{"id":"fcf1445b-d135-4a45-9c24-5be91e607683","arxiv_id":"2501.09065","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Long-lived heavy neutral leptons produced via top-quark effective operators could be probed at the HL-LHC up to new-physics scales around 12 TeV at ATLAS and 4.5 TeV at MATHUSLA or ANUBIS.","lead":"This paper calculates how often the upgraded LHC could produce a hypothetical heavy neutral lepton together with top quarks, and how many such particles ATLAS or proposed off-axis detectors would catch. The projections give concrete new-physics scales, up to about 12 TeV at ATLAS and 4.5 TeV at far detectors, that experiments could probe.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Background-free assumption for the proposed lower-threshold ATLAS DV+jets selection is unvalidated; the 3-event contours that set the headline Λ≈12 TeV reach rely on it.","rationale":"The paper is a well-executed phenomenological study. The operator basis, production and decay calculations, and the simulation pipeline are coherent; the one-loop width for O13_QuNL is cross-checked against Ref. [60]. The central claim—Λ reaches of 12 (4.5) TeV—rests on the number of expected signal events. The most fragile step is the translation of signal-event counts into exclusion limits: the 3-event contours are derived under the explicit assumption of zero SM background at ATLAS after lowering the jet-pT thresholds relative to the 13 TeV search, and under 100% detection efficiency at the far detectors. This assumption is plausible (displaced-vertex selections are powerful) but not demonstrated. The paper does not provide a background estimate for the modified selection, and the 13 TeV search's jet cuts were not lowered in the original analysis without reason. If even O(1) background events survive, the 3-event requirement undercovers; the 30-event contours, which would correspond to ~225 background events, show the possible loss in reach. The far-detector efficiencies are secondary because the reach scales only as the fourth root of efficiency. I considered the EFT-validity/unitarity concern (s_hat could exceed Λ^2 for some events), but the high-x PDF suppression makes this a smaller effect for the quoted Λ values. Therefore, the unvalidated background assumption is the single most load-bearing concern. Since the paper itself provides higher-event contours that partially mitigate the issue, and the underlying physics is sound, the reader's CONDITIONAL verdict is appropriate and should be kept.","tokens_in":22111,"tokens_out":11899,"duration_ms":123628,"concrete_test":"Use a validated ATLAS fast simulation (e.g., Delphes with the ATLAS card, or the public recast framework for the 13 TeV DV+jets search) to generate the dominant SM backgrounds at √s=14 TeV with HL-LHC pileup (QCD multijet, ttbar, W/Z+jets, single top, and a sample with detector-material interactions) and apply the full proposed event-level selection (4/5/6 jets with pT>90/65/55 GeV) followed by the DV-level requirements (4mm<Rxy<300mm, |z|<300mm, ≥5 tracks with |d0|>2mm, pT/|q|>1 GeV, and invariant mass >10 GeV) and the vertex-level parameterized efficiencies from Ref. [73]. Count the surviving background events. If the expected background is <0.1 events for 3 ab^-1, the 3-event contours are valid; otherwise, recompute the 95% CL exclusion contours using the Poisson upper limit with the estimated background and quantify the resulting reduction in the Λ reach.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline sensitivity of Λ≈12 TeV at ATLAS (and 4.5 TeV at far detectors) comes from the 3-event zero-background contours in Figs. 7 and 8. Section 3.1 justifies zero background by arguing that the vertex-level selection criteria suppress all backgrounds, citing that in the original 13 TeV ATLAS DV+jets search [73] background suppression is driven by DV requirements rather than jet cuts. However, the proposed event selection lowers the jet-pT thresholds to 90/65/55 GeV and changes the multi-jet requirement to 4/5/6 jets. At √s=14 TeV with HL-LHC pileup, the lower thresholds can admit SM events (QCD multijet, ttbar, W/Z+jets, and detector-material interactions) that were not present in the original search's signal region. No SM background simulation is provided for the modified selection; the paper only states the assumption. If even a few background events survive, the 95% CL exclusion requires more than 3 signal events, and the Λ reach scales roughly as (N_required/3)^{1/4} (e.g., 30 events would reduce Λ≈12 TeV to about 7 TeV). The paper does present 10- and 30-event contours, but the abstract and Section 4 quote the 3-event numbers as the primary result, making the central claim conditional on an unvalidated experimental assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies long-lived heavy neutral leptons (HNLs) in the NRSMEFT, concentrating on dimension-6 four-fermion operators that couple HNLs to top quarks. Four benchmark scenarios are defined: the pair-NR operator O_uN^13 and the single-NR operators O_QuNL^13, O_duNe^13, and O_duNe^33. Production cross sections in pp collisions at 14 TeV and HNL decay widths (including a loop-induced three-body decay cross-checked against Ref. [60]) are computed, and the sensitivities of ATLAS and the far detectors MATHUSLA, ANUBIS, CODEX-b, and MAPP2 are estimated via displaced-vertex searches. The headline claim is that for O_uN^13 with c/Λ^2 = 1 TeV^-2, under zero-background and 100%-efficiency assumptions, ATLAS could probe scales up to Λ ≈ 12 TeV and MATHUSLA/ANUBIS up to Λ ≈ 4.5 TeV, with complementary HNL-mass coverage. The paper also provides 10- and 30-event contours to illustrate the impact of less optimistic assumptions.","tokens_in":22321,"tokens_out":8162,"duration_ms":89710,"significance":"If the projected sensitivities hold, the paper would establish a new and previously unstudied class of top-philic HNL operators as promising targets for HL-LHC displaced-vertex searches, with competitive scale reaches and complementary mass coverage between ATLAS and the far detectors. The paper is technically careful in several respects: the production and decay formulas are standard or cross-checked against the literature, the loop-induced three-body width is explicitly compared with Ref. [60], the ATLAS recast uses public parameterized efficiencies, and the far-detector geometry is implemented in detail. The authors are also transparent about the idealized nature of their event-count assumptions, and they show 10- and 30-event contours as a robustness check. However, the headline numbers quoted in the abstract and Section 4 are the 3-event, zero-background, 100%-efficiency projections, and these assumptions are not validated for the modified ATLAS selection or for the far-detector efficiencies. The significance is therefore real but conditional on assumptions that need further justification or a more cautious presentation.","major_comments":[{"comment":"The zero-background assumption for the proposed ATLAS DV+jets selection is load-bearing and currently unvalidated. The proposed selection lowers the jet-pT thresholds to 90/65/55 GeV and uses 4/5/6-jet multiplicities at 14 TeV, whereas the Run-2 search in Ref. [73] used harder selections; no SM background simulation is provided for the modified selection. Because the 3-event contours in Figs. 7 and 8 produce the abstract's Λ ≈ 12 TeV reach, even a small residual background could materially reduce the claimed exclusion power. For example, since the production rate scales roughly as Λ^-4 in the production-dominated regime, requiring 30 signal events instead of 3 would lower the Λ reach by about a factor of (3/30)^(1/4) ≈ 0.56, bringing 12 TeV to about 7 TeV. I ask the authors to either provide a background estimate for the modified selection or, in the abstract and Section 4, use the more robust 10- or 30-event contours as the primary stated reach, with the 3-event contours clearly labeled as idealized projections.","section":"Section 3.1, after Eq. (3.1), and footnote 13"},{"comment":"The far-detector sensitivity estimates assume ε = 100% detection efficiency and vanishing background, and the paper acknowledges this explicitly. This assumption is load-bearing for the quoted reach of Λ ≈ 4.5 TeV at MATHUSLA and ANUBIS, because reconstruction and selection efficiencies for HNL decays inside these detectors are not modeled. The 30-event contours provide some robustness against changes in acceptance or background level, but the abstract and Section 4 quote the 3-event, 100%-efficiency numbers as the primary results. Please either include more realistic efficiency assumptions or clearly present the far-detector numbers as upper-bound projections, with the efficiency dependence stated in the abstract and conclusions.","section":"Section 3.2, Eq. (3.3)"}],"minor_comments":[{"comment":"The abstract's '12 TeV' reach is not explicitly derived in the text of Section 4, which states that ATLAS can probe scales 'in excess of 10 TeV' for 40 GeV ≲ m_N ≲ 80 GeV in scenario 1. Please specify the exact contour and parameter point that gives Λ ≈ 12 TeV, or adjust the abstract to match the stated numerical result.","section":"Section 4 and abstract"},{"comment":"Please clarify whether the cross section σ in Eq. (3.1) is the sum over all relevant production channels and charge conjugates, e.g., pp → t N Nbar and pp → tbar N Nbar for the pair-NR case. The text discusses these channels separately, and an unambiguous normalization is important for reproducing the event counts.","section":"Eq. (3.1)"},{"comment":"The text says that MATHUSLA and ANUBIS reach HNL masses 'in excess of 1 TeV' and immediately adds that the plot range is limited by EFT validity for Λ = 1 TeV. Please reconcile these statements, since a reach above 1 TeV would appear to lie outside the valid region for the assumed Λ.","section":"Section 4, upper-left panel of Fig. 7"},{"comment":"The caption for Fig. 8 states that solid lines correspond to |V_eN|^2 = 10^-10 except for scenario 2, where they correspond to 10^-9. This is consistent with footnote 18, but the main text would benefit from stating this exception in the same sentence where the mixing values are first introduced, to avoid apparent inconsistency.","section":"Section 4, Fig. 8 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is transparent and technically sound at the level of the EFT calculations and the simulation framework. My main concern is the gap between the headline reach claims (3-event, zero-background, 100%-efficiency) and the absence of validated background and efficiency estimates for the modified ATLAS selection and the far detectors. This is fixable by additional estimates or by reframing the abstract and conclusions around the more robust event-count contours. The paper fits the journal's scope, and I do not see a load-bearing error that would require rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take: this is a well-executed projection study that breaks genuinely new ground by simulating displaced-vertex sensitivities for long-lived HNLs from d=6 top-quark operators. The headline numbers (Lambda ~ 12 TeV at ATLAS, 4.5 TeV at far detectors) are conditional on a zero-background assumption that the paper does not fully validate, but the paper is transparent about this and provides 10- and 30-event contours that largely defuse the worry.\n\nThe genuinely new content is the detector-level sensitivity study itself. The operator list and tree-level widths mostly come from Ref. [59], and the one-loop three-body decay reproduces Ref. [60], but the ATLAS low-threshold jet strategy and the far-detector acceptance calculations for these operators are new. I checked the derivations and they are coherent; the loop width is explicitly cross-checked against the literature. The benchmark scenarios are well chosen and the mass/mixing/scale structure of the results is clear.\n\nThe soft spot is exactly what the stress-test flags: the ATLAS contours are set by requiring 3 signal events because the authors assert that the vertex-level cuts suppress all backgrounds, but they lower the jet thresholds relative to the original 13 TeV search and provide no SM background estimate for the modified selection. If residual backgrounds appear, the reach shrinks roughly as (N_required/3)^{1/4} — 30 events would bring 12 TeV down to about 7 TeV. That said, the paper itself quotes the 10- and 30-event contours, so the reader is not misled. The far-detector 100% efficiency/zero-background assumptions are standard for these proposals, though optimistic. No code or model files are shipped, which is a minor drawback for reproducibility.\n\nOverall, the central argument holds up: the projections are reasonable and the paper is careful. It is primarily for the HNL/LLP and EFT phenomenology community, and it deserves a serious referee. I'd recommend publishing after the authors add an explicit discussion of background robustness — even a simple scaling of the reach versus assumed background count, or a caveat on the 3-event contours, would be enough.","headline":"Solid, genuinely new LLP sensitivity projections for top-philic HNL operators; headline reach rests on an unvalidated zero-background assumption, but the paper's own 10/30-event contours keep it honest.","tokens_in":22970,"tokens_out":1965,"would_cite":true,"duration_ms":18806,"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 claims that displaced-vertex searches at the HL-LHC could probe top-HNL effective couplings up to new-physics scales of about 12 TeV at ATLAS and 4.5 TeV at MATHUSLA and ANUBIS.","keywords":["heavy neutral leptons","long-lived particles","displaced vertices","effective field theory","top quark","HL-LHC","MATHUSLA","ANUBIS"],"falsifier":"Simulate a full Standard Model background sample through the proposed ATLAS search chain — the 4/5/6-jet thresholds with $p_T > 90/65/55$ GeV, vertices with $4 < R_{xy} < 300$ mm, $|d_0| > 2$ mm, at least five tracks, and a vertex mass above 10 GeV — and count events at $L = 3$ ab$^{-1}$; more than a handful of events would invalidate the 3-event exclusions. For ANUBIS, the corresponding check is a complete background study of the service-shaft location, since its proximity to the IP could admit non-negligible backgrounds.","tokens_in":21805,"feed_emoji":"⚛️","tokens_out":11198,"duration_ms":97446,"temperature":0.7,"pith_summary":"This paper tries to establish that long-lived heavy neutral leptons (HNLs) coupled to top quarks through dimension-six effective operators could be probed or excluded at the high-luminosity LHC using displaced vertices. The authors classify the production and decay of HNLs from four-fermion top-HNL operators, and compute sensitivity projections for ATLAS and the proposed far detectors MATHUSLA, ANUBIS, CODEX-b, and MAPP2. For the pair-production operator $\\mathcal{O}^{13}_{uN}$ with coefficient $c/\\Lambda^2 = 1\\,\\mathrm{TeV}^{-2}$, they estimate that ATLAS could probe new-physics scales up to about 12 TeV, while MATHUSLA and ANUBIS reach about 4.5 TeV, over complementary HNL mass ranges. A sympathetic reader would care because these operators escape present top-width constraints, so a displaced-vertex search would turn the HL-LHC into a probe of few-TeV top-philic new physics.","feed_headline":"Displaced vertices could probe top-heavy neutrino physics to 12 TeV","feed_subtitle":"New EFT study has ATLAS reaching 12 TeV for top-coupled heavy neutral leptons, with MATHUSLA and ANUBIS at 4.5 TeV.","key_machinery":"The load-bearing object is the set of dimension-six four-fermion operators coupling a right-handed heavy neutral lepton $N_R$ to top quarks, each entering with Wilson coefficient $c/\\Lambda^2$, together with the displaced-vertex machinery used to count signal events. In the pair-NR operator case, production is controlled by the operator while decay is controlled by the active-sterile mixing $|V_{eN}|^2$; in the single-NR operator case the same operator feeds both production and decay, including loop-induced three-body decays. ATLAS sensitivities are obtained by applying jet and vertex requirements to simulated events and using the ATLAS DV+jets parameterized efficiencies, while far-detector sensitivities come from the probability that a simulated HNL decays inside each detector's fiducial volume, all evaluated at $\\sqrt{s}=14$ TeV and $L=3\\,\\mathrm{ab}^{-1}$.","core_discovery":"The central claim is that at the HL-LHC with $L=3\\,\\mathrm{ab}^{-1}$, displaced-vertex searches can probe dimension-six four-fermion operators containing a top quark and a heavy neutral lepton much more strongly than earlier estimates based on stable-HNL final states. For the pair-NR operator $\\mathcal{O}^{13}_{uN}$, the dominant production is $pp \\to tN\\bar{N}$ in association with a top quark, and the HNL decays only through active-sterile mixing; the authors estimate that ATLAS can exclude scales $\\Lambda \\lesssim 12\\,\\mathrm{TeV}$ for $40\\,\\mathrm{GeV} \\lesssim m_N \\lesssim 80\\,\\mathrm{GeV}$ at $|V_{eN}|^2 = 10^{-10}$, while MATHUSLA (ANUBIS) reach about 4.5 (4) TeV for $m_N$ near 20 GeV. For single-NR operators such as $\\mathcal{O}^{13}_{QuNL}$ and $\\mathcal{O}^{13}_{duNe}$, the operator itself drives both production and decay, producing a funnel-shaped region of sensitivity where even negligibly small mixings are probed, and ATLAS reaches scales of 7 to 8 TeV. All reach estimates assume the vertex-level selection and far-detector shielding reduce backgrounds to zero, so the 95% exclusion is set at three signal events, with additional contours shown for ten or thirty events.","pith_inferences":["Editorial inference: the headline 12 (4.5) TeV reach rests on the zero-background premise; if a realistic Standard Model background simulation finds even a handful of events passing the lower jet-threshold selection, the 3-event contours would shrink, and the 10- or 30-event contours give a rough measure of that loss.","Editorial inference: the same search strategy could be applied to CMS or to data-parked low-threshold triggers, potentially delivering an independent cross-check and roughly doubling the ATLAS-like coverage.","Editorial inference: in the single-NR funnel regions a signal would not by itself determine the mixing parameter; measuring the HNL lifetime distribution or charge asymmetry would be needed to separate operator-induced from mixing-induced decays.","Editorial inference: the far-detector projections assume 100% detection efficiency, so folding in realistic tracking efficiencies would reduce the quoted far-detector scale reach proportionally to the efficiency."],"forward_implications":["If correct, the HL-LHC displaced-vertex program becomes a direct probe of top-philic HNL effective operators at scales an order of magnitude above present top-width limits.","For the pair-NR operator $\\mathcal{O}^{13}_{uN}$, MATHUSLA and ANUBIS could reach active-sterile mixings as low as $|V_{eN}|^2 \\sim 10^{-20}$ for $m_N \\gtrsim 120$ GeV, far below current searches.","ATLAS and the far detectors cover complementary HNL mass ranges, so a combined analysis fills the mass-lifetime plane that neither experiment covers alone.","For single-NR operators, the funnel-shaped sensitivity means the operator alone can produce a displaced signal even with vanishing mixing, making the search sensitive to operators rather than only to mixing.","The flavor-diagonal variant $\\mathcal{O}^{33}_{duNe}$ yields a weaker reach, at most about 1.5 TeV at ATLAS, so not all top-HNL operators are equally discoverable."],"supporting_citations":[{"why":"Supplies the operator basis and the analytic formulas for flavor-violating top decays used for single- and pair-NR production.","marker":"[59]"},{"why":"The ATLAS DV+jets search whose parameterized event- and vertex-level efficiencies are applied in the recast.","marker":"[73]"},{"why":"The recast of the ATLAS DV+jets search that motivates the proposed jet and vertex selection strategy.","marker":"[74]"},{"why":"The 8 TeV ATLAS displaced-vertex search whose lower jet pT thresholds replace the 13 TeV thresholds.","marker":"[75]"},{"why":"Connects the pair-NR HNL production cross sections in the EFT and the Dirac/Majorana comparison.","marker":"[39]"},{"why":"Provides the active-sterile mixing HNL decay width used for pair-NR and mixing-induced decays.","marker":"[70]"},{"why":"Supplies the loop-induced three-body HNL decay rates for single-NR operators that dominate at low mass.","marker":"[60]"},{"why":"Defines the MATHUSLA geometry used to compute the fiducial-volume decay probability.","marker":"[90]"},{"why":"Defines the ANUBIS geometry used for the far-detector sensitivity estimates.","marker":"[62]"}],"fun_headline_variants":["Displaced vertices probe top-heavy neutrinos to 12 TeV","ATLAS and far detectors reach 12 TeV for top-coupled HNLs","Top-heavy HNL EFT: ATLAS 12 TeV, MATHUSLA 4.5 TeV","Displaced vertex searches reach 12 TeV for top-coupled HNLs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the displaced-vertex selections at ATLAS and the shielding of the far detectors eliminate all Standard Model backgrounds, so three observed signal events suffice for a 95% confidence exclusion.","fun_headline_variants_meta":{"raw":{"variants":["Displaced vertices probe top-heavy neutrinos to 12 TeV","ATLAS and far detectors reach 12 TeV for top-coupled HNLs","Top-heavy HNL EFT: ATLAS 12 TeV, MATHUSLA 4.5 TeV","Displaced vertex searches reach 12 TeV for top-coupled HNLs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000476,"raw_usage":{"total_tokens":2389,"prompt_tokens":1005,"completion_tokens":1384,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":1294}},"tokens_in":621,"tokens_out":1384,"duration_ms":11559,"temperature":1.0,"reasoning_tokens":1294,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:12:17.482790+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate a full Standard Model background sample through the proposed ATLAS search chain — the 4/5/6-jet thresholds with $p_T > 90/65/55$ GeV, vertices with $4 < R_{xy} < 300$ mm, $|d_0| > 2$ mm, at least five tracks, and a vertex mass above 10 GeV — and count events at $L = 3$ ab$^{-1}$; more than a handful of events would invalidate the 3-event exclusions. For ANUBIS, the corresponding check is a complete background study of the service-shaft location, since its proximity to the IP could admit non-negligible backgrounds.","supporting_citations":[{"cited_title":"Heavy neutral leptons in effective field theory and the high-luminosity LHC","cited_arxiv_id":"2105.13851","evidence_quote":"Connects the pair-NR HNL production cross sections in the EFT and the Dirac/Majorana comparison."},{"cited_title":"Topportunities at the LHC: Rare Top Decays with Light Singlets","cited_arxiv_id":"2307.11154","evidence_quote":"Supplies the loop-induced three-body HNL decay rates for single-NR operators that dominate at low mass."}],"review_version":1}