{"id":"8f929802-272e-44ec-afed-a07f3012be92","arxiv_id":"2510.24499","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The paper projects that ATLAS with lead-lead ultra-peripheral collisions could exclude or discover light charged vector bosons decaying to heavy neutral leptons in the 5 to 105 GeV mass range.","lead":"This paper simulates ultra-peripheral heavy-ion and proton collisions at the LHC to see if ATLAS could detect a new charged boson that decays into a heavy neutral lepton, leaving two muons and missing energy. It claims the experiment could exclude or discover such particles in a low-mass range (5 to 105 GeV) that other searches have not covered.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Signal sensitivity relies on a single W+W- background; omitted gamma-gamma -> tau+tau- -> mu+mu-+4nu likely changes N_B and the claimed exclusions.","rationale":"The reader's weakest_assumption correctly identifies the incomplete background model and the trials factor from the random-search optimization as the main risks to the quantitative claims. I agree with that assessment, and I single out the omission of gamma-gamma -> tau+tau- with leptonic tau decays because it is a concrete, Standard-Model process with a larger expected rate than W+W- in the UPC environment, and it produces the exact same dimuon+MET signature. The paper's Fig. 2 and Section IV.A present only W+W- as background; no cross-section, cutflow, or yield is given for any other SM process. Since N_b after optimization is quoted as 1 for the most sensitive scenarios, even a small admixture of tau-pair events changes the conclusion. Additionally, the 2x10^5 random searches on the same Monte Carlo sample constitute a look-elsewhere effect; the true significance should be evaluated on an independent sample or with a trials correction. The HNL escape assumption is actually well-motivated by the Z2 symmetry (N_L is a dark matter candidate), so I do not object to it. Footnote 3 (Pb-Pb luminosity may be exceeded) only strengthens the experimental reach, but is irrelevant to the background deficit. Overall, the qualitative idea of using UPCs to search for sub-100 GeV charged vectors is plausible and not refuted by my concern, but the specific 95% C.L. exclusions and 5-sigma discovery are projections that need a full background budget and a validation of the cut-optimization procedure. Therefore the CONDITIONAL verdict stands unchanged.","tokens_in":16581,"tokens_out":10176,"duration_ms":87000,"concrete_test":"Using MadGraph with the same chff photon flux for Pb-Pb UPC at sqrt(s_NN) = 5.02 TeV, generate the Standard Model process gamma-gamma -> tau+tau- with tau -> mu nu nu, apply the basic cuts of Eq. 5, then apply the optimized cuts reported in Table II for the (20,10), (30,10), and (30,20) mass scenarios. Scale the surviving events to L_int = 3.48 nb^-1 with a 0.8 muon efficiency and compute N_B(tau). If N_B(tau) >= 2 in any scenario, recompute the significance from Eq. 6 with the combined background; if the (20,10) scenario no longer reaches exclusion at 95% C.L. at this luminosity, the central claim is not supported. If N_B(tau) is negligible, the check also validates the W+W--only approximation and allows the remaining overfitting concern to be treated separately.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The analysis assumes that gamma-gamma -> W+W- -> mu+mu- + neutrinos (Fig. 2a) is the only SM background in the dimuon+MET final state, and this assumption feeds Tables II and III and the significance formula (Eq. 6). In Pb-Pb UPC at sqrt(s_NN) = 5.02 TeV, the same photon-photon luminosity also produces gamma-gamma -> tau+tau-, and with BR(tau -> mu nu nu) ~ 17% per tau this yields the same dimuon + 4-neutrino signature with a rate expected to be much larger than W+W- at low invariant masses. This process is not generated, not included in the background set, and does not enter the cut optimization or the quoted N_b = 1. If even one or two tau-pair events survive the optimized cuts for e.g. (20,10) (alpha < 0.346, MET > 22.8 GeV), the 95% C.L. exclusions for (30,20), (30,10), (20,10) and the 5-sigma claim for (20,10) would not survive at L_int = 3.48 nb^-1. The paper itself concedes 'background arising from the intricate structure of lead nuclei' at the end of Section IV but provides no estimate of any process beyond W+W-. The 2e5 random searches over the same Monte Carlo sample further inflate significance, but the missing QED background is sufficient on its own to undermine the headline sensitivity.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a search for charged vector bosons V± of the Vector Scotogenic Model decaying into heavy neutral leptons N_L, using the dimuon+MET final state in ultra-peripheral Pb-Pb collisions at ATLAS (L_int = 3.48 nb^-1) and in pp UPCs at the HL-LHC. Signal and a γγ→W+W- background are generated with MadGraph using the chff photon flux, and a random search over 2×10^5 cut combinations on six observables is used to maximize a significance formula. The authors claim 95% C.L. exclusion of the mass scenarios (30,20), (30,10), and (20,10) GeV and a 5σ discovery for (20,10) GeV at current Pb-Pb luminosity, with broader coverage at the HL-LHC. The central quantitative claims rest on the background model, the optimized cuts, and the statistical interpretation of Eq. (6).","tokens_in":16828,"tokens_out":7584,"duration_ms":69607,"significance":"If the sensitivity projections are correct, the paper would open a low-mass window for charged-vector-boson and HNL searches that complements LEP-II, and the use of UPCs to suppress QCD backgrounds is a timely and interesting idea. The authors provide a concrete FeynRules/MadGraph implementation of the VSM and consider angular observables that are well suited to exclusive photon-induced events. However, the reported exclusions and discovery reach depend on a background estimate that contains only γγ→W+W-, on an uncalibrated optimization over many cut combinations, and on an unspecified exclusion criterion. These issues must be resolved before the quantitative claims can be accepted; the qualitative idea, namely that UPCs may probe light charged vectors in this final state, remains worth pursuing.","major_comments":[{"comment":"The background estimate for the dimuon+MET final state includes only γγ→W+W-→μ+μ-νν (Fig. 2a). In Pb-Pb UPCs at √s_NN = 5.02 TeV, γγ→τ+τ- has a much larger cross-section at the low invariant masses relevant here, and with BR(τ→μνν) ≈ 17.4% per tau it yields the same μ+μ- + 4ν signature. This process is not generated, not included in the 2×10^5 cut optimization, and not reflected in the N_B ≈ 1 values of Tables II and III. Because the 95% C.L. exclusions and the 5σ claim for (20,10) GeV follow from inserting those N_B values into Eq. (6), the central sensitivity result is not established until this irreducible background, along with other sources such as γγ→μ+μ- with mismeasured MET and the 'background arising from the intricate structure of lead nuclei' acknowledged at the end of Section IV, is quantified with a realistic systematic uncertainty.","section":"Section IV A, Fig. 2(a), Tables II and III"},{"comment":"The significance is evaluated after selecting the best of 2×10^5 random cut combinations on the same simulated samples. The N_S and N_B entries in Tables II and III are therefore maxima over a large search, and the quoted significances do not include a trials factor or a separation between optimization and evaluation samples. With N_B ≈ 1–6 and N_S ≈ 8–57, the probability of an upward fluctuation over 2×10^5 trials is non-negligible. The analysis should either fix the cuts a priori, use a validation sample for the final significances, or report a global significance that accounts for the scan.","section":"Section IV B, Eq. (6), Tables II and III"},{"comment":"The paper uses one significance formula (Eq. 6) for both the discovery and exclusion statements, but a 95% C.L. exclusion of a signal hypothesis is not equivalent to a signal significance threshold. The statistical criterion behind the exclusion contours in Figs. 7 and 13 is never stated, and no CL_s or profile-likelihood construction is provided. Please specify the limit-setting procedure, including how the systematic ε_sys^B is treated as a nuisance parameter and what threshold defines 'exclude at 95% C.L.'.","section":"Section IV B, Fig. 7, Section V B, Fig. 13"},{"comment":"The analysis applies a flat muon detection efficiency of 0.8 and does not include muon momentum resolution, trigger efficiency, MET resolution, or acceptance effects beyond |η|<2.4. The background counts N_B ≈ 1 are at the level where detector-level smearing and efficiency losses can change the outcome by factors of order unity. A detector-level or smeared analysis is needed before the projected exclusions and 5σ reach can be taken at face value.","section":"Section IV A, Tables II and III"}],"minor_comments":[{"comment":"The mass ranges quoted in the abstract are inconsistent with the body: the abstract says 5 GeV < M < 105 GeV while Section IV B restricts the Pb-Pb analysis to 5–50 GeV, and the abstract's 100–350 GeV range for the HL-LHC disagrees with Section V's 100 GeV < M_V± < 200 GeV.","section":"Abstract, Sections IV and V"},{"comment":"The column headers are ambiguous; entries such as cosθ = 0.034±0.96 and ΔR = 1.55±1.40 are not clearly defined as cut windows, fitted values, or cut bounds, and the pT μ and |E_T|miss cut values should be explicitly labeled as lower bounds.","section":"Tables II and III"},{"comment":"The random search algorithm is described only qualitatively; please provide the number of search iterations, the sampling distributions over the six observables, and the exact cost function so that the quoted efficiencies and cut values are reproducible.","section":"Section IV B"},{"comment":"There are numerous typos, including 'archive' for 'achieve', 'liminosity' for 'luminosity', 'exceded' for 'exceeded', and the title's 'A TLAS' spacing; a careful proofread is needed.","section":"Footnote 3 and general text"},{"comment":"The text says only mass values near the maximum γγ energy of about 160 GeV were selected, yet Fig. 3 displays masses up to 350 GeV; please clarify the selection criterion and the role of masses above 160 GeV.","section":"Fig. 3 and Section IV A"},{"comment":"The LEP-II exclusion argument cites the PDG review [62]; a direct citation to the LEP W-pair cross-section measurements would let the reader verify the claimed M_W' < 105 GeV constraint.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is a phenomenological sensitivity study with an incomplete background model. The missing γγ→τ+τ- contribution and the lack of a trials-factor correction are the main obstacles; both are fixable in a revision, but the quantitative claims should be re-derived. The paper is within the scope of a phenomenological journal, though the title and abstract overstate what is currently established by referring to 'proving' the existence of the new boson rather than projecting sensitivity. I would encourage the authors to add the missing background, validate the cut optimization on an independent sample, and state a proper exclusion statistical procedure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper argues that ATLAS UPC Pb-Pb collisions can probe a charged vector boson V± decaying to heavy neutral leptons (HNLs) via dimuon+MET, covering a mass window roughly 5–100 GeV that LEP-II didn't directly search and pp collisions drown in QCD background. The application of UPCs to this model is new, and the simulation setup (FeynRules/MadGraph with chff photon PDF) is standard and appropriately documented. The authors are also honest that their Pb-Pb luminosity assumption is conservative.\n\nThe soft spot is the background budget. The analysis includes only gamma-gamma -> W+W- -> mu+mu- + neutrinos as the SM background. But in UPCs, gamma-gamma -> tau+tau- has a much larger cross section at low invariant mass, and each tau decaying to mu+nu nu gives the same dimuon + 4-neutrino signature. This process is not generated, not included in the cut optimization, and not quoted in N_B. Even a few surviving events would wash out the claimed 95% C.L. exclusions for (30,20), (30,10), (20,10) and the 5-sigma claim for (20,10) at L_int = 3.48 nb^-1. The paper mentions 'background arising from the intricate structure of lead nuclei' but provides no estimate of any process beyond W+W-.\n\nThere's also an overfitting concern: the random search over 2e5 cut combinations is performed on the same Monte Carlo sample used for the significance evaluation, with no trials correction. That inflates the reported significances. The significance formula treats systematics very crudely, and no code or data is provided for reproduction.\n\nNone of this kills the core idea. The missing background is identifiable and fixable, and the paper's qualitative conclusion that UPCs open a new search window for light charged vectors and HNLs would likely survive an improved analysis. But the specific exclusions and 5-sigma claim should be treated as optimistic projections, not established sensitivities.\n\nThis paper deserves a serious referee, but with the expectation of heavy revision. I'd send it to peer review and ask the authors to add the tau-pair background (and ideally other SM processes like gamma-gamma -> mu+mu- with fake MET), re-optimize on an independent sample, and provide a trials-corrected significance. The idea is worth publishing in a corrected form.","headline":"Interesting proposal with a real physics gap: the missing tau+tau- background undercuts the headline sensitivity, but the UPC channel idea is worth pursuing.","tokens_in":17497,"tokens_out":3316,"would_cite":false,"duration_ms":29281,"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 claims that ATLAS, using ultra-peripheral lead-ion collisions, can directly test a new charged vector boson and heavy neutral lepton in the 5–100 GeV mass range, excluding three mass scenarios at 95% C.L.","keywords":["ultra-peripheral collisions","heavy ion collisions","charged vector boson","heavy neutral leptons","Vector Scotogenic Model","dimuon plus missing energy","LHC searches","beyond Standard Model"],"falsifier":"Generate the same signal and background with a detector-level simulation that adds exclusive $\\gamma\\gamma\\to \\tau^+\\tau^-\\to \\mu^+\\mu^- +$ neutrinos and fake or mismeasured missing energy, then apply the optimized cuts of Table II; if the expected background at $L_{\\rm int}=3.48$ nb$^{-1}$ is above about one event in the (30, 20), (30, 10), or (20, 10) GeV signal regions, the claimed 95% exclusions fail. In data, applying the same Table II cuts to the existing ATLAS Pb-Pb UPC sample would directly count the excess and settle whether the 5$\\sigma$ discovery claim holds.","tokens_in":16272,"feed_emoji":"⚛️","tokens_out":11784,"duration_ms":101007,"temperature":0.7,"pith_summary":"This paper argues that ultra-peripheral collisions of lead ions at the LHC, where two nuclei exchange quasi-real photons without touching, give the ATLAS experiment a nearly background-free way to search for new physics in the dimuon-plus-missing-energy channel. Within the Vector Scotogenic Model, it claims ATLAS can directly probe a new charged vector boson decaying to a heavy neutral lepton with masses between about 5 and 100 GeV, a region LEP-II did not directly search. At the current lead-ion luminosity, ATLAS could exclude the mass scenarios (30, 20), (30, 10), and (20, 10) GeV at 95% confidence and reach 5σ for (20, 10) GeV. The paper further claims that at the HL-LHC, proton-proton ultra-peripheral collisions would extend the reach to a 100–200 GeV charged boson with a heavy lepton up to 200 GeV, excluding all considered scenarios and reaching 5σ for most. If correct, this opens a low-mass window that conventional proton-proton searches cannot access because of QCD background.","feed_headline":"Lead-ion photon collisions could reveal a 20 GeV boson at ATLAS","feed_subtitle":"It offers a clean channel to probe the 5-100 GeV mass range left untested by LEP-II.","key_machinery":"The load-bearing object is the Vector Scotogenic Model's charged vector boson doublet $V_\\mu=(V^\\pm_\\mu,V^0_\\mu)^T$ and its interaction $\\mathcal{L}=-\\sum_{i=e,\\mu,\\tau}\\beta_i \\bar{L}_i\\gamma^\\mu \\tilde V_\\mu N_L + \\mathrm{h.c.}$, with $\\beta_2=0.5$ adopted in the scans; this couples the new boson to SM leptons and the heavy neutral lepton. The search machinery is the ultra-peripheral photon-fusion environment, where two quasi-real photons from lead nuclei fuse into $V^+V^-$ (signal) or $W^+W^-$ (background), together with a random-search cut optimizer that maximizes a significance defined in Eq. (6) over six muon observables: $p_T(\\mu^\\pm)$, missing transverse energy, dimuon invariant mass, $\\Delta R$, acoplanarity, and $\\cos\\theta$. These cuts exploit the back-to-back, angularly separated muons produced by light $V$ decays, while the $W$-pair background is angularly unstructured.","core_discovery":"The central claim is that the exclusive two-photon process $\\mathrm{Pb}+\\mathrm{Pb}\\to \\mathrm{Pb}^{(*)} + \\mathrm{Pb}^{(*)} + \\mu^+\\mu^- + \\not E_T$ carries a usable signal for a charged vector boson $V^\\pm$ decaying into a heavy neutral lepton $N_L$, with the main irreducible background coming from $\\gamma\\gamma\\to W^+W^-\\to \\mu^+\\mu^-\\nu\\bar{\\nu}$ vector-boson fusion. Using $L_{\\rm int}=3.48$ nb$^{-1}$ of lead-ion UPC data, the paper claims ATLAS can exclude the mass scenarios $(M_{V^\\pm},M_{N_L})=(30,20)$, $(30,10)$, and $(20,10)$ GeV at 95% C.L. and reach 5$\\sigma$ for $(20,10)$ GeV. It further claims that at the HL-LHC with proton-proton ultra-peripheral collisions and 150 fb$^{-1}$, the exclusion extends to $100< M_{V^\\pm}<200$ GeV and $5< M_{N_L}<200$ GeV, with 5$\\sigma$ discovery possible for most scenarios.","pith_inferences":["The quoted significances are computed on the same Monte Carlo samples used by the random-search cut optimizer; a blinded, independent validation sample would be the natural next check before claiming discovery power.","The background model includes only $\\gamma\\gamma\\to W^+W^-$; adding exclusive $\\gamma\\gamma\\to \\tau^+\\tau^-$ with leptonic tau decays, or detector-level fake missing energy, would likely raise the background above the roughly one-event level and shrink the excluded region.","The same dimuon-plus-MET selection is a reusable template: any model producing a charged resonance that decays to a muon plus an invisible or long-lived fermion could be constrained by reinterpreting these cut sets.","The clean separation between signal and background in acoplanarity and $\\Delta R$ suggests that a shape-based or machine-learned discriminant on these observables could outperform the box cuts and extend the reach beyond the specific mass points quoted."],"forward_implications":["If the (30, 20), (30, 10), and (20, 10) GeV scenarios are absent in the current 3.48 nb$^{-1}$ Pb-Pb UPC sample, ATLAS would already exclude a part of the Vector Scotogenic Model parameter space that no other experiment has directly tested.","A 5$\\sigma$ excess in the (20, 10) GeV channel would be direct evidence of a new charged vector boson and a heavy neutral lepton produced through photon fusion, without the QCD background that obscures conventional proton-proton searches.","If the Pb-Pb search comes back empty, the HL-LHC proton-proton UPC program with 150 fb$^{-1}$ would extend the exclusion to 100–200 GeV for $M_{V^\\pm}$ and up to 200 GeV for $M_{N_L}$, covering most of the remaining low-mass window.","The same six-observable cut sets could be published as a reusable analysis recipe for light charged resonances decaying to dileptons plus missing energy.","The mass hierarchy matters: the analysis shows the smallest scenarios, especially (20, 10) GeV, give the strongest signal, so future searches should prioritize low $V^\\pm$ masses with significantly lighter heavy neutral leptons."],"supporting_citations":[{"why":"Defines the Vector Scotogenic Model whose $V^\\pm$ and $N_L$ are the objects being searched for.","marker":"[49–52]"},{"why":"Supplies the numerical values of the VSM couplings and parameters adopted in the cross-section scans.","marker":"[51]"},{"why":"Provides the UPC photon flux distribution (chff) used to generate signal and background events.","marker":"[67]"},{"why":"Supplies the Monte Carlo generation chain that produces the signal and background pseudo-events.","marker":"[63–66]"},{"why":"Gives the LEP-II-motivated mass bound that defines the unexplored low-mass window the paper targets.","marker":"[62]"},{"why":"Sets the $\\gamma\\gamma$ center-of-mass reach and integrated luminosities used to choose the Pb-Pb and HL-LHC search ranges.","marker":"[31–33, 66]"}],"fun_headline_variants":["Photon fusion at ATLAS could reveal a 20 GeV boson","Ultra-peripheral collisions may expose a new charged boson","ATLAS UPCs can exclude a 20 GeV boson scenario","New boson search: lead-ion photons at ATLAS","UPCs at LHC might detect a boson at 20 GeV"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes the dimuon-plus-MET sample after the basic cuts contains only the $\\gamma\\gamma\\to W^+W^-$ background and that the cuts optimized on the same Monte Carlo sample describe real detector data; if either assumption fails by even a few events, the claimed 95% exclusions and 5$\\sigma$ discovery do not survive.","fun_headline_variants_meta":{"raw":{"variants":["Photon fusion at ATLAS could reveal a 20 GeV boson","Ultra-peripheral collisions may expose a new charged boson","ATLAS UPCs can exclude a 20 GeV boson scenario","New boson search: lead-ion photons at ATLAS","UPCs at LHC might detect a boson at 20 GeV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000575,"raw_usage":{"total_tokens":2796,"prompt_tokens":1107,"completion_tokens":1689,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":723,"completion_tokens_details":{"reasoning_tokens":1598}},"tokens_in":723,"tokens_out":1689,"duration_ms":12785,"temperature":1.0,"reasoning_tokens":1598,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:42:19.824430+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Generate the same signal and background with a detector-level simulation that adds exclusive $\\gamma\\gamma\\to \\tau^+\\tau^-\\to \\mu^+\\mu^- +$ neutrinos and fake or mismeasured missing energy, then apply the optimized cuts of Table II; if the expected background at $L_{\\rm int}=3.48$ nb$^{-1}$ is above about one event in the (30, 20), (30, 10), or (20, 10) GeV signal regions, the claimed 95% exclusions fail. In data, applying the same Table II cuts to the existing ATLAS Pb-Pb UPC sample would directly count the excess and settle whether the 5$\\sigma$ discovery claim holds.","supporting_citations":[{"cited_title":"Probing Left-handed Heavy Neutral Leptons in the Vector Scotogenic Model","cited_arxiv_id":"2211.09753","evidence_quote":"Supplies the numerical values of the VSM couplings and parameters adopted in the cross-section scans."},{"cited_title":"Vidovic, M","cited_arxiv_id":null,"evidence_quote":"Provides the UPC photon flux distribution (chff) used to generate signal and background events."}],"review_version":2}