{"id":"5a64996b-ce75-4630-98ec-b3317fde5aea","arxiv_id":"2412.14752","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Using Ulysses and SOHO MeV electron data, the authors set the strongest direct upper bound U_e^2 ~ 10^-6 on solar-produced heavy neutral leptons, reaching that value at a mass near 10 MeV.","lead":"This paper searches for electrons and positrons in interplanetary space that would be produced by hypothetical heavy neutrinos created inside the Sun, and finds no convincing excess. It sets the strongest direct limit yet on how strongly such particles can mix with ordinary electron neutrinos, narrowing the options for new physics beyond the Standard Model.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline limit depends on a signal-flux normalization from a simplified 1D diffusion model that the paper itself estimates can be off by half an order of magnitude or more; a factor ≥9 error in flux would erase the claimed factor-3 margin over Borexino.","rationale":"The reader's conditional verdict is appropriate. The central physics — HNL production in 8B decays, the decay to e+e- nu, and the general idea of searching in heliospheric electron data — is sound and the paper is transparent about its main caveat. The decisive weakness is genuinely the signal transport model: Eq. (11) is an approximate analytic solution under assumptions the authors acknowledge are not reliable at the factor-of-a-few level, and the quoted limit has no systematic band. Because the claimed advantage over Borexino is only a factor ~3 in Ue^2, the headline claim survives only if the flux is not overestimated by more than about an order of magnitude. That is a real, concrete, and central vulnerability, but it does not collapse the paper; it makes the result conditional on a numerical transport validation. The background-calibration concern is real but less dangerous, since fitting the background normalization to data that might contain signal makes the limit conservative rather than aggressive. I see no internal mathematical error in the derivation, and the comparison with Borexino is reasonably argued, so the appropriate outcome is to keep the reader's conditional verdict rather than reject or accept outright.","tokens_in":16769,"tokens_out":12205,"duration_ms":117760,"concrete_test":"Run the numerical heliospheric transport code used for the Jovian electron background (Refs. [82,97]) with the HNL decay source q(R,Ee) from Eqs. (7)-(9) for, e.g., MN = 10 MeV, and compare the predicted 1 AU electron flux with Eq. (11) using the same D. Then recompute the 90% CL Ue^2 limit with the numerical flux. If the limit shifts upward by more than the factor separating the paper's curve from the Borexino curve at that mass (about 3 in Ue^2, i.e. a flux ratio of about 9 or more), the headline claim 'strongest direct bound' is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Sec. VI: Ue^2 ~ 10^-6, strongest direct bound in 2-12 MeV) is load-bearing only if the predicted 1 AU electron flux from HNL decays is correct to within roughly a factor of a few. The prediction rests on Eq. (11), which assumes spherical symmetry, a constant D = 2e22 cm^2/s, no convection, and the factorized source Eq. (9). Sec. V B concedes that R-dependent diffusion and convection change fluxes by up to half an order of magnitude, and that anisotropic perpendicular transport (with mean free path up to two orders smaller) can give similar or larger corrections. Since Ue^2 scales as sqrt(flux), a flux overestimate by a factor ~9 (well within the stated range of combined uncertainties) shifts the 90% limit up by a factor ~3, exactly the claimed margin over Borexino. The paper does not propagate these uncertainties into the quoted limit, so the 'strongest direct bound' is not demonstrated at the required precision. Footnote 6's acknowledgment that the Jovian background model was calibrated to data that may contain the signal is secondary, because a free background normalization tends to weaken, not strengthen, an exclusion limit.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper searches for MeV-scale heavy neutral leptons (HNLs) produced by solar nuclear reactions and decaying into e+e-ν, using archival MeV electron/positron spectra from the Ulysses KET and SOHO EPHIN instruments at 1 AU. The signal is computed from standard HNL production and decay rates, propagated to Earth with a simplified one-dimensional steady-state diffusion model, and compared to a Jovian-electron background with a profiled mean free path. The authors report a 90% CL upper limit reaching Ue^2 ~ 1e-6 at MN ~ 10 MeV and claim that this is the strongest direct bound in the 2-12 MeV mass range, about a factor of 3 stronger than the Borexino limit, with a tentative ~3 sigma preference for a signal at MN ~ 8.2 MeV.","tokens_in":16965,"tokens_out":8247,"duration_ms":76515,"significance":"If the propagation-model uncertainties were under control, the paper would provide a genuinely new direct probe of MeV HNL parameter space using existing heliospheric data, with a clean particle-physics input, careful decay kinematics, and a sensible comparison to Borexino and PIENU. The treatment of the Jovian background as a fitted nuisance is a reasonable first step, and the use of two independent datasets is a strength. However, the headline claim inherits an unquantified normalization uncertainty from the transport model, so the significance of the result is at present conditional rather than established.","major_comments":[{"comment":"The central claim of Sec. VI, that the result is the strongest direct upper bound in the 2-12 MeV range, is not supported at the required precision by the signal-flux normalization. The 1 AU flux is computed from the one-dimensional diffusion solution Eq. (11) with a constant D = 2e22 cm^2/s, while Sec. V B itself states that R-dependent diffusion and convection change fluxes by up to half an order of magnitude and that anisotropic perpendicular transport, with mean free path up to two orders of magnitude smaller, can give similar or larger corrections. In the small-mixing regime the electron flux scales as Ue^4, so the 90% limit on Ue^2 scales as the inverse square root of the flux normalization; a flux overestimate by a factor ~9, well within the range of corrections named in Sec. V B, would shift the limit up by a factor ~3 and erase the claimed factor-3 improvement over Borexino. Because these systematic uncertainties are not propagated into Fig. 3 or into the quoted value Ue^2 ~ 1e-6, the paper should present the limit as a band over the allowed transport-model choices, adopt an explicitly conservative flux normalization that accounts for perpendicular transport and convection, or weaken the 'strongest direct limit' claim to be conditional on the transport model.","section":"Sec. V B, Eq. (11)"},{"comment":"The analytic solution Eq. (11) and the source factorization Eq. (9) are asserted rather than derived or validated. Since Eq. (11) is the only bridge between the standard HNL decay calculation and the observed electron flux, the paper should provide the derivation in an appendix and compare Eq. (11) with a numerical solution of Eq. (6) under the stated assumptions, including the finite-volume boundary conditions described in Footnote 4. Without such a validation the reader cannot distinguish a genuine solution from an approximation error that could be comparable to the propagation uncertainties already discussed in Sec. V B.","section":"Sec. III C, Eqs. (9)-(11)"},{"comment":"The background model is not fully independent of the signal hypothesis: Footnote 6 acknowledges that the Jovian-electron modeling was calibrated to data that could include HNL decays, and the analysis then profiles over lambda_parallel for each dataset. This procedure is likely conservative for an exclusion, but the paper should quantify the degeneracy between the signal normalization and lambda_parallel, for example by reporting the fitted lambda_hat_parallel as a function of Ue^2 and comparing with the limit obtained when lambda_parallel is fixed to the value from Ref. [97]. If a signal of similar spectral shape can be partially absorbed by the background normalization, the claimed sensitivity relies on spectral-shape differences that should be demonstrated explicitly.","section":"Sec. V A, Footnote 6, Eq. (13)"}],"minor_comments":[{"comment":"The phrase 'vanishing density at R -> infinity and R0 -> infinity' is unclear, because R0 is a parameter of the source model and not a spatial boundary; the two integration constants and their boundary conditions should be stated explicitly for the physical coordinate R and for the finite-volume radius RV.","section":"Footnote 4"},{"comment":"The notation d(phi_N)/(dR dE_N) is confusing because phi_N in Eq. (5) is already a flux at fixed R, not a cumulative distribution; please define the derivative with respect to R explicitly.","section":"Eq. (7)"},{"comment":"The text 'Ionisation Energy Losstechnique' appears to be a typographical error for 'Ionisation Energy Loss technique'.","section":"Sec. IV A"},{"comment":"The right panel of Fig. 4 and the accompanying text describe a ~3 sigma preference for a signal at MN ~ 8.2 MeV; the caption or text should state explicitly that this preference is not claimed as a discovery and that no trials factor has been applied, so that readers do not over-interpret the simultaneous presentation of an exclusion limit and a best-fit excess.","section":"Sec. V B and Fig. 4"},{"comment":"The definition sigma_bd = f_d * D_bd uses the observed data rather than the predicted flux as the error template; because this can bias the chi-squared in low-count bins, a brief justification or a cross-check with an error based on the model prediction would be helpful.","section":"Eq. (13)"}],"recommendation":"major_revision","confidential_remarks":"I think the paper is promising and the particle-physics input is solid, but the headline bound is currently conditional on an unquantified normalization of the heliospheric transport model. I do not see grounds for rejection: the issue is fixable by presenting the limit as a band over transport-model assumptions, by adopting a conservative normalization, or by softening the 'strongest direct bound' claim. I would also ask the authors to validate Eq. (11) numerically and to clarify the background-signal degeneracy. The 3 sigma preference should be handled carefully to avoid over-interpretation, although the authors already note that incomplete background modeling is a likely explanation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: the new thing here is the detection channel, not the physics idea. Solar HNL production and the N -> e+e- nu decay were already probed by Borexino; what is new is using Ulysses/SOHO MeV electron data as a heliospheric detector and modeling the HNL decay products' propagation with a 1D diffusion equation. That is a legitimate idea, and the paper does several things well: the particle-physics input is standard, the comparison with Borexino is reproduced honestly, and the authors explicitly flag the 3-sigma preference near 8.2 MeV as likely due to incomplete background modeling.\n\nBut the headline claim—'strongest direct upper bound' with a factor-of-three margin over Borexino—is not demonstrated at the required precision. The signal flux at 1 AU rests on an analytic solution, Eq. (11), that is asserted without derivation or numerical validation. The source factorization in Eq. (9) is uncontrolled. The paper itself concedes that R-dependent diffusion and convection change fluxes by up to half an order of magnitude, and that anisotropic perpendicular transport, with mean free path up to two orders smaller, could cause similar or larger corrections. Since the limit scales as the square root of the flux, a flux error of a factor of a few is enough to shift the limit by the same amount as the claimed margin over Borexino. The stress-test note is right: a factor ~9 overestimate in flux would erase the margin entirely. The authors call the limit 'realistic,' but they do not propagate these uncertainties into a band on the exclusion curve.\n\nA secondary concern is footnote 6: the Jovian background model was calibrated on data that may contain the signal. The authors fit the background normalization simultaneously, which is the right thing to do, but it means the exclusion rests on the background shape. That said, this likely weakens rather than strengthens the limit.\n\nOverall, the paper is honest and the particle physics is sound. The soft spot is the transport model, and it is load-bearing. I would send it to peer review—the channel is worth airing—but a referee should require a systematic uncertainty estimate on the propagation model and at least a numerical check of Eq. (11) before the 'strongest bound' claim is allowed to stand. Without that, I would treat the result as a proof-of-principle rather than a definitive limit.","headline":"A clever new heliospheric detection channel for solar-produced heavy neutrinos, but the claimed factor-of-three edge over Borexino sits inside an unquantified transport-model uncertainty.","tokens_in":17589,"tokens_out":3254,"would_cite":false,"duration_ms":26786,"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 Sun's $^8$B neutrino flux would produce MeV heavy neutral leptons that decay to $e^+e^-$ pairs, and a comparison of the predicted heliospheric electron flux with Ulysses and SOHO data yields the strongest direct upper bound on…","keywords":["heavy neutral leptons","sterile neutrinos","solar neutrinos","neutrino mixing","heliospheric electron propagation","Jovian electrons","Ulysses","SOHO"],"falsifier":"A positron-sensitive MeV detector near 1 AU could settle the claim directly: positrons from HNL decays would appear as an excess over a background that is up to three orders of magnitude smaller than the electron background, so a null result after profiling the Jovian background would confirm the exclusion, while an excess matching $M_N \\simeq 8.2$ MeV with $U_e^2 \\simeq 7.9 \\times 10^{-7}$ would confirm the tentative signal. In parallel, multi-spacecraft tracking of Jovian electrons during magnetic connection events could fix $D$ and the parallel mean free path, converting the current half-order-of-magnitude propagation uncertainty into a definite prediction for the 1 AU flux and shifting the limit as the square root of the flux.","tokens_in":16460,"feed_emoji":"☀️","tokens_out":10457,"duration_ms":70857,"temperature":0.7,"pith_summary":"The paper asks whether heavy neutral leptons (HNLs) with masses in the MeV range and a small mixing $U_e^2$ with electron neutrinos could be produced in nuclear reactions in the Sun and then decay into electron-positron pairs inside the heliosphere. It computes the resulting electron and positron flux at 1 AU using a steady-state diffusion model and compares it with MeV electron spectra recorded by the Ulysses and SOHO satellites during quiet solar periods. The absence of an excess in the measured fluxes yields an upper limit that reaches $U_e^2 \\simeq 10^{-6}$ at $M_N = 10$ MeV, which the authors state is the strongest direct upper bound on $U_e^2$ in the HNL mass range $(2\\text{--}12)$ MeV, surpassing Borexino and PIENU. A side result is a $3\\sigma$ preference in the fit for a small signal at $M_N \\simeq 8.2$ MeV, $U_e^2 \\simeq 7.9 \\times 10^{-7}$, which the authors ascribe to incomplete background modeling rather than evidence for HNLs. The result matters because laboratory limits are weak in this mass window and cosmological bounds depend on assumptions about the early universe.","feed_headline":"Solar electrons set tightest bound on MeV heavy neutrinos yet","feed_subtitle":"Ulysses and SOHO data surpass Borexino and PIENU in the 2-12 MeV range, reaching U_e^2 ~ 10^-6.","key_machinery":"The load-bearing object is the steady-state solution of the spherically symmetric diffusion equation for the electron phase-space density $\\psi(R,E_e)$, Eq. (11): with a constant diffusion coefficient $D$, a factorized source term, and no convection or energy losses, the predicted flux follows as $\\psi(R,E_e) = \\xi(E_e)/(DR)\\left(1-e^{-R/R_0} - (R/R_0)\\operatorname{Ei}(-R/R_0)\\right)$. This solution converts the solar HNL production flux, Eq. (5), into the electron/positron flux at 1 AU that is compared with data. The diffusion coefficient $D = 2 \\times 10^{22}\\,{\\rm cm}^2/{\\rm s}$ is taken from Jovian-electron transport analyses, and the background is modeled with the same transport framework, so the same machinery sets both the signal prediction and the dominant background.","core_discovery":"The paper's central claim is that, for a Majorana HNL that mixes only with the electron flavor and has a mass between roughly 1 and 16 MeV, the MeV electron and positron data from Ulysses and SOHO exclude mixing strengths above $U_e^2 \\simeq 10^{-6}$ at $M_N = 10$ MeV at 90% CL. The exclusion is obtained from a $\\Delta\\chi^2$ analysis that fits the signal and a Jovian-electron background model, profiling the parallel mean free path $\\lambda_\\parallel$ as a nuisance parameter, to the two datasets. In the mass range $(2\\text{--}12)$ MeV this is claimed to be the strongest direct bound on $U_e^2$, stronger than the Borexino bound by roughly a factor of three and also excluding a region at larger $M_N$ that Borexino cannot reach. The same data prefer a signal at the $3\\sigma$ level with $M_N \\simeq 8.2$ MeV and $U_e^2 \\simeq 7.9 \\times 10^{-7}$, but the authors interpret this as a likely artifact of astrophysical background uncertainties.","pith_inferences":["The same diffusion-based signal pipeline could be reused for any long-lived particle produced in the Sun that decays into electrons or positrons, since the inputs are just the solar production spectrum and the decay kinematics; the constraints would then apply with rescaled couplings.","The tentative $3\\sigma$ excess is a concrete target: a positron-sensitive measurement with exposure similar to Ulysses or SOHO would either confirm the $8.2$ MeV peak or rule it out, because the positron background is expected to be negligible.","If future heliospheric transport modeling pins down the radial dependence of $D$ and the role of convection, the same datasets could yield a bound several times stronger without any new detector, since the current limit is set by propagation and background uncertainties rather than by counting statistics."],"forward_implications":["Solar $^8$B reactions do not produce an observable electron-positron excess from $N \\to e^+e^-\\nu$ decays for electron mixing $U_e^2$ above the reported line in the $(2\\text{--}12)$ MeV mass window, and the previous direct bounds from Borexino and PIENU are superseded there.","Heavy neutral leptons in this mass range remain viable only if their electron mixing lies below roughly $10^{-6}$, or if they mix predominantly with muon or tau flavor, since the reported bound applies to electron-flavor coupling.","A detector that can separate positrons from electrons near 1 AU, or an ab initio model of the heliospheric electron background, would improve the sensitivity and could test whether the $3\\sigma$ feature at $8.2$ MeV is real.","Because the limit scales as the square root of the predicted flux, each factor-two improvement in the diffusion-modeling uncertainty translates into a factor $\\sqrt{2}$ improvement in the reach in $U_e^2$."],"supporting_citations":[{"why":"Supplies the solar-HNL production flux formula and the Borexino bound on the same $N \\to e^+e^-\\nu$ process that this analysis compares against and improves.","marker":"[7]"},{"why":"Provides the PIENU-based accelerator bound that this analysis supersedes in the 2-12 MeV mass range.","marker":"[19]"},{"why":"Gives the $N \\to e^+e^-\\nu$ and $N \\to 3\\nu$ decay widths and the phase-space factor needed for the signal calculation.","marker":"[77]"},{"why":"Supplies the solar neutrino spectrum, in particular the $^8$B flux, which is the HNL production source in the relevant energy range.","marker":"[80]"},{"why":"Provides the Jovian-electron background model and the diffusion coefficient and mean free path estimates used for both the background fit and the signal propagation.","marker":"[82]"},{"why":"Provides the cosmic-ray transport and diffusion framework on which the analytic solution Eq. (11) is based.","marker":"[83]"},{"why":"Supplies the residence-time model for Jovian electrons that fixes the parallel mean free path $\\lambda_\\parallel = 0.15$ used as the background normalization prior.","marker":"[97]"},{"why":"Provides the Ulysses dataset, recorded with the COSPIN/KET detector, that is one of the two electron spectra in the fit.","marker":"[49]"},{"why":"Provides the SOHO mission dataset, recorded with the COSTEP/EPHIN detector, that is the second electron spectrum in the fit.","marker":"[50]"}],"fun_headline_variants":["Solar electrons set tightest limit on MeV heavy neutrinos","Ulysses and SOHO beat Borexino for solar heavy neutrino bounds","Heliospheric electron data tighten heavy neutrino mixing to 10^-6","Strongest direct limit on solar heavy neutrinos from MeV electrons","New record bound on heavy neutrino mixing from solar electron spectra"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The bound rests on a simplified one-dimensional diffusion model for MeV electrons in the inner heliosphere, with a single constant diffusion coefficient, no convection or energy losses, and a Jovian-electron background model whose fitted normalization can absorb part of a real signal; if the true electron transport or the background differs substantially, the predicted flux, and therefore the excluded mixing which scales as the square root of the flux, shifts correspondingly.","fun_headline_variants_meta":{"raw":{"variants":["Solar electrons set tightest limit on MeV heavy neutrinos","Ulysses and SOHO beat Borexino for solar heavy neutrino bounds","Heliospheric electron data tighten heavy neutrino mixing to 10^-6","Strongest direct limit on solar heavy neutrinos from MeV electrons","New record bound on heavy neutrino mixing from solar electron spectra"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000764,"raw_usage":{"total_tokens":3378,"prompt_tokens":924,"completion_tokens":2454,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":540,"completion_tokens_details":{"reasoning_tokens":2365}},"tokens_in":540,"tokens_out":2454,"duration_ms":15237,"temperature":1.0,"reasoning_tokens":2365,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:58:02.798460+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A positron-sensitive MeV detector near 1 AU could settle the claim directly: positrons from HNL decays would appear as an excess over a background that is up to three orders of magnitude smaller than the electron background, so a null result after profiling the Jovian background would confirm the exclusion, while an excess matching $M_N \\simeq 8.2$ MeV with $U_e^2 \\simeq 7.9 \\times 10^{-7}$ would confirm the tentative signal. In parallel, multi-spacecraft tracking of Jovian electrons during magnetic connection events could fix $D$ and the parallel mean free path, converting the current half-order-of-magnitude propagation uncertainty into a definite prediction for the 1 AU flux and shifting the limit as the square root of the flux.","supporting_citations":[{"cited_title":"Remarks on the KARMEN Anomaly","cited_arxiv_id":"hep-ph/9503295","evidence_quote":"Provides the Jovian-electron background model and the diffusion coefficient and mean free path estimates used for both the background fit and the signal propagation."}],"review_version":1}