{"id":"b53c7809-1f33-4682-b65b-f99a9674f1f2","arxiv_id":"2607.23166","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A topology-based six-muon search at the HL-LHC could probe vector-like muon dark-portal partners up to about 1.9 TeV.","lead":"A collider study shows that heavy 'vector-like' muons connected to a dark sector can produce six clean muons per collision. A dedicated search could reach masses near 1.9 TeV at the HL-LHC with essentially zero Standard Model background.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Six-muon efficiencies for light V' are likely inflated because Delphes does not model track merging for ΔR~0.003-0.01 dimuons; this overstates the claimed reach.","rationale":"The paper's central claim is that a dedicated six-muon search is nearly background-free and can probe m_mu' up to ~1.9 TeV. For that to hold, signal events must survive the detector-level selection at the quoted efficiency, and backgrounds must remain far below one event. I examined both pillars. The background side is actually robust: after the topology cuts (Table VI), the total Category-1 background at 3000 fb^-1 is 3.8e-5 events, dominated by irreducible 6mu (2.2e-5) and ZZW+W (1.6e-5), with heavy-flavour ttZ/4t below 1.5e-7. Even a 1000-fold error in the Appendix A heavy-flavour filter would leave the background below ~1e-2 events, so the reader's weakest assumption, while unvalidated, is not the limiting factor. The signal side is less secure. For mV' of order 1-10 GeV, the V'->mu+mu- muons have ΔR of order 0.003-0.017 and cannot be resolved as two isolated muons by current LHC detectors. The paper uses Delphes, which does not simulate track merging, and obtains ~90% six-muon efficiency for BP1 (Table IV) and flat 80-90% efficiency across mV' (Fig. 18). This likely overestimates the signal exactly in the light-mediator region where the six-muon branching fraction is largest. Since the 1.9 TeV reach uses 'favourable spectra' that include mV'=10 GeV, a real detector simulation could reduce the reach. I therefore keep the CONDITIONAL verdict — the reader reached the same verdict but for a different, less consequential reason. My concern is a concrete detector-modeling issue that would need to be settled by a tracking-aware simulation before the projected reach is used.","tokens_in":26313,"tokens_out":20546,"duration_ms":210910,"concrete_test":"For BP1 (mV'=1 GeV), BP2 (mV'=10 GeV), and the m_mu'=1.8 TeV, mV'=10 GeV, mHD=100 GeV high-yield point used in Fig. 17, re-run Delphes after (i) rejecting/merging any muon pair with ΔR<0.02 and (ii) recomputing muon isolation with the companion muon included in the cone sum. If the six-muon selection efficiency falls by more than ~20% relative to Fig. 18/Table IV, the broad-parameter-space and ~1.9 TeV reach claims are overestimated and should be rescaled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires signal events to survive a six-muon, isolated-lepton selection (Sec. IV.D). For the light-mediator benchmarks BP1/BP2 (mV'=1,10 GeV at m_mu'=1200 GeV), the V' from mu'->mu V' is highly boosted, so the two muons from V'->mu+mu- have ΔR ~ mV'/p_T(V') ~ 0.003-0.017 (see Fig. 9, left). Real LHC muon reconstruction cannot resolve two muons separated by <~0.01-0.02, and standard isolation would also fail because each muon sees the other in its isolation cone. Delphes 3.5 does not simulate track merging, yet Table IV quotes N_mu>=6 efficiency 0.906 for BP1 and Fig. 18 shows ~80-90% six-muon efficiency flat down to mV'~0.1 GeV, which is physically implausible. The projected reach is defined by N_sig=3 after these efficiencies, and the 'favourable spectra' in Fig. 17 include mV'=10 GeV points; an overestimated muon-reconstruction efficiency directly inflates the ~1.9 TeV reach. Note that the reader's heavy-flavour concern is less load-bearing: Table VI shows the reconstructed background is dominated by irreducible 6mu (2.2e-5) and ZZW+W (1.6e-5) events at Category 1, with ttZ/4t <1.5e-7, so even a large error in the Appendix A filter would not overturn the background-free conclusion.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies six-muon final states from pp → μ'+μ'− production in the Muonic Portal to Vector Dark Matter (MPVDM) model. After fixing the model parameters to benchmark points and adopting the relation g_D = 10 m_V'/m_μ' from the authors' earlier relic-density study, the authors compute branching fractions for the 4μ, 6μ, 8μ and 10μ channels, identify the 6μ final state as the most promising target, and develop a topology-based reconstruction with two exclusive categories: a symmetric (3+3)μ category and an asymmetric (1+5)μ category. Signal and SM backgrounds are simulated with CalcHEP/MG5_aMC + Pythia + Delphes, including a dedicated repeated-decay filter for rare heavy-flavour muons. A CheckMATE recast of CMS-SUS-16-039 is used to estimate current Run-2 constraints, and an HL-LHC projection at 3000 fb−1 is made using a background-free N_sig=3 criterion, giving a reach up to about 1.9 TeV for favourable spectra. The central claims are that the 6μ channel is essentially background-free after topology reconstruction and that a dedicated search substantially extends the existing Run-2 coverage.","tokens_in":26668,"tokens_out":5293,"duration_ms":54732,"significance":"If the efficiency and background estimates are reliable, this is a useful and original contribution. The paper identifies a six-lepton cascade topology from vector-like lepton pair production that has not been systematically exploited, proposes an explicit assignment-based reconstruction rather than a black-box classifier, and provides public model implementation and supplementary code. The heavy-flavour filtering procedure is a thoughtful attempt to address a difficult Monte Carlo problem. The main value is in demonstrating a concrete, reconstructable multilepton target with a quantitative LHC projection. However, the central reach claim depends on detector-level muon reconstruction efficiencies that are likely overestimated for light mediators, and the LO cross-section treatment lacks the systematic uncertainties needed to support a precise mass reach.","major_comments":[{"comment":"The six-muon reconstruction efficiencies for light dark vectors are not physical. For BP1/BP2 with mV'=1,10 GeV at mμ'=1200 GeV, the two muons from V'→μ+μ− have ΔR ~ mV'/p_T(V') ~ 0.003–0.017 (Fig. 9, left). Delphes 3.5 does not simulate track merging or isolation-cone overlap, yet Table IV quotes N_mu≥6 efficiencies of 0.906 and 0.863 and Fig. 18 shows ~80–90% efficiency flat down to mV'~0.1 GeV. In a real muon system these pairs would be merged into a single track or each muon would fail isolation because the other lies inside its isolation cone. Since the reach is defined by N_sig=3 after these efficiencies (Eq. 4.15) and the 'favourable spectra' in Fig. 17 include mV'=10 GeV, the 1.9 TeV projection for light-mediator scenarios is overestimated. Please re-evaluate with a track-merge veto (e.g. ΔR>0.02) or an equivalent detector-level overlap treatment and show how the reach changes.","section":"IV.C, IV.D, Table IV, Fig. 18"},{"comment":"All signal and background cross sections are leading order, with the scale fixed to Q=m_μ' and no PDF/scale uncertainties or K-factors. The Drell–Yan signal cross section falls steeply with mμ', and the exclusion criterion is a three-event threshold. A typical DY NLO K-factor of ~1.2–1.3, or even the PDF uncertainty alone, changes the event count by tens of percent and therefore shifts the derived 1.9 TeV reach noticeably. The paper should either include NLO signal cross sections with scale/PDF uncertainties, or present the reach as a band with an explicit statement of the LO-induced systematic uncertainty.","section":"IV.A, Fig. 5, Eq. 4.15"}],"minor_comments":[{"comment":"The heavy-flavour filtering efficiencies (e.g. ϵ_HF(t-tbar)=3.0×10−11) are quoted without statistical uncertainties and without validation against an independent generator or data. This is not load-bearing for the background-free conclusion because Table VI shows the post-reconstruction background is dominated by irreducible 6μ and ZZWW, with heavy-flavour contributions below ~10−7 events, but the procedure should still be documented with uncertainties for reproducibility.","section":"Appendix A, Eqs. (A9)–(A14)"},{"comment":"The 'corrected CMS Delphes card' used for the CheckMATE recast is not described in detail. Please specify the correction and, if possible, validate it against a published CMS efficiency or resolution curve.","section":"IV.E"},{"comment":"The 'Maximal Signal Yield' curve is not clearly defined in the text. If it is the pointwise maximum over the (mV',mHD) plane, the authors should state which parameter values produce it, particularly because some of the light-mediator points are affected by the track-merging issue in Major Comment 1.","section":"Fig. 17"},{"comment":"The claim that the six-lepton signature from vector-like-lepton pair production 'has not previously been explored at the LHC' should be supported by explicit references to recent ATLAS/CMS multilepton searches and existing VLL phenomenology, so the novelty statement is checkable.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper is promising and the topic is timely, but the central reach estimate is currently vulnerable to the unresolved track-merging issue for light dark vectors. The heavy-flavour background concern is much less serious than it first appears because the reconstructed background is dominated by irreducible 6μ and ZZWW. I would send the revised version to a phenomenological referee with experience in detector-level muon reconstruction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading, with one caveat that matters. The paper identifies the six-muon final state as the most promising channel in the MPVDM model and builds an explicit topology-based reconstruction (Category 1 for 3+3, Category 2 for 1+5) that is genuinely new and sensible. It ships code and model files, uses a CheckMATE recast for current constraints, and is honest about LO cross sections, the indicative mass window, and the conservative treatment of unequal dimuon masses. That is good practice.\n\nThe soft spot that bothers me most is the signal efficiency for light V'. For BP1 (mV'=1 GeV) and BP2 (mV'=10 GeV), the muons from V'->mu+mu- have DeltaR around 0.003–0.017. Real LHC muon systems cannot resolve pairs separated by much less than a few hundredths, and the isolation requirement alone would make each muon fail when the other sits inside its cone. Delphes 3.5 does not model track merging, yet Table IV quotes N_mu>=6 efficiencies of 0.85–0.9 for these benchmarks, and Fig. 18 shows a flat high efficiency down to small mV'. That is physically implausible. Since the HL-LHC reach curves in Fig. 17 include mV'=10 GeV points, the 1.9 TeV claim is probably inflated. The authors should either apply a realistic DeltaR separation or isolation treatment, restrict the reach claim to mediator masses where the dimuons are resolved, or show that the maximal-yield curve is driven by points without this problem.\n\nThe heavy-flavour background filter in Appendix A is indeed unvalidated, but I agree with the stress-test note that it is not load-bearing: Table VI shows the irreducible 6mu and ZZW+W processes dominate the post-selection background, so even a large error in the ttZ/4t efficiencies would not overturn the background-free conclusion. LO cross sections with no scale/PDF uncertainties are a real but minor concern — they could shift the mass reach by maybe ten to thirty percent, which matters but does not change the qualitative message. The novelty claim that the six-lepton signature has never been explored at the LHC is also stated too strongly without a systematic literature survey.\n\nFor whom: phenomenologists working on vector-like leptons or dark portals, and experimental colleagues thinking about high-multiplicity muon searches. It deserves a serious referee: the method is explicit, the code is available, and the topology-based reconstruction is a useful template. But the referee should ask for a realistic treatment of collimated muons before the reach claim is taken at face value.","headline":"A well-executed, transparent collider study whose headline 1.9 TeV reach is likely optimistic because the detector simulation does not handle collimated dimuons from light dark vectors.","tokens_in":27291,"tokens_out":3870,"would_cite":true,"duration_ms":40928,"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 paper shows that six-muon cascades from pair-produced heavy vector-like muons are an essentially background-free signature at the HL-LHC, and that a dedicated reconstruction can probe vector-like muon masses up to about 1.9 TeV.","keywords":["vector-like leptons","fermionic portal dark matter","six-muon final state","multilepton searches","HL-LHC","dark vector boson","cascade decay reconstruction","heavy-flavour backgrounds"],"falsifier":"Run an independent Monte Carlo simulation, or a top-pair control region in existing LHC data, to measure the probability that a top-quark-pair event passes the six-muon selection with four additional isolated muons (p_T > 10 GeV, |η| < 2.5, isolation). If that efficiency times the top-pair cross section exceeds about 10⁻⁵ fb — more than roughly 0.03 expected events at 3000 fb⁻¹ — the background-free premise fails and the 1.9 TeV reach estimate would need recomputation. A secondary check: a dedicated six-muon search should show a clustered trimuon invariant-mass peak at the parent mass; a flat","tokens_in":26091,"feed_emoji":"⚛️","tokens_out":8363,"duration_ms":78399,"temperature":0.7,"pith_summary":"This paper identifies the six-muon final state as the most powerful collider signature of a class of dark-matter models in which heavy vector-like muons are the fermionic portal to a dark gauge sector. The claim is that this channel is simultaneously generic, remaining sizeable across a broad region of the dark-sector mass plane, and exceptionally clean, because after a topology-based reconstruction the Standard Model background is negligible. The paper develops an explicit reconstruction that assigns reconstructed muons to the two cascade branches using repeated dimuon and trimuon resonance structures, and shows that a dedicated search at the High-Luminosity LHC could probe vector-like muon masses up to about 1.9 TeV for favourable spectra. A sympathetic reader would care because this turns an exotic multilepton signature into a concrete, nearly background-free search target that carries measurable information about the dark sector.","feed_headline":"Six-muon cascades could reveal the dark portal to 1.9 TeV","feed_subtitle":"Heavy vector-like muons decaying through a dark sector leave a nearly background-free HL-LHC signature.","key_machinery":"The load-bearing structure is the decay chain μ' → μV' or μH_D with V' → μ⁺μ⁻, H_D → μ⁺μ⁻, and H_D → V'V' → 4μ, whose combinatorics produce the repeated dimuon, trimuon, and five-muon resonance patterns. The machinery that carries the argument is the topology-based reconstruction: a χ² assignment procedure using logarithmic mass-difference compatibility terms that pairs the six muons into two charge-compatible trimuon systems (Category 1) or into one five-muon system with a single isolated muon on the other branch (Category 2), followed by a ±20% parent-mass window. The classification of single-chain probabilities P1, P3, P5 with the multiplicity branching fractions B_2μ = P1², B_4μ = 2P1P3,","core_discovery":"In the muonic realisation of the fermionic portal to vector dark matter, Drell–Yan pair production of the heavy vector-like muon μ' is followed by cascade decays through the dark vector V' and the dark scalar H_D, producing final states with four, six, eight, or ten visible muons. The paper's central claim is that the six-muon channel is the best target: it combines a sizeable branching fraction over a broad region of the (m_V', m_H_D) plane, a very clean experimental signature, and enough internal resonance structure to reconstruct the intermediate states and the parent heavy lepton. Six muons arise from either the symmetric (3+3)μ topology, in which each μ' chain yields three muons, or the","pith_inferences":["Extending the topology-based strategy to tau or electron final states would test whether the fermionic portal is flavour-universal; a six-tau channel trades rate for a harder background but carries the same repeated-resonance structure.","The Category-1 reconstruction assumes equal dimuon masses on the two branches — a choice the paper flags as conservative — so a third category allowing different intermediate masses on each chain would likely recover lost events and push the reach beyond 1.9 TeV.","A data-driven cross-check of the heavy-flavour background, such as counting top-pair events with four additional isolated muons in existing data, is the most direct way to test the 3×10⁻¹¹ efficiency on which the background-free premise rests.","The model predicts correlated signatures across regimes: the six-muon channel studied here, the two-muon plus missing-energy regime of earlier work, and displaced-vertex signatures at smaller coupling; observing one and not the others would constrain the μ'–μ_D mass splitting."],"forward_implications":["A dedicated six-muon search at the HL-LHC can probe vector-like muon masses up to about 1.9 TeV for favourable spectra, substantially extending the low-mass region already constrained by generic Run-2 multilepton searches.","Because the analysis is effectively background-free, the excluded cross section scales roughly as 1/L, so increases in luminosity translate directly into mass reach without needing tighter event selections.","The fully visible (3+3)μ topology means the search must stay inclusive in missing transverse momentum; an E/T requirement would discard precisely the events in which both cascade branches can be reconstructed.","If a signal is observed, the reconstructed dimuon, trimuon, and five-muon masses measure the dark-sector spectrum — the masses of V', H_D, and the parent μ' — turning the signature into a spectroscopy tool rather than a counting excess.","The same reconstruction logic applies to any pair-produced vector-like fermion that decays through resonant two-body steps, so six-lepton cascade topologies become reconstructable targets beyond this specific model."],"fun_headline_variants":["Six muons reveal vector dark matter via fermionic portal","Heavy muon pairs decay to six muons, probing dark vector up to 1.9 TeV","Six-muon cascade from vector-like muons: cleanest dark matter probe","New six-muon signature targets fermionic portal to dark vector","Vector-like muon cascades: six muons beat four and ten"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The entire sensitivity estimate rests on the Monte Carlo repeated-decay procedure correctly predicting extremely rare heavy-flavour muon rates — in particular, that a top-quark-pair event supplies four additional isolated muons with a probability of 3×10⁻¹¹; if the true rate were orders of magnitude larger, the 'negligible background' conclusion would break.","fun_headline_variants_meta":{"raw":{"variants":["Six muons reveal vector dark matter via fermionic portal","Heavy muon pairs decay to six muons, probing dark vector up to 1.9 TeV","Six-muon cascade from vector-like muons: cleanest dark matter probe","New six-muon signature targets fermionic portal to dark vector","Vector-like muon cascades: six muons beat four and ten"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000238,"raw_usage":{"total_tokens":1434,"prompt_tokens":914,"completion_tokens":520,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":418}},"tokens_in":658,"tokens_out":520,"duration_ms":5594,"temperature":1.0,"reasoning_tokens":418,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T03:24:32.339845+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an independent Monte Carlo simulation, or a top-pair control region in existing LHC data, to measure the probability that a top-quark-pair event passes the six-muon selection with four additional isolated muons (p_T > 10 GeV, |η| < 2.5, isolation). If that efficiency times the top-pair cross section exceeds about 10⁻⁵ fb — more than roughly 0.03 expected events at 3000 fb⁻¹ — the background-free premise fails and the 1.9 TeV reach estimate would need recomputation. A secondary check: a dedicated six-muon search should show a clustered trimuon invariant-mass peak at the parent mass; a flat","supporting_citations":[],"review_version":1}