{"id":"d6a4dc46-a003-47fc-92e5-7fa20fc64622","arxiv_id":"2608.01251","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"CMS sees no evidence for a top-philic Z' boson decaying to top-quark pairs and sets the tightest published cross-section limits for masses between 0.5 and 3 TeV.","lead":"The CMS experiment searched for a hypothetical top-philic Z' boson in LHC proton-proton collisions, using 13 TeV data recorded from 2016 to 2018. No signal was found, and the new cross-section limits on this particle are the most stringent available for masses from 0.5 to 3 TeV.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CR-to-SR background transfer relies on a single b-tag-inverted control region validated only in simulation; a data/MC transfer-factor mismatch could bias the normalization and hence the headline limits.","rationale":"The paper is a standard CMS search with appropriate signal and background modeling, and the limits appear internally consistent. The reader's conditional verdict is driven by a title/abstract discrepancy (13.6 vs 13 TeV), which is a trivial metadata error. The more substantive issue is the CR-to-SR transfer. It is common in HEP searches to validate such transfers with simulation, and the paper does quote the relevant systematic uncertainties (b-tagging scale factors, top-tagging uncertainties, and a 10-20% background estimation uncertainty), so this is not a fatal flaw. Nevertheless, the central claim of 'most stringent limits to date' rests on the limits being unbiased, and the CR transfer is the least directly tested ingredient of the background model. The suggested checks--an SR-only fit and a data/MC transfer-factor comparison--would settle whether the concern lands. I therefore agree with the reader's CONDITIONAL verdict and do not propose changing it; the condition should include addressing the title error and providing transfer validation.","tokens_in":38920,"tokens_out":10298,"duration_ms":91145,"concrete_test":"Recompute the 95% CL upper limits for all four width hypotheses using a fit to the SR data only (no CR constraint), and compare the fitted background normalization and the resulting limits with the nominal values. Also evaluate the SR-to-CR background yield ratio in data and in MC in a top-tag-inverted sideband; if the ratio differs from MC by more than the quadrature sum of the b-tagging and transfer uncertainties, or if the SR-only limits shift outside the 68% expected band, the CR-based normalization is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption is the transfer of the background normalization from the zero-b-tag control region (CR) to the signal region (SR) with at least one b-tagged jet, described in Section 4. A single freely floating parameter scales the sum of all simulated backgrounds simultaneously in SR and CR. The paper states that the background composition and the m_rec(Z') distribution are consistent between the two regions, but this is only a simulation-based check; no data-driven closure test is shown. Because the SR and CR differ precisely in b-tag multiplicity, the SR/CR background ratio is sensitive to the modeling of b-tagging efficiency and the mistag rate for light-flavor jets. A multiplicative miscalibration of that ratio would directly bias the fitted background normalization in the SR, shifting the 95% CL upper limits on sigma(pp->ttZ')*B and the derived mass exclusions. Since the central claim is that these limits are the most stringent to date, the accuracy of the CR-to-SR transfer is the linchpin of the result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a CMS search for a top-philic Z' boson produced in association with a top-antitop pair and decaying into a tt pair, using 138 fb^-1 of 13 TeV pp collision data. Events are selected with a single lepton, at least six small-radius jets, HT > 700 GeV, at least one b-tagged jet, and at least two large-radius jets tagged as boosted hadronic top decays by the ParticleNet algorithm. The invariant mass of the two highest-pT top-tagged jets is used as the discriminant. The dominant tt background is constrained with a zero-b-tag control region via a simultaneous SR+CR maximum-likelihood fit with a common background normalization parameter. No significant excess is found; 95% CL upper limits on sigma(pp->ttZ')B(Z'->tt) are set for relative widths of 4, 10, 20, and 50%, with observed limits ranging from 170 to 3 fb for the 4% width, and mass exclusions of 560, 850, and 1130 GeV for the 10, 20, and 50% widths. The paper claims these are the most stringent limits to date on a top-philic Z' boson.","tokens_in":39059,"tokens_out":8760,"duration_ms":76577,"significance":"If the results are correct, the paper provides the most stringent constraints to date on a top-philic Z' boson in the adopted simplified model, extending previous ATLAS limits to lower masses and to widths of 4-50%. The analysis uses the full Run 2 dataset, a state-of-the-art jet tagger, and a standard CMS statistical framework; the HEPData record and detailed systematic accounting support reproducibility. The main qualification is that the background estimate in the signal region relies on a transfer from a zero-b-tag control region that is validated only in simulation, so the central exclusion claim is vulnerable to a data/MC mismatch in the CR/SR background ratio.","major_comments":[{"comment":"The CR-to-SR background transfer is validated only in simulation. The text states that the background composition and m_rec(Z') shape are consistent between the two regions, but the only supporting evidence shown is the prefit CR distribution in Fig. 2; no data-driven closure test is provided. Because the SR requires at least one b-tagged small-radius jet and the CR requires exactly zero, the ratio of background yields in CR and SR is directly sensitive to the data/MC modeling of b-tagging efficiency and the light-flavor mistag rate. A single common normalization parameter fitted to CR+SR will absorb any data/MC difference in the CR yield and translate it into a biased SR prediction if the simulated CR/SR ratio is wrong. The systematic uncertainties quoted in Section 5 (b-tagging scale factors of 0.3-4.9% for the background, and the 10-20% background normalization uncertainty from the fit) are not demonstrated to cover this transfer-ratio effect. The authors should add a data-driven validation, for example a region with exactly one b-tagged jet or a top-tag-inverted sideband, and quantify the resulting systematic uncertainty on the SR background normalization.","section":"Section 4 (Background estimation)"},{"comment":"The mass exclusions of 560, 850, and 1130 GeV are derived by comparing the observed upper limits with the LO theoretical cross-section curve from Ref. [44], but the theory curve is shown without any uncertainty band. Section 5 quotes the signal renormalization/factorization scale uncertainty as approximately 40%. The paper should state whether these quoted mass exclusions are robust under this theory uncertainty, or display the theory band in Fig. 4 so that the reader can assess the sensitivity of the mass bounds to the missing k-factor and scale choices.","section":"Section 6 (Results) and Fig. 4"}],"minor_comments":[{"comment":"The title states sqrt(s) = 13.6 TeV, but the analysis uses 2016-2018 data at sqrt(s) = 13 TeV; this inconsistency should be corrected.","section":"Title and Abstract"},{"comment":"The text 'The lower limits on the ttZ' production cross section... range between 170 and 3 fb' should read 'upper limits', since the paper sets upper limits on the cross section and lower limits on the Z' boson mass.","section":"Section 7 (Summary)"},{"comment":"The term 'anogenophilic' is nonstandard and undefined; it should be replaced with a standard phrase such as 'third-generation-philic' or defined at first use.","section":"Section 1 (Introduction)"},{"comment":"The correlation structure of the background normalization nuisance parameters is unclear: the text states that a single freely floating parameter scales the total background in SR and CR, but also that the normalization of each individual background process is allowed to vary within uncertainties. Please clarify whether these are separate constrained nuisance parameters or a single parameter, and state the correlation assumptions between processes and data-taking years.","section":"Section 4 and 5"},{"comment":"To support the claim that the limits are 'the most stringent to date', the paper should directly compare the present observed limits with the ATLAS results of Ref. [49] in the overlapping mass and width region, rather than only quoting the ATLAS ranges in the introduction.","section":"Section 6 (Results)"}],"recommendation":"major_revision","confidential_remarks":"The analysis appears technically sound and the central search strategy is standard for CMS. The main issue is the CR-to-SR transfer validation, which is load-bearing for the headline limits; this is fixable with additional closure tests or a systematic uncertainty. The title/abstract energy mismatch and the cross-section limit direction typo in Section 7 should be corrected during revision. No concerns about scope or citation patterns; the paper fits JHEP."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a solid, conventional CMS search. What is actually new: first CMS constraints on a top-philic Z' in the ttZ' channel for masses down to 0.5 TeV, with four width assumptions (4, 10, 20, 50%). The previous ATLAS search stopped at 1 TeV, so this fills the low-mass gap. The reported limits (170–3 fb for 4% width over 0.5–3 TeV, mass exclusions of 560, 850, and 1130 GeV for 10, 20, and 50% widths) are plausible and clearly compared with the theoretical prediction.\n\nThe analysis is done to the usual CMS standard: MC signal and background, ParticleNet for boosted tops, a zero-b-tag control region to constrain the dominant tt background, a simultaneous SR+CR fit with nuisance parameters, and HEPData tables. The data agree with the SM. I see no statistical or procedural abuse.\n\nThe soft spots are minor but real. First, the title says sqrt(s)=13.6 TeV while the abstract, body, and summary consistently say 13 TeV with 2016–2018 data. That is an obvious error and must be fixed. Second, the CR-to-SR transfer: the zero-b-tag region is used to normalize the background in the signal region with at least one b-tag. The paper states the kinematic shapes and composition are consistent in simulation, but there is no data-driven closure test. Because the two regions differ precisely by b-tag multiplicity, a mismodeled b-tag efficiency or mistag rate could bias the normalization and shift the limits. This is a legitimate concern, but it is tempered: the analysis includes b-tagging scale-factor uncertainties and the overall background normalization is free in the fit. I would not call it load-bearing; I would ask the authors to add a data/MC comparison in a signal-depleted subset of the zero-tag region (e.g., high m_reco or low top-tag multiplicity) or otherwise demonstrate closure. Also, \"anogenophilic\" in the introduction looks like a typo.\n\nOverall, the physics conclusion is credible and the limits are likely the most stringent to date in this channel for the low-mass range. The paper deserves a serious referee. I would accept it with minor revisions: fix the title energy, address the CR transfer with a closure check or an explicit discussion of its limitation, and clean up small typographical issues.","headline":"A standard, well-executed CMS search that extends top-philic Z' limits down to 0.5 TeV and across four width hypotheses; the main fix needed is the title's wrong center-of-mass energy, and the CR-to-SR transfer assumption deserves a data closure check but is not a fatal flaw.","tokens_in":39620,"tokens_out":1851,"would_cite":true,"duration_ms":20170,"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":"A search of 138 fb$^{-1}$ of CMS data finds no top-philic $Z'$ boson; the new 95% CL upper limits reach 3 fb and, depending on the assumed width, exclude masses below 560$-$1130 GeV.","keywords":["top-philic Z' boson","boosted top quarks","four-top-quark final state","ParticleNet","CMS experiment","single-lepton search","95% confidence level limits","LHC Run 2"],"falsifier":"If, in a data sideband with at least one bottom-tagged jet defined by looser selection (for example, relaxing the top-tagging requirement or lowering the $H_T$ threshold), the reconstructed $m_{\\mathrm{rec}}^{Z'}$ distribution and its ratio to the zero-bottom-tag control region disagree with simulation by more than the floating normalization can absorb, the background transfer underpinning the limits would be falsified and the reported cross-section bounds would shift.","tokens_in":38670,"feed_emoji":"⚛️","tokens_out":10154,"duration_ms":82979,"temperature":0.7,"pith_summary":"The paper searches for a $Z'$ boson that couples only to top quarks, produced together with a top-antitop pair and decaying into another top-antitop pair, giving a four-top signature with one lepton. Using 138 fb$^{-1}$ of 13 TeV proton-proton collisions recorded by CMS, it looks for a resonance in the invariant mass of the two highest transverse-momentum boosted top-quark jets. The data agree with the standard model background, so no signal is claimed. The result is a set of 95% confidence-level upper limits on the production cross section times branching fraction: for a width of 4% of the mass, the limits range from 170 fb at 0.5 TeV down to 3 fb at 3 TeV, and masses below 560, 850, and 1130 GeV are excluded for widths of 10, 20, and 50%, respectively. These are the most stringent constraints so far on such a top-philic $Z'$ boson.","feed_headline":"CMS finds no top-only Z-prime boson; limits reach 3 fb","feed_subtitle":"For the narrowest width, 95% CL limits run 170-3 fb across 0.5-3 TeV; masses below 560-1130 GeV are excluded.","key_machinery":"The discriminant is the invariant mass $m_{\\mathrm{rec}}^{Z'}$ of a $Z'$ candidate built from the two highest-$p_T$ large-radius jets that pass a machine-learning top tagger called ParticleNet, which identifies hadronically decaying boosted top quarks from their jet constituents and substructure. The event sample requires one isolated electron or muon, at least six jets, $H_T>700$ GeV, $p_T^{\\mathrm{miss}}>60$ GeV, and at least one bottom-tagged jet. The background is modeled with simulated $m_{\\mathrm{rec}}^{Z'}$ templates rescaled by a single floating normalization factor determined in a simultaneous maximum-likelihood fit of the signal region and an orthogonal control region with zero bottom-tagged jets; the control region also validates the background shape in data. The $Z'$ width is parameterized through $\\Gamma_{Z'}/m_{Z'} \\simeq c_t^2/(8\\pi)$, so the four width assumptions correspond to different top-quark couplings.","core_discovery":"The central claim is that no excess of events beyond the standard model appears in the single-lepton, four-top-jet final state, and that the resulting upper limits improve on all previous constraints. For $\\Gamma_{Z'}/m_{Z'}=4\\%$, the observed 95% CL upper limits on $\\sigma(\\mathrm{pp}\\to\\mathrm{t\\bar{t}}Z')\\,\\mathcal{B}(Z'\\to\\mathrm{t\\bar{t}})$ fall from 170 fb at $m_{Z'}=0.5\\,\\mathrm{TeV}$ to 3 fb at 3 TeV; for widths of 10, 20, and 50%, the corresponding limits range over 160$-$6, 160$-$9, and 110$-$17 fb. Interpreting these against the leading-order production prediction, the analysis excludes $Z'$ masses below 560, 850, and 1130 GeV for the 10, 20, and 50% width cases. The paper states these are the most stringent limits to date on a $Z'$ boson that couples exclusively to top quarks. The claim rests on a shape analysis of the reconstructed mass spectrum in a signal region requiring at least one bottom-tagged jet, with the background normalization constrained by a zero-bottom-tag control region.","pith_inferences":["The reported cross-section limits can be recast as upper limits on the $Z'$–top coupling $c_t$ using $\\Gamma_{Z'}/m_{Z'} \\simeq c_t^2/(8\\pi)$; the paper does not quote such coupling bounds, but they follow directly from the width assumptions.","Re-running the same analysis on Run-3 data at the higher collision energy would extend the mass reach, because the selection efficiency for boosted top quarks rises with the $Z'$ mass and the higher collision energy increases the production rate; the current search already provides its best limits at the highest masses.","The zero-bottom-tag control region could be supplemented by a data-driven validation region containing bottom-tagged jets but with relaxed top-tagging requirements, which would test the transfer assumption in data rather than only in simulation."],"forward_implications":["A top-philic $Z'$ boson with $\\Gamma_{Z'}/m_{Z'}=10\\%$ is excluded below 560 GeV, and with 20% (50%) width below 850 (1130) GeV, assuming 100% branching fraction to $\\mathrm{t\\bar{t}}$.","For the narrow 4% width, production cross sections above 170 fb at 0.5 TeV and above 3 fb at 3 TeV are ruled out at 95% CL, constraining the coupling $c_t$ through the width relation.","The observed data being consistent with the standard model means the four-top final state shows no sign of the resonantly produced heavy state, complementing the observed standard-model four-top production at 13 TeV.","The combined electron and muon channels and the three data-taking years provide the search's statistical reach; the limits supersede the earlier ATLAS single-lepton limits, which ranged from 21 to 119 fb."],"supporting_citations":[{"why":"Provides the simplified top-philic $Z'$ model, its Lagrangian, the production mechanism, and the width-to-coupling relation used for signal samples.","marker":"[44]"},{"why":"ParticleNet, the machine-learning tagger used to identify the boosted hadronic top-quark jets that build the $Z'$ candidates.","marker":"[50]"},{"why":"The previous ATLAS search in the same single-lepton final state, whose limits this analysis compares against and claims to supersede.","marker":"[49]"},{"why":"A previous CMS four-top-quark measurement whose event-selection criteria the analysis extends with tighter jet and $H_T$ requirements.","marker":"[48]"},{"why":"POWHEG next-to-leading-order simulation of the dominant $\\mathrm{t\\bar{t}}$ background.","marker":"[68–71]"},{"why":"Next-to-next-to-leading-order QCD calculations, including soft-gluon terms, used to normalize the $\\mathrm{t\\bar{t}}$ background cross section.","marker":"[72–78]"},{"why":"DeepJet, the bottom-quark jet tagging algorithm that defines the signal-region requirement of at least one b-tagged jet and the zero-b-tag control region.","marker":"[88]"}],"fun_headline_variants":["CMS sets tightest limits on top-philic Z' boson","No top-only Z' in CMS data: limits down to 3 fb","CMS: No top-philic Z' boson, strongest limits","Top-philic Z' boson absent in CMS run-2 data","CMS data show no top-philic Z' boson"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the zero-bottom-tag control region reproduces the reconstructed-mass shape and background composition of the one-bottom-tag signal region, a transfer checked only in simulation.","fun_headline_variants_meta":{"raw":{"variants":["CMS sets tightest limits on top-philic Z' boson","No top-only Z' in CMS data: limits down to 3 fb","CMS: No top-philic Z' boson, strongest limits","Top-philic Z' boson absent in CMS run-2 data","CMS data show no top-philic Z' boson"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000891,"raw_usage":{"total_tokens":3947,"prompt_tokens":1150,"completion_tokens":2797,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":766,"completion_tokens_details":{"reasoning_tokens":2705}},"tokens_in":766,"tokens_out":2797,"duration_ms":17202,"temperature":1.0,"reasoning_tokens":2705,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:09:30.272482+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If, in a data sideband with at least one bottom-tagged jet defined by looser selection (for example, relaxing the top-tagging requirement or lowering the $H_T$ threshold), the reconstructed $m_{\\mathrm{rec}}^{Z'}$ distribution and its ratio to the zero-bottom-tag control region disagree with simulation by more than the floating normalization can absorb, the background transfer underpinning the limits would be falsified and the reported cross-section bounds would shift.","supporting_citations":[],"review_version":2}