{"id":"08b75ad4-26a6-4507-8687-c92a873aacd5","arxiv_id":"1908.09882","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"In the SLIM model, lepton flavour violation experiments like MEG exclude most of the parameter space that fits the dark matter relic density, making electron recoil detection hopeless.","lead":"The paper checks an existing dark matter model, where a light right-handed neutrino is the dark matter, against lepton flavour violation and direct detection experiments. It finds the lepton flavour violation limits already rule out most of the model, and that electron recoil signals are far too small to see.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The real-θ Casas–Ibarra scan misses complex-angle cancellations that can suppress μ→eγ, so the MEG-based upper bound σ_e ≤ 10^-52 cm^2 may not cover the full SLIM parameter space.","rationale":"The reader's weakest assumption was the normal hierarchy with a massless lightest neutrino. That is a legitimate caveat, but it likely changes the LFV and cross-section bounds by a factor of order one to a few and does not threaten the central conclusion. The more serious issue is the unstated restriction to a real Casas–Ibarra angle. In the two-right-handed-neutrino seesaw, the R matrix is complex orthogonal in general. Complex θ (equivalently, relative phases in λ8) can exponentially enhance or suppress individual Yukawa couplings while preserving the neutrino mass matrix through Eq. (3.6). Because the μ→eγ amplitude is a sum of two terms with different loop functions for the MeV and GeV right-handed neutrinos, a cancellation can make BR(μ→eγ) arbitrarily small even for |λ_e1| values that the paper's real-θ scan would exclude. If such cancellations exist, the statement that MEG currently excludes λ_e1 > 6×10^-3 (Sec. 5) and the resulting σ_e ≤ 10^-52 cm^2 upper bound (Sec. 6.1) are not valid for the full parameter space. The paper's broad conclusion that XENON1T cannot observe this DM is less affected, because even the larger pre-LFV cross sections around 10^-46 cm^2 lie far below the estimated sensitivities in Fig. 9. However, the paper's main new quantitative result—that LFV experiments severely constrain the SLIM parameter space—would need to be rederived with a complex R or explicitly restricted to the real-θ slice. The proposed scan over z = x + i y is a concrete, computationally straightforward check that would settle whether the concern lands. Therefore the appropriate verdict is CONDITIONAL: accept only after the complex-angle slice is included or the LFV claims are appropriately qualified.","tokens_in":16673,"tokens_out":17023,"duration_ms":185434,"concrete_test":"Rerun the Sec. 3.3 scan replacing Eq. (3.9) by R = [[0, cos z, sin z], [0, -sin z, cos z]] with z = x + i y, x ∈ [0,2π], y ∈ [0,3], keeping all other scan ranges and imposing the same neutrino masses and mixings. For each point, compute λ8 from Eq. (3.11), the relic density as in Sec. 4, and BR(μ→eγ), BR(μ→3e), and CR(μ-Ti → e-Ti) with SPheno. Record the maximal |λ_e1| among points that satisfy the Planck relic density, perturbativity (|λ8| < 4π), and all current LFV limits. If this maximum exceeds 6×10^-3, the MEG exclusion and the σ_e ≤ 10^-52 cm^2 upper bound in Sec. 6.1 fail for the full parameter space; if no such points pass all constraints, the concern is settled and the current claims stand.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing assumption is not the mass hierarchy but the restriction to a real Casas–Ibarra rotation. In Sec. 3.3, R is written in Eq. (3.9) with real cos/sin and θ ∈ [0,2π]. For two right-handed neutrinos, the general solution of Eq. (3.8) has a complex orthogonal R; the complex angle (or relative phases in λ8) changes the size of the Yukawa couplings without changing Mν = λ8^T M λ8. The μ→eγ amplitude in Sec. 5 is A = Σ_i λ_ei λ_μi^* f(m_Ni^2/m_η^2); because m_N1 ≈ MeV and m_N2 ≈ GeV have different loop functions, this two-term sum can vanish by cancellation even when |λ_e1| is O(0.1). The plotted lower bound BR(μ→eγ) ≳ 10^-17 and the statement that MEG already excludes λ_e1 > 6×10^-3 are then artifacts of the real-θ slice. If a complex R is allowed, the LFV constraints of Sec. 5 need not reduce the electron-recoil cross section to 10^-52 cm^2; the model may sit at the larger, still-undetectable values near 10^-46 cm^2 of Fig. 8. Thus the quantitative LFV claim, which is the paper's main new result, is not yet demonstrated over the full model parameter space. The qualitative conclusion that XENON1T cannot see this DM may still hold because σ_e remains bounded by the LEP mass and relic density, but the MEG-based bound is conditional on the real-θ restriction.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a scotogenic/SLIM extension of the Standard Model containing two right-handed Majorana neutrinos (a light DM candidate N1 and a heavier N2) plus a complex scalar doublet and singlet. It imposes collider, cosmological and neutrino mass/mixing constraints, uses the Casas-Ibarra parametrization to fix the Yukawa couplings λ8, computes the DM relic density with micrOMEGAs, and then studies one-loop lepton flavour violation (μ→eγ, μ→3e, μ-Ti→e-Ti) and the DM-electron recoil cross section for XENON1T and near-future detectors. The central quantitative claims are that the correct relic density forces sizeable couplings to charged leptons, that MEG already excludes |λ_e1| larger than about 6×10^-3, that lepton flavour violation reduces the electron recoil cross section to at most about 10^-52 cm^2 (with at most about 10^-46 cm^2 even before LFV constraints), and that the model is therefore unobservable in current and near-future electron-recoil searches.","tokens_in":17064,"tokens_out":7945,"duration_ms":80148,"significance":"If the central claims hold, this is a useful and falsifiable study: it closes a gap in the SLIM literature by imposing the Planck relic density and neutrino mixing data simultaneously, and it identifies lepton flavour violation as a much stronger constraint than previously appreciated. The detailed XENON1T sensitivity estimate, including the S2-only analysis and realistic threshold/background assumptions, is also a valuable comparison for the sub-GeV dark matter community. Strengths of the paper are its use of public numerical tools (SARAH, SPheno, micrOMEGAs), the explicit parameter ranges in Table 1, and the honest evaluation of irreducible backgrounds and energy thresholds. The qualitative conclusion that the electron recoil signal is far below experimental reach is robust to the issue raised below; the quantitative LFV exclusion is not yet demonstrated over the full model parameter space.","major_comments":[{"comment":"The parametrization is restricted to a real orthogonal matrix R with θ ∈ [0, 2π]. For two right-handed neutrinos and one massless active neutrino, the general solution of Eq. (3.8) is a complex 2×3 matrix (equivalently a complex 2×2 rotation in the non-zero block); the complex angle changes the magnitudes and relative phases of the entries of λ8 without changing the light-neutrino mass matrix. Since the μ→eγ amplitude in Sec. 5 is A = Σ_i λ_ei λ_μi^* f(m_Ni^2/m_η^2), and f(m_N1^2) differs from f(m_N2^2), the two terms can interfere destructively for a complex R, so |λ_e1| can be O(0.1) with BR(μ→eγ) below the MEG bound. The statements that MEG excludes λ_e1 > 6×10^-3 and that LFV limits the electron recoil cross section to at most 10^-52 cm^2 are therefore not demonstrated on the full parameter space. Please extend the scan to a complex Casas-Ibarra angle, or prove that cancellations cannot occur, and re-derive the LFV and recoil limits accordingly.","section":"Sec. 3.3, Eq. (3.9)"},{"comment":"Only the normal hierarchy with m_1 = 0 is implemented, and this restriction is stated as an assumption 'for simplicity'. An inverted hierarchy with m_3 = 0 is also compatible with two right-handed neutrinos and would generically produce different λ8 matrices, different LFV rates, and different electron-recoil cross sections. Since the paper repeatedly refers to a comprehensive scan and to upper bounds on the recoil cross section, the authors should either include the inverted-hierarchy case or explicitly restrict the exclusion statements and the 10^-52 cm^2 bound to the normal hierarchy.","section":"Sec. 3.3, Eq. (3.5)"}],"minor_comments":[{"comment":"The text says that M is a 3×3 diagonal matrix, but with λ8 a 2×3 matrix the sum over the two right-handed neutrinos in Eq. (3.4) requires a 2×2 diagonal matrix; please correct the dimension or clarify the notation.","section":"Sec. 3.3, Eq. (3.6)"},{"comment":"There is a typo in the sentence describing the b→sγ constraint: 'braching ratio' should be 'branching ratio'.","section":"Sec. 4"},{"comment":"The statement that non-zero neutrino masses impose a lower bound BR(μ→eγ) ≳ 10^-17 should specify whether this is a hard bound over all scanned points or only a feature of the real-θ slice shown in Fig. 5; as written it reads like a general theorem.","section":"Sec. 5"},{"comment":"The idealized red dotted sensitivity curve is scaled by exposure only, and the text correctly notes that this neglects the irreducible neutrino background; please state explicitly in the caption that this curve is an idealized projection rather than a limit, and consider separating projected sensitivities from published limits to improve legibility.","section":"Sec. 6.2 and Fig. 9"}],"recommendation":"major_revision","confidential_remarks":"The core model and numerical setup largely follow the authors' earlier work, but the LFV and electron-recoil analysis is a genuine new step. The main issue is the real-θ Casas-Ibarra slice: if complex angles allow cancellations in μ→eγ, the paper's headline exclusion (MEG bound on λ_e1 and the 10^-52 cm^2 recoil ceiling) is not established for the full model. This is fixable by extending the scan, and the qualitative no-signal conclusion for XENON1T would likely survive; hence I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a competent constraints paper for MeV-scale right-handed neutrino DM in the SLIM/scotogenic framework. The genuinely new content is the numerical relic-density scan with Casas-Ibarra neutrino constraints and the first LFV analysis (mu->e gamma, mu->3e, mu-e conversion) for this model. The main quantitative result is that MEG already cuts the lambda_e1 coupling above a few 10^-3 and, if applied, pushes the electron-recoil cross section below 10^-52 cm^2. The paper also does a careful job on XENON sensitivities; the conclusion that this model is invisible to electron recoil is robust because it is driven by the LEP bound m_eta+ > 98.5 GeV, which enters to the fourth power.\n\nCredit where due: the authors use public codes (SARAH/SPheno/micrOMEGAs), the scan is described well enough to reproduce, and the LFV plots are physically sensible because neutrino masses force a lower floor on lambda8. The abstract's BBN sentence is stronger than what the scan actually shows--most viable models have m_N1 above a few MeV, but the scan allowed lower masses--and no scan code or data files were released. Those are minor.\n\nThe real soft spot is the one the stress test flags. In Sec. 3.3 the Casas-Ibarra rotation is taken real (Eq. 3.9). With two right-handed neutrinos, the general solution of Eq. (3.8) allows a complex 2x2 orthogonal block, and that complex phase can make the mu->e gamma amplitude vanish by cancellation between the N1 and N2 loop contributions. The claimed MEG bound and the resulting sigma_e <= 10^-52 cm^2 are therefore only established on a real-theta slice, not over the full lambda8 parameter space. This does not kill the paper's main point: because the charged scalar is heavy, electron recoil stays far below current and planned sensitivities even at the 10^-46 cm^2 level. But the headline \"MEG excludes these couplings\" is conditional, and the paper should either extend the scan to complex R or explicitly state the restriction.\n\nThe neutrino-mass assumption (normal hierarchy, massless lightest neutrino) is also a simplification and is not tested; it is standard, so I would call that a minor caveat rather than a flaw.\n\nFor whom: phenomenologists working on sub-GeV DM or LFV in radiative seesaw models will get value here, and the XENON sensitivity discussion is useful reference material. It deserves a serious referee. I would send it to review with a request to address the complex-Casas-Ibarra issue before publication, but I would not desk-reject it. My recommendation: engage with it, ask for a clarification or a rerun with complex R, and if that cannot be done, ask them to soften the MEG-based claim accordingly.","headline":"Useful, mostly sound constraints study for MeV neutrino DM, but the MEG-based electron-recoil bound is only proven for a real Casas-Ibarra angle, so the headline needs a caveat or a complex-angle rerun.","tokens_in":17639,"tokens_out":3118,"would_cite":true,"duration_ms":34497,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper shows that MeV-to-GeV right-handed neutrino dark matter in the SLIM model can reproduce the observed relic density and neutrino masses, but lepton-flavour-violation limits from MEG and SINDRUM push the electron-recoil cross…","keywords":["dark matter","right-handed neutrino dark matter","SLIM model","lepton flavour violation","mu to e gamma","electron recoil","relic density","Casas-Ibarra parametrization"],"falsifier":"A decisive test would be to rerun the full parameter scan with the Casas-Ibarra inversion for a massive lightest active neutrino and for the inverted neutrino ordering. If any point reproduces the observed relic density and neutrino data, passes the LEP and cosmological constraints, and has $|\\lambda_8^{e1}|>6\\times 10^{-3}$ (or an electron-recoil cross section above $10^{-52}\\,\\mathrm{cm}^2$) while respecting the MEG, SINDRUM and SINDRUM II bounds, the paper's central exclusion would be contradicted. On the experimental side, the MEG upgrade to $2\\times 10^{-15}$ on $\\mathrm{BR}(\\mu\\to e\\gamma)$ is the sharpest near-term test: a positive signal would need to be accompanied by the correlated $\\mu\\to 3e$ and $\\mu\\to e$ conversion rates predicted here, while a null result would close most of the remaining window.","tokens_in":16460,"feed_emoji":"🌌","tokens_out":19972,"duration_ms":179619,"temperature":0.7,"pith_summary":"The paper asks whether right-handed neutrinos with masses of a few MeV to a few GeV can be the dark matter, in a minimal extension of the Standard Model called the SLIM model, where one scalar singlet and one scalar doublet generate neutrino masses in a loop. It finds that matching the observed dark-matter relic density and the measured neutrino mass differences and mixings forces the dark matter's couplings to charged leptons to be sizeable, between a few $10^{-4}$ and a few $10^{-1}$. Those same couplings inevitably generate lepton-flavour-violating processes such as $\\mu\\to e\\gamma$, and the current MEG limit already rules out couplings above about $6\\times 10^{-3}$. As a result, the DM-electron recoil cross section reaches at most a few $10^{-46}\\,\\mathrm{cm}^2$ under all low-energy, collider, cosmological and neutrino constraints, and drops to at most $10^{-52}\\,\\mathrm{cm}^2$ once lepton-flavour-violation limits are imposed—far below the realistic sensitivity of XENON1T. The broader point is that lepton-flavour violation, rather than direct detection, is the bottleneck that decides whether this class of MeV neutrino dark-matter models survives.","feed_headline":"Electron signal of this dark matter is 10 orders too faint","feed_subtitle":"Lepton-flavour violation already shrinks the viable space, and the MEG upgrade may nearly close it.","key_machinery":"The load-bearing machinery is the radiative-seesaw connection between neutrino masses and dark-sector couplings. The active neutrino mass matrix is a one-loop sum over the two right-handed neutrinos and the real and imaginary neutral scalars, Eq. (3.4), controlled by the small mass splitting $m_4$; inverting this matrix with the Casas-Ibarra parametrization—a standard inversion that converts measured neutrino masses and mixings into the Yukawa couplings of a radiative seesaw—gives $\\lambda_8 = M^{-1/2} R D_\\nu^{1/2} U_\\nu^\\dagger$, Eq. (3.11), so the same coupling matrix drives relic density, LFV and direct detection. The cross-section side is carried by the non-relativistic electron-recoil formula $\\bar\\sigma_{\\chi e}=\\mu_{\\chi e}^2(\\lambda_8^{e1})^4/(\\pi m_{\\eta^\\pm}^4)$, where the charged scalar mass is bounded by LEP at 98.5 GeV; because this mass enters to the fourth power, the predicted signals are strongly suppressed even when the couplings are large.","core_discovery":"The paper's central finding is that in the SLIM model with right-handed neutrino DM, the same Yukawa matrix $\\lambda_8$ that generates the observed active neutrino masses at one loop also sets the relic abundance, the lepton-flavour-violating rates, and the electron-recoil cross section. The one-loop neutrino mass matrix of Eq. (3.4) is inverted with the Casas-Ibarra parametrization, taking a massless lightest active neutrino and normal ordering, so that the measured PMNS mixing and mass splittings fix $\\lambda_8$ up to the dark-sector masses, the mixing angles, and one free angle $\\theta$. Scanning over these parameters while imposing the observed relic density and the LEP, LHC and cosmological constraints, the viable models have DM masses above a few MeV (consistent with BBN) and DM-lepton couplings between a few $10^{-4}$ and a few $10^{-1}$. Those couplings make $\\mu\\to e\\gamma$, $\\mu\\to 3e$, and $\\mu\\to e$ conversion in titanium unavoidable: neutrino masses alone force $\\mathrm{BR}(\\mu\\to e\\gamma)$ above about $10^{-17}$, the MEG bound of $4.2\\times 10^{-13}$ excludes $|\\lambda_8^{e1}|>6\\times 10^{-3}$, and the planned MEG upgrade at $2\\times 10^{-15}$ would test almost all of the remaining parameter space. For electron recoil the charged scalar mediator enters as $\\bar\\sigma_{\\chi e}=\\mu_{\\chi e}^2(\\lambda_8^{e1})^4/(\\pi m_{\\eta^\\pm}^4)$; with the LEP limit $m_{\\eta^\\pm}>98.5$ GeV, the largest allowed couplings give at most a few $10^{-46}$ cm$^2$, and after the LFV limits at most $10^{-52}$ cm$^2$.","pith_inferences":["The paper's bounds are conditional on the chosen neutrino sector; testing an inverted mass ordering or a non-zero lightest neutrino mass is the natural next calculation, and could shift the $\\lambda_8$ values and the derived cross-section ceilings.","The charged-scalar-mass suppression mechanism is not specific to SLIM: any MeV-to-GeV neutrino dark-matter model with a charged scalar mediator above the LEP bound will inherit the same fourth-power suppression, so the qualitative no-detectable-electron-recoil conclusion should extend to similar radiative-seesaw constructions.","If MEG's upgrade observes $\\mu\\to e\\gamma$, the rate alone will not identify the model; the decisive cross-check is whether $\\mu\\to 3e$ and $\\mu\\to e$ conversion in titanium appear at the correlated rates shown in the paper, which would distinguish this scenario from other sources of lepton flavour violation.","Reaching this model class experimentally would seem to require either sub-keV detectors with large exposure or a way to evade the LEP bound on the charged scalar, neither of which is present in the minimal setup studied here."],"forward_implications":["If the SLIM model with right-handed neutrino dark matter is correct, MEG's current limit already excludes DM-electron couplings above about $6\\times 10^{-3}$, and the upgraded MEG experiment at $2\\times 10^{-15}$ will probe nearly all of the remaining parameter space.","The predicted electron-recoil cross sections, at most $10^{-52}\\,\\mathrm{cm}^2$ after lepton-flavour-violation limits, lie far outside the reach of current and near-future liquid-xenon detectors, including the S2-only XENON1T analysis.","Viable dark-matter masses in the scan automatically sit above a few MeV, so the model is consistent with big-bang nucleosynthesis bounds without extra cosmological tuning.","The three lepton-flavour-violating channels $\\mu\\to e\\gamma$, $\\mu\\to 3e$ and $\\mu\\to e$ conversion are strongly correlated, so a discovery in one channel predicts a narrow range of rates in the other two.","Because the suppression comes from the fourth power of the charged scalar mass, any observable electron-recoil signal would require a mediator lighter than the LEP bound, which this minimal model does not allow."],"supporting_citations":[{"why":"It supplies the Casas-Ibarra inversion used to fix the Yukawa couplings λ8 from measured neutrino masses and mixings.","marker":"[36]"},{"why":"It defines the SLIM model with MeV neutrino DM, including the scalar mass matrix and the mass-splitting conditions that set the scan.","marker":"[25]"},{"why":"It provides the observed relic-density value that the parameter scan is required to reproduce.","marker":"[2]"},{"why":"It sets the current upper limit on BR(µ→eγ) that excludes DM-electron couplings above about 6×10^-3.","marker":"[41]"},{"why":"It gives the upgraded MEG sensitivity of 2×10^-15, used to show that nearly all of the remaining parameter space can be tested.","marker":"[42]"},{"why":"It provides the SINDRUM limit on BR(µ→3e), a second lepton-flavour-violating constraint on the same couplings.","marker":"[43]"},{"why":"It gives the SINDRUM II limit on muon-to-electron conversion in titanium, a third constraint on the same couplings.","marker":"[44]"},{"why":"It sets the LEP lower bound of 98.5 GeV on the charged scalar mass, which enters the electron-recoil cross section with the fourth power.","marker":"[32]"},{"why":"It supplies the XENON1T data and energy-threshold model used as the benchmark for the electron-recoil sensitivity comparison.","marker":"[9]"},{"why":"It gives the XENON1T S2-only analysis with a 0.4 keV threshold, the realistic sensitivity curve the predictions are compared with.","marker":"[63]"}],"fun_headline_variants":["Lepton flavour violation shrinks dark matter space","Electron signal too faint for XENON1T to see","MEG upgrade could nearly close dark matter window","Heavy mediator makes dark matter invisible to detectors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the lightest active neutrino is massless and the neutrino masses follow the normal ordering; if the lightest neutrino has a non-zero mass or the ordering is inverted, the Casas-Ibarra inversion produces different couplings $\\lambda_8$, and the quoted lepton-flavour-violation and electron-recoil limits would have to be recomputed.","fun_headline_variants_meta":{"raw":{"variants":["Lepton flavour violation shrinks dark matter space","Electron signal too faint for XENON1T to see","MEG upgrade could nearly close dark matter window","Heavy mediator makes dark matter invisible to detectors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0002,"raw_usage":{"total_tokens":1488,"prompt_tokens":1171,"completion_tokens":317,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":787,"completion_tokens_details":{"reasoning_tokens":254}},"tokens_in":787,"tokens_out":317,"duration_ms":3704,"temperature":1.0,"reasoning_tokens":254,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:00:17.567599+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be to rerun the full parameter scan with the Casas-Ibarra inversion for a massive lightest active neutrino and for the inverted neutrino ordering. If any point reproduces the observed relic density and neutrino data, passes the LEP and cosmological constraints, and has $|\\lambda_8^{e1}|>6\\times 10^{-3}$ (or an electron-recoil cross section above $10^{-52}\\,\\mathrm{cm}^2$) while respecting the MEG, SINDRUM and SINDRUM II bounds, the paper's central exclusion would be contradicted. On the experimental side, the MEG upgrade to $2\\times 10^{-15}$ on $\\mathrm{BR}(\\mu\\to e\\gamma)$ is the sharpest near-term test: a positive signal would need to be accompanied by the correlated $\\mu\\to 3e$ and $\\mu\\to e$ conversion rates predicted here, while a null result would close most of the remaining window.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the Casas-Ibarra inversion used to fix the Yukawa couplings λ8 from measured neutrino masses and mixings."},{"cited_title":"Arhrib, C","cited_arxiv_id":null,"evidence_quote":"It defines the SLIM model with MeV neutrino DM, including the scalar mass matrix and the mass-splitting conditions that set the scan."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It sets the current upper limit on BR(µ→eγ) that excludes DM-electron couplings above about 6×10^-3."},{"cited_title":"Renga [MEG Collaboration], Hyperfine Interact.239 (2018) 58","cited_arxiv_id":null,"evidence_quote":"It gives the upgraded MEG sensitivity of 2×10^-15, used to show that nearly all of the remaining parameter space can be tested."},{"cited_title":"Bellgardt et al","cited_arxiv_id":null,"evidence_quote":"It provides the SINDRUM limit on BR(µ→3e), a second lepton-flavour-violating constraint on the same couplings."},{"cited_title":"Dohmen et al","cited_arxiv_id":null,"evidence_quote":"It gives the SINDRUM II limit on muon-to-electron conversion in titanium, a third constraint on the same couplings."},{"cited_title":"Abbiendi et al","cited_arxiv_id":null,"evidence_quote":"It sets the LEP lower bound of 98.5 GeV on the charged scalar mass, which enters the electron-recoil cross section with the fourth power."},{"cited_title":"Aprile et al","cited_arxiv_id":null,"evidence_quote":"It supplies the XENON1T data and energy-threshold model used as the benchmark for the electron-recoil sensitivity comparison."}],"review_version":1}