets $\\mathrm{BR}(\\mu^+\\to e^+\\chi\\bar\\chi)\\lesssim 10^{-6}$ would falsify the claimed O(TeV) reach for vector operators, since at $\\Lambda\\sim 1$\\text{--}$2$ TeV the vector branching ratio is roughly $10^{-3}$ to $10^{-4}$.","supporting_citations":[],"review_version":2}