{"id":"e92e0bbe-cdae-49e7-9266-bb204c067eb2","arxiv_id":"2607.27307","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"ATLAS finds no evidence for heavy Majorana neutrinos in same-sign tau-lepton final states and sets new 95% CL upper limits on |V_τN|², reaching 0.30 at m_N = 900 GeV and extending to m_N = 6.5 TeV.","lead":"ATLAS searched for a hypothetical heavy neutrino that decays into tau leptons in proton collisions at the LHC and saw no sign of it. The result sets the strongest limits yet on how strongly this particle could mix with ordinary tau neutrinos for masses above 1 TeV.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Jet→τ fake background is extrapolated from a non-VBS W+jets control region without a closure test in the signal-region phase space; the absolute |V_τN|^2 limits depend on this extrapolation.","rationale":"The paper is a well-executed ATLAS search with standard methods, and the no-excess conclusion is likely robust. However, the absolute limits on |V_τN|^2 are the central quantitative result, and those limits scale directly with the dominant jet-fake background. The fake-factor method is credible, but the WCR used to measure it is not in the VBS phase space, and the validation regions shown do not close that loop: the OS VR is enriched in real leptons and the AntiVBS VR explicitly inverts the VBS selection. The uncertainty derived by inverting EmissT/mT is a reasonable first estimate, but it does not directly test whether the jet-fake rate is the same when m_jj>300 GeV and |Δη_jj|>4 are required. Because the same-sign τhadτhad channel is the most sensitive channel for high m_N, and because the paper's claimed extension above 1.5 TeV depends on that channel's background estimate, this is the most load-bearing assumption in the analysis. The proposed closure test is a standard, feasible check with existing data and MC; if it passes, the limits stand, and if it fails, the limits and possibly the headline constraints would need adjustment. The reader's weakest assumption also flagged the fake-factor jet-composition issue, so this stress-test partly agrees, while focusing on the absence of a VBS-like closure rather than the model-dependence of the single-neutrino interpretation.","tokens_in":46148,"tokens_out":18677,"duration_ms":167925,"concrete_test":"Construct a VBS-like fake-factor closure region using the same μτhad or τhadτhad preselection, requiring m_jj>300 GeV and |Δη_jj|>4, with the τhad candidate failing the medium identification but passing the loose identification (Anti-ID). Subtract the MC-predicted real-τ, lepton-fake, and other backgrounds, then apply the nominal WCR fake factor to predict the jet-fake yield. Compare this prediction with observed data in bins of leading τhad pT, for both opposite-sign and same-sign sub-samples. If the data/prediction ratio deviates from unity by more than the currently assigned 30% jet-composition uncertainty, the fake-factor extrapolation and the derived |V_τN|^2 limits would need to be revised, e.g., by re-deriving the fake factor in a VBS-like control region or by enlarging the systematic uncertainty.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The dominant background in all three signal regions is Jet→τhad fake, estimated with a fake factor derived in a WCR that explicitly omits the VBS requirements (no m_jj>300 GeV, no |Δη_jj|>4) and requires opposite-sign μτ, EmissT>20 GeV, mT>20 GeV. The text notes that the SR jet-fake sample is dominated by QCD multijet events, especially in the τhadτhad channel, while the WCR is enriched in W+jets. The only systematic assigned for the WCR→SR jet-composition difference is evaluated by inverting EmissT and mT; this changes the W+jets/QCD mixture but does not reproduce the two-forward-jet, large-m_jj, same-sign topology of the SR. The validation regions are either opposite-sign (enriched in real leptons) or anti-VBS (m_jj<300 GeV, |Δη_jj|<4), so neither validates the fake factor in the actual VBS signal phase space. Since jet fakes dominate the background, a shift in the fake-factor normalization or shape in the VBS phase space directly shifts the fitted background yields and therefore the 95% CL cross-section limits from which the |V_τN|^2 constraints are derived. The observed agreement in the SR distributions is reassuring, but the profile-likelihood fit can absorb moderate shifts through nuisance parameters; the absolute normalization of the derived limits still rests on the unvalidated WCR-to-SR extrapolation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a search for heavy Majorana neutrinos coupling to third-generation leptons in vector boson scattering (VBS) with same-sign ττ, eτ, and μτ final states, using 140 fb^{-1} of 13 TeV pp collisions recorded by ATLAS. Three signal regions are defined with VBS topology (m_jj>300 GeV, |Δη_jj|>4), a b-jet veto, and a same-sign lepton–τhad or τhad–τhad pair. The dominant jet→τhad fake background is estimated with a data-driven fake-factor method derived from a W+jets control region, with smaller contributions from lepton fakes and prompt-τ processes taken from MC. A simultaneous profile-likelihood fit to the leading-τ pT distributions finds no significant excess: observed yields of 887, 579, and 270 events agree with predicted backgrounds of 886±29, 579±23, and 264±14 in the eτ, μτ, and ττ channels, respectively. Upper limits at 95% CL are set on |V_τN|^2 for heavy Majorana neutrino masses between 92 GeV and 6.5 TeV, with the strongest exclusion at m_N=900 GeV (observed/expected 0.30/0.31), and two-dimensional limits are presented for m_N=1 and 5 TeV.","tokens_in":46451,"tokens_out":11060,"duration_ms":89374,"significance":"If the result holds, this is the first search for tau-flavored heavy Majorana neutrinos in vector boson scattering and extends the excluded mass range above the reach of previous same-sign dilepton searches. The analysis uses the full Run 2 dataset, a data-driven fake-factor estimate for the dominant background, two validation regions, a simultaneous fit, and a detailed systematic model (Table 1). These are genuine strengths. The main caveat is that the absolute background normalization, and therefore the numerical values of the derived limits, rests on an extrapolation of the jet→τhad fake factor from a non-VBS W+jets control region into the VBS signal phase space; the closure of this extrapolation is not demonstrated in the signal-region topology. This concern is load-bearing because jet fakes dominate all three signal regions.","major_comments":[{"comment":"The dominant Jet→τhad fake background is estimated with a fake factor derived in a W+jets control region that omits the VBS requirements (no m_jj>300 GeV, no |Δη_jj|>4), requires an opposite-sign μτ pair, and applies EmissT and mT cuts. The only systematic for the WCR→SR jet-composition difference is evaluated by inverting EmissT and mT, which changes the W+jets/QCD mixture but does not reproduce the same-sign, two-forward-jet, high-m_jj topology of the signal regions; in the τhadτhad channel the WCR also requires a muon that is not present in that SR. The AntiVBS validation region covers m_jj<300 GeV and |Δη_jj|<4, not the VBS phase space. Since jet fakes dominate all three SRs, a normalization or shape misestimate of this background propagates directly into the fitted yields and therefore into the derived |V_τN|^2 limits; Table 1 assigns only 1.3/1.2/6.4% fake-factor systematics. The o","section":"Background estimation (fake-factor method and validation regions)"},{"comment":"The conversion from cross-section limits to |V_ℓN|^2 uses σ_ℓℓ' ∝ |V_ℓN|^2|V_ℓ'N|^2, but the text lists only signal uncertainties that affect the acceptance (renormalization/factorization scales, ISR/FSR, multiple parton interactions). It is not stated whether the uncertainty on the predicted total signal cross-section (PDF and scale) is included when translating the cross-section limit into a bound on |V|^2. If it is omitted, the quoted limits on a model parameter may be over-constrained. Please state explicitly the treatment of the signal cross-section normalization uncertainty and, if it is not included, provide its effect on the observed limits.","section":"Systematic uncertainties and cross-section interpretation"}],"minor_comments":[{"comment":"The phrase 'same-sign ττ, eτ or μτ pair' is imprecise: in the eτ and μτ channels the e/μ is a prompt lepton rather than a tau decay. Suggest rewording to 'same-sign lepton–τhad pair'.","section":"Abstract"},{"comment":"The sentence 'The feature around 900 GeV is consistent with a statistical fluctuation found in the reconstruction efficiency of the signal events' is vague. Please specify the nature of the efficiency fluctuation and its statistical significance, or remove the statement.","section":"Results paragraph"},{"comment":"The two-dimensional exclusion contours are shown only for m_N=1 and 5 TeV. If feasible, include a few more mass points or state that additional masses are available in auxiliary material.","section":"Figure 4"},{"comment":"A table of observed and expected limits as a function of m_N would improve reproducibility and make the numerical claims easier to verify than relying on the figure alone.","section":"Limits presentation"},{"comment":"The data-driven correction for the ℓ→τhad fake background from Z→ee events is mentioned, but no uncertainty for the correction is specified. Please clarify whether it is included in the quoted systematic uncertainties.","section":"ℓ→τhad fake correction"},{"comment":"Reference [33] appears as 'Refs. [33]' (plural); it should be singular.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main risk to this paper is the unvalidated WCR→SR fake-factor extrapolation into the VBS phase space, which directly affects the absolute normalization of the dominant background and hence the numerical limits. I would like the authors to add a closure test or a more robust jet-composition systematic before publication. The analysis is otherwise well executed and the physics case is appropriate for this journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is a standard, well-executed ATLAS search that gives the first direct limits on heavy Majorana neutrino VBS production in same-sign tau-tau, e-tau, and mu-tau final states. The null result is believable, and the paper is honest about its methods. The one real caveat is the dominant jet-to-tau fake background: its normalization rests on an extrapolation from a W+jets control region that does not include the VBS requirements, and the paper doesn't demonstrate closure in the signal-region phase space.\n\nWhat's new: extending the VBS heavy-neutrino search to taus. That matters because the previous ATLAS and CMS VBS searches used only electrons and muons, so the tau mixing element is largely unconstrained above a TeV. The analysis itself uses routine Run-2 tools: NLO signal MC, data-driven fake factors, profile likelihood with systematics. The yields match beautifully—887 vs 886±29, 579 vs 579±23, 270 vs 264±14—which speaks for the overall background modeling. I also credit the paper for stating the model dependence (type-I seesaw, sigma proportional to |V|^4) rather than burying it.\n\nThe soft spot: the fake-factor method. The WCR is opposite-sign, has no m_jj or Delta_eta_jj cut, and is enriched in W+jets. The signal regions, especially tauhad-tauhad, are dominated by QCD multijet fake taus. The only systematic on the composition difference is derived by inverting EmissT and mT, which changes the W/Z-to-QCD ratio but does not create the two-forward-jet, large-mass, same-sign environment of the signal region. The OS and anti-VBS validation regions also don't test the fake factor in the actual VBS phase space. So if the fake factor's shape or normalization differs there, the limits on |V_tauN|^2 shift directly. The profile-likelihood fit can soak up some of this through nuisance parameters, but the absolute scale of the limits is set by this unvalidated extrapolation. It's a standard limitation for the method, and ATLAS's internal review may have auxiliary studies that address it, but the PRL text as written doesn't close the gap.\n\nThere's nothing circular in the limit derivation, and the citation pattern looks proper. The claim of 'first direct constraints above 1.5 TeV' is fair for the tau coupling.\n\nBottom line: this deserves serious peer review. The search is valid and worth publishing, but the referee should ask for better closure of the fake-factor estimate in the VBS phase space, or at least a quantitative bound on how much a QCD-multijet contamination could shift the limits.","headline":"First VBS search with same-sign tau final states, well executed, but the dominant jet-to-tau fake background is not validated in the signal-region phase space, so trust the null result but treat the absolute limits as provisional.","tokens_in":47008,"tokens_out":3846,"would_cite":true,"duration_ms":32375,"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":"This paper searches for heavy Majorana neutrinos coupling to tau leptons in vector boson scattering at the LHC and finds no excess, setting new upper limits on the tau mixing element for masses from 92 GeV to 6.5 TeV.","keywords":["heavy Majorana neutrino","type-I seesaw","vector boson scattering","same-sign tau pairs","tau lepton","LHC","ATLAS detector","mixing matrix"],"falsifier":"A concrete test would be to reconstruct the invariant mass of the same-sign tau-pair system in the selected events: the paper fits only the leading tau pT, so a narrow resonance peak at any m_N where |V_tauN|^2 is claimed to be excluded would directly falsify the limit claim.","tokens_in":45987,"feed_emoji":"⚛️","tokens_out":6531,"duration_ms":55995,"temperature":0.7,"pith_summary":"Using 140 fb^-1 of 13 TeV proton-proton collisions recorded by the ATLAS detector, the paper searches for heavy Majorana neutrinos (hypothetical particles that are their own antiparticles, proposed in seesaw models to explain tiny neutrino masses) produced in vector boson scattering and decaying through same-sign tau pairs, including e-tau and mu-tau final states. No significant deviation from Standard Model backgrounds is observed. From this absence, the paper derives upper limits on the squared mixing element |V_tauN|^2 for masses between 92 GeV and 6.5 TeV, with the strongest exclusion at 900 GeV of 0.30 observed (0.31 expected). These are the first direct constraints for masses above 1.5 TeV and improve existing limits for tau-coupled heavy neutrinos above 1.0 TeV. A sympathetic reader would care because this probes a previously unexplored mass range for a third-generation seesaw mechanism.","feed_headline":"Set new exclusion limits on tau-coupled heavy neutrinos to 6.5 TeV","feed_subtitle":"First direct limits above 1.5 TeV for third-generation neutrinos; at 900 GeV, mixing above 0.30 is excluded.","key_machinery":"The search exploits the same-sign W-boson scattering topology mediated by a heavy Majorana neutrino. Events are selected with two forward jets with dijet mass above 300 GeV and large rapidity separation, plus a same-sign tau-tau, e-tau, or mu-tau pair with the leading hadronic tau pT above 55 GeV. The dominant background, jets misidentified as hadronic taus, is estimated with a data-driven fake-factor method derived in a W+jets control region; a profile likelihood fit to the leading-tau pT distribution then yields the limits. The tau-hadron tau-hadron channel provides the best sensitivity.","core_discovery":"The central claim is that if a heavy Majorana neutrino couples to tau leptons, it would be produced via same-sign W±W± vector boson scattering and yield same-sign tau-tau, e-tau, or mu-tau events; comparing the data to simulation finds the data consistent with background. Assuming the type-I seesaw model, in which the cross section scales as sigma ∝ |V_ellN|^2 |V_ell'N|^2, the paper sets 95% CL upper limits on |V_tauN|^2. The limits lie below 1 for m_N from 92 GeV to 6.5 TeV, with the strongest exclusion at m_N = 900 GeV: an observed (expected) limit of 0.30 (0.31). Two-dimensional exclusions in (|V_eN|^2, |V_tauN|^2) and (|V_muN|^2, |V_tauN|^2) are also derived at m_N = 1 and 5 TeV.","pith_inferences":["If the heavy neutrino is not Majorana, or if more than one heavy state contributes, the quoted |V_tauN|^2 bounds do not directly apply because the production rate scaling would change.","Combining this search with existing e-e and mu-mu vector-boson-scattering searches could yield a global fit to all three |V_lN|^2 simultaneously, going beyond the pairwise exclusions presented here.","A dedicated measurement of the jet-composition dependence of the fake factor, for instance using gluon-enriched control samples, would test the main background-systematic assumption and could sharpen the limits.","At higher luminosity or higher collision energy, the same same-sign-tau signature could push the excluded mixing values down by roughly the square root of the integrated-luminosity gain."],"forward_implications":["For a type-I seesaw neutrino with m_N = 900 GeV, mixing |V_tauN|^2 ≥ 0.30 is now excluded at 95% CL.","The search covers an unexplored mass range for third-generation couplings, with first direct constraints above 1.5 TeV and sensitivity extending to 6.5 TeV.","The two-dimensional exclusions in (|V_eN|^2, |V_tauN|^2) and (|V_muN|^2, |V_tauN|^2) allow flavor-dependent seesaw models to be tested.","These results complement low-energy searches for lepton-number violation by directly probing the third-generation sector at collider energies.","The tau-hadron tau-hadron final state proves to be a powerful probe, motivating continued use of tau-rich signatures in future searches."],"fun_headline_variants":["Tau-coupled heavy neutrino limits pushed to 6.5 TeV","No Majorana neutrinos seen in tau-lepton VBS events","First exclusions above 1.5 TeV for tau-neutrino mixing","ATLAS excludes tau-neutrino mixing below 0.30 at 900 GeV","Heavy neutrinos with tau couplings ruled out to 6.5 TeV"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The limit interpretation assumes a single heavy Majorana neutrino whose production cross section scales exactly as |V_lN|^2 |V_l'N|^2, and also assumes that jets mimicking tau decays have the same composition in the control region as in the signal region; if either assumption fails, the extracted bounds would not directly apply.","fun_headline_variants_meta":{"raw":{"variants":["Tau-coupled heavy neutrino limits pushed to 6.5 TeV","No Majorana neutrinos seen in tau-lepton VBS events","First exclusions above 1.5 TeV for tau-neutrino mixing","ATLAS excludes tau-neutrino mixing below 0.30 at 900 GeV","Heavy neutrinos with tau couplings ruled out to 6.5 TeV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000311,"raw_usage":{"total_tokens":1664,"prompt_tokens":859,"completion_tokens":805,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":603,"completion_tokens_details":{"reasoning_tokens":717}},"tokens_in":603,"tokens_out":805,"duration_ms":6769,"temperature":1.0,"reasoning_tokens":717,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T09:47:27.643217+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete test would be to reconstruct the invariant mass of the same-sign tau-pair system in the selected events: the paper fits only the leading tau pT, so a narrow resonance peak at any m_N where |V_tauN|^2 is claimed to be excluded would directly falsify the limit claim.","supporting_citations":[],"review_version":1}