{"id":"5ed7b655-5e4e-4eb3-82de-44eebc632b7f","arxiv_id":"1909.02588","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A lepton-specific two-Higgs-doublet model can accommodate the LHC's 28 GeV dimuon excess through a light CP-odd Higgs boson decaying to muons, though the reported significance is under 2 sigma.","lead":"This paper studies a two-Higgs-doublet model in which one Higgs doublet talks only to quarks and the other only to leptons. It finds that a light 'pseudoscalar' Higgs boson around 28 GeV in this model could explain a small excess of muon pairs seen at the LHC.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The abstract's 1.5σ and 2σ dimuon-excess significances are not derivable from the manuscript: Section 4 shows only MC signal/background curves, with no data, luminosity, systematic uncertainties, or statistical method.","rationale":"The reader's weakest_assumption correctly identifies the most load-bearing weakness: the quantitative excess claim is not backed by any reproducible statistical analysis. The paper's central novelty is not merely that a light CP-odd Higgs can exist in an LS-2HDM (which the mass-spectrum scan plausibly demonstrates), but that it can explain the CMS dimuon excess at a quoted significance. Without the data comparison, luminosity, systematics, and statistical method, the 1.5σ/2σ numbers are unsupported. The reader's verdict of CONDITIONAL is appropriate: the issue is fixable by supplying the omitted analysis, but as written the paper does not establish the headline claim. I also note a secondary internal inconsistency: the scalar potential in Eq. (2) and the mass formulas in Eq. (5) omit lambda4, yet Table 2 lists nonzero lambda4 values for both benchmark points. This is a serious documentation flaw, but it does not necessarily invalidate the numerical scan if SARAH/SPheno included lambda4; the primary obstacle remains the missing significance derivation. Agreement with the reader is 'agree' because the same weakest assumption was identified, and the verdict should remain CONDITIONAL rather than move to ACCEPT or REJECT.","tokens_in":11475,"tokens_out":6277,"duration_ms":69476,"concrete_test":"Take the CMS dimuon search datasets from arXiv:1808.01890 (8 and 13 TeV with their published luminosities), reproduce the Section 4 selection (pT(mu1)>25 GeV, |eta(mu1)|<2.1, pT(mu2)>5 GeV, |eta(mu2)|<2.4, pT(b-jet)>20 GeV, pTmiss<40 GeV) in MadGraph for Point 1 (mA=27.9 GeV) and the listed SM backgrounds, normalize to the CMS luminosities, and compute a binned profile-likelihood ratio including systematic uncertainties on the background estimate. Compare the resulting local significance at m_mu mu ~ 28 GeV with the claimed 1.5σ (8 TeV) and 2σ (13 TeV); if the recomputed values differ by more than 0.5σ, the abstract's quantitative claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is quantitative: an LS-2HDM with mA ~ 28 GeV and tan beta ~ 12 produces a 1.5σ excess at 8 TeV and 2σ at 13 TeV in pp -> bbA -> bb mu mu. For this claim to hold, the MadGraph event generation and the applied cuts must reproduce the CMS analysis in [9], and a valid statistical comparison against the observed data must yield those significances. Section 4 states only 'we follow similar analyses represented in [9]' and lists a few kinematical cuts; it does not give the integrated luminosity, the background normalization, the number of observed events, systematic uncertainties, or the statistical procedure. Figure 5 shows magenta (SM background), blue (LS-THDM without light Higgses), and red (signal) curves, but no data points or uncertainty bands. Consequently, the numbers '1.5 sigma' and '2 sigma' in the abstract cannot be reproduced or checked from the text. If the background model or the data comparison is wrong, the claimed excess disappears; if the statistical treatment is absent, the claim is unfalsifiable as presented. This is not a minor presentational issue but a gap in the evidence for the paper's main quantitative conclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates the Higgs mass spectrum in the lepton-specific two-Higgs-doublet model (LS-2HDM), where one doublet couples only to quarks and the other only to leptons. Using a SPheno/SARAH scan, the authors apply constraints from the 125 GeV Higgs mass, LEP, B_s to mu+ mu- and b to s gamma, and find that a CP-odd Higgs boson as light as about 10 GeV can be consistent with these constraints. They present two benchmark points: Point 1 with m_A = 27.9 GeV and h_1 SM-like at 125.3 GeV, and Point 2 with m_h1 = 27.3 GeV and h_2 SM-like at 123.4 GeV. They then compute pp -> b bbar phi -> b bbar mu+ mu- event distributions with MadGraph, compare them with the SM background, and claim that the model can explain the CMS 28 GeV dimuon excess, quoting significances of about 1.5 sigma at 8 TeV and 2 sigma at 13 TeV for tan beta about 12.","tokens_in":11723,"tokens_out":13262,"duration_ms":119072,"significance":"If fully substantiated, the paper would provide a concrete and economic new-physics interpretation of the 28 GeV dimuon excess, and the mass-spectrum result that m_A can be as low as about 10 GeV while evading B-physics and Higgs constraints is independently interesting. The use of standard public tools (SPheno, SARAH, MadGraph) and the inclusion of the B_s to mu+ mu- and b to s gamma constraints are strengths. However, the headline quantitative claim is not verifiable from the manuscript as written, and several internal inconsistencies in the model definition and mass-spectrum summary must be resolved before the results can be accepted.","major_comments":[{"comment":"The 1.5 sigma and 2 sigma significances stated in the abstract are not derivable from the presented analysis. Section 4 says only that the authors 'follow similar analyses represented in [9]' and lists a few kinematic cuts; it gives no integrated luminosity, no data points, no background normalization, no systematic uncertainties, and no statistical procedure. Figure 5 shows only model and background curves without data or uncertainty bands. Without these elements, the significance numbers cannot be reproduced or checked, leaving the paper's main quantitative conclusion unsupported.","section":"Section 4 / Abstract"},{"comment":"The abstract states that when h1 is the SM-like Higgs boson, 'the heavier CP-even (h2) and charged Higgs bosons (H+) become above 600 GeV.' This contradicts Section 3, which finds that the light-m_A region yields m_h2 less than about 350 GeV, and Point 1 in Table 2 has m_h2 = 178.6 GeV and m_H+ = 173.8 GeV. The mass-spectrum summary in the abstract is inconsistent with the results in the body and must be corrected.","section":"Section 3 / Abstract / Table 2"},{"comment":"The scalar potential written in Eq. (2) omits the lambda4 (Phi1^dagger Phi2)(Phi2^dagger Phi1) term, yet Table 2 lists nonzero lambda4 values for both benchmark points, and the mass formulas in Eq. (5) do not include lambda4. If the numerical scan uses a model with lambda4, the text is incomplete and the mass formulas are missing a term; if lambda4 is supposed to be zero, the table entries are incorrect. This ambiguity directly affects the computed mass spectrum, decay branching ratios, and all derived constraints.","section":"Section 2, Eq. (2) / Table 2"},{"comment":"The benchmark points are under-specified: tan beta, quoted as about 12 in the abstract, is not listed in Table 2, nor is the CP-even mixing angle alpha. In addition, the centre-of-mass energy at which the cross-sections in Table 2 are evaluated is not identified; Section 3 refers to 14 TeV LHC energies, while the abstract cites 8 and 13 TeV. Without tan beta, alpha, and the collision energy, the benchmark points cannot be reproduced or independently checked.","section":"Table 2 / Reproducibility"},{"comment":"Use of the word 'predict' is too strong for the role the model actually plays. In Point 1, m_A = 27.9 GeV is an input parameter chosen to match the 28 GeV excess, so the model accommodates the excess rather than independently predicting it. The paper should state transparently that the benchmark is tuned to the observed mass, and that the scan is not a fit to the dimuon data.","section":"Abstract"}],"minor_comments":[{"comment":"The introduction contains two broken cross-references: 'we consider a case in which the CP-odd Higgs boson of mass about 28 GeV can yield an excess ... while we also consider the possibility for a CP-even Higgs boson leading to excess in Section ??.' The placeholder 'Section ??' should be replaced with the correct section number.","section":"Section 1"},{"comment":"The text references the wrong panel in Figure 5: it says the A-boson signal 'is shown in the right panel' and 'The left panel of Figure 5 shows that Point 2', but the caption labels Point 1 as left and Point 2 as right. The panel references should be corrected or the caption changed.","section":"Section 4 / Figure 5"},{"comment":"The caption states 'The hyphens in sigma(A to mu+ mu-) of Point 1 and sigma(h1 to mu+ mu-) of Point 2 mean the branching ratios ... are lower than about 10^-5.' This is reversed: for Point 1 the hyphen appears in sigma(h1 to mu+ mu-), and for Point 2 it appears in sigma(A to mu+ mu-).","section":"Table 2 caption"},{"comment":"The pseudorapidity cut on the b-jet is written as '|eta_mu1| < 2.4'; this should be the pseudorapidity of the b-jet, not of a muon. Also, the symbols '&' and 'less-than-or-similar' are used where 'greater than' and 'less than' are meant, which should be clarified.","section":"Section 4, cut list"},{"comment":"The text refers to 'the current LHC experiments with 14 TeV COM energy', but Run 2 of the LHC operates at 13 TeV; 14 TeV is the design energy. This should be corrected to avoid ambiguity about the cross-section predictions.","section":"Section 3"},{"comment":"The constraint '123 <= m_hi <= 127 GeV' with i = 1,2 cannot hold for both CP-even states simultaneously. The text later clarifies that only the SM-like state is constrained, so the notation should be modified to indicate that the bound applies to the SM-like CP-even Higgs boson only.","section":"Eq. (6)"},{"comment":"The 1% accuracy of the approximation in Eq. (7) is attributed to Ref. [34], which is an unrelated manuscript 'in preparation' about gluino searches. A published source or a direct calculation should be cited instead.","section":"Footnote 1 / Ref. [34]"}],"recommendation":"major_revision","confidential_remarks":"The manuscript reads like an early draft: it has broken cross-references, swapped panel descriptions, an internal contradiction between the abstract and the body regarding the h2 mass, and an inconsistency between the potential in Eq. (2) and the parameters in Table 2. More importantly, the central quantitative claim—the 1.5 and 2 sigma significances—is not backed by any statistical analysis in the text. I would advise the editor that acceptance cannot be considered until the authors either provide a complete data/MC comparison with a stated statistical method or remove the significance claims and restrict the paper to the mass-spectrum and cross-section study. If the statistical analysis is added, the paper may be publishable; otherwise the quantitative conclusion is unfalsifiable as presented."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is two papers in one. The mass-spectrum analysis is workmanlike and likely right. Using SPheno/SARAH with standard constraints, the authors credibly show that a lepton-specific 2HDM can have a CP-odd Higgs as light as ~28 GeV, with tanβ around 12, while satisfying Bs→μμ, b→sγ, and the 125 GeV Higgs measurements. The bound that the lighter CP-even Higgs cannot go below about 55 GeV when mA~28 GeV is a concrete, useful result.\n\nThe second thing: the abstract's headline numbers are not in the paper. Section 4 describes MadGraph event generation and lists kinematical cuts, but gives no integrated luminosity, no observed event counts, no background normalization, no systematic uncertainties, and no statistical method. Figure 5 shows magenta, blue, and red curves with no data points and no uncertainty bands. So the claimed 1.5σ at 8 TeV and 2σ at 13 TeV are not derivable from the manuscript. That is not a cosmetic flaw; it is the central claim of the paper.\n\nWhat is genuinely new is the application to the lepton-specific 2HDM and the specific benchmark: mA=27.9 GeV, tanβ~12, with the lightest CP-even Higgs at 125 GeV. Those numbers are consistent with the mass-spectrum scan and with the B-meson constraints, as far as I can tell. But note the circularity: the resonance location is put in by hand when mA is set to 27.9 GeV, so the paper accommodates the bump rather than predicting it; the abstract's “predict” overstates that.\n\nThere is also a technical inconsistency worth flagging: Table 2 lists λ4=-0.82 for Point 1, but the scalar potential in Eq. (2) explicitly sets λ4=0 along with λ6–λ10. Either the model definition or the benchmark point is wrong, and that needs to be fixed before the numbers can be trusted. Minor issues: a dangling “Section ??” placeholder for the CP-even case, and the approximation in Eq. (7) is validated by a reference that is “in preparation” rather than published.\n\nSo my take: the spectrum part deserves a serious referee, but the excess claim as written does not. If this were submitted to a journal, I would not desk-reject it; I would send it to review with a clear request that the authors either provide a full statistical comparison with the CMS data or remove the σ numbers from the abstract and conclusions. In the latter case, the paper becomes a modest but acceptable parameter-space study.\n\nFor a reading group, it is a useful case study in how easily “we follow a similar analysis” can hide an absent statistical framework. I would not cite the excess claim, though I might cite the mass-spectrum bound if I worked on light 2HDM phenomenology.","headline":"Solid 2HDM mass-spectrum scan undercuts its quantitative claim: the abstract quotes 1.5σ and 2σ excess significances that the text never derives.","tokens_in":12322,"tokens_out":2345,"would_cite":false,"duration_ms":27101,"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 lepton-specific two-Higgs-doublet model with a 28 GeV CP-odd Higgs boson can account for the observed dimuon excess while evading rare B-decay constraints.","keywords":["two Higgs doublet model","lepton-specific 2HDM","28 GeV dimuon excess","CP-odd Higgs boson","tan beta enhancement","rare B meson decays","LHC Higgs searches","light Higgs spectrum"],"falsifier":"Plot the measured $b\\bar b\\mu^+\\mu^-$ invariant-mass spectrum around 28 GeV from the 13 TeV dataset with the same cuts; if the data show no excess above the simulated background at the level the model predicts, the claim is falsified.","tokens_in":11220,"feed_emoji":"⚛️","tokens_out":12384,"duration_ms":122668,"temperature":0.7,"pith_summary":"This paper tries to show that a lepton-specific two-Higgs-doublet model—one doublet coupling only to quarks, the other only to leptons—can explain the observed excess of dimuon events at an invariant mass near 28 GeV. The explanation runs through the CP-odd Higgs boson $A$: produced in association with bottom quarks and decaying to $\\mu^+\\mu^-$, it adds a small resonance to the dimuon spectrum. The authors find a consistent spectrum with $m_A\\sim 28$ GeV and $\\tan\\beta\\sim12$, where the rare-decay constraints $\\mathrm{Br}(B_s\\to\\mu^+\\mu^-)$ and $\\mathrm{Br}(b\\to s\\gamma)$ stay within their measured ranges, and they quote the resulting excess as about $1.5\\sigma$ at 8 TeV and $2\\sigma$ at 13 TeV. If true, this is the simplest way to turn a modest collider bump into a specific prediction about the Higgs sector.","feed_headline":"Lepton-specific two-Higgs model explains the 28 GeV dimuon excess","feed_subtitle":"A 28 GeV pseudoscalar with tan β ≈ 12 passes rare-decay limits and leaves a small bump at both 8 and 13 TeV.","key_machinery":"The mechanism that carries the argument is the lepton-specific Yukawa structure imposed by the $Z_2$ symmetry: quarks receive mass only from $\\Phi_2$ and leptons only from $\\Phi_1$, so the CP-odd Higgs couplings are $\\cot\\beta$ to quarks and $\\tan\\beta$ to leptons. This ratio suppresses the dangerous contributions to $B_s\\to\\mu^+\\mu^-$ (which grows like $(\\tan\\beta)^6/m_A^4$) while keeping $A\\to\\mu^+\\mu^-$ sizeable. The analysis also relies on the tree-level mass relation $m_A^2=m_3^2/(\\sin\\beta\\cos\\beta)-\\lambda_5 v^2$, which ties $m_A$ to the other Higgs masses, and on a parameter scan plus parton-level event generation using the selection cuts of the 28 GeV dimuon search. The signal rate is estimated through $\\sigma(pp\\to A\\to\\mu^+\\mu^-)\\approx\\sigma(pp\\to A)\\,\\mathrm{Br}(A\\to\\mu^+\\mu^-)$, with the full process $pp\\to b\\bar b A\\to b\\bar b\\mu^+\\mu^-$ checked against the dominant Drell-Yan and top backgrounds.","core_discovery":"The central discovery claim is that a lepton-specific 2HDM (LS-2HDM) can accommodate a CP-odd Higgs boson of about 28 GeV that is produced through $pp\\to b\\bar b A$ and decays to muons, producing a small bump in the dimuon invariant-mass distribution that matches the reported excess at $m_{\\mu\\mu}\\sim28$ GeV. Because the quarks couple to $A$ with strength $\\cot\\beta$ while leptons couple with strength $\\tan\\beta$, the state has suppressed production but enhanced leptonic decays; at $\\tan\\beta\\sim12$ the balance is enough to give a visible signal without violating the constraints from $B_s\\to\\mu^+\\mu^-$, $b\\to s\\gamma$, and the 125 GeV Higgs measurements. The paper also demonstrates that a consistent spectrum can contain a CP-odd state as light as about 10 GeV, while the lightest CP-even Higgs boson cannot be lighter than about 55 GeV in the region $m_A\\sim28$ GeV. Two benchmark points are given, one with the CP-odd state at 28 GeV and one with a light CP-even state at 28 GeV; both are shown to produce similar excesses in the associated-production channel.","pith_inferences":["A direct comparison of the model curves to the published data points around $m_{\\mu\\mu}=28$ GeV, including systematic uncertainties, would convert the quoted $1.5$ and $2\\sigma$ values into a testable significance; the paper itself does not display the data.","The same $A\\to\\mu^+\\mu^-$ resonance could be searched for in $\\tau^+\\tau^-$ and $e^+e^-$ final states; measuring the ratio of leptonic branching fractions would distinguish the CP-odd Higgs explanation from a $Z'$, dark photon, or another resonance at the same mass.","Because $A$ couples to quarks through $\\cot\\beta$, any measurement that shifts $\\tan\\beta$ away from about 12—for example from $b\\to s\\gamma$—changes the predicted significance, so the model is tightly predictive in other observables."],"forward_implications":["If the 28 GeV excess is a real CP-odd Higgs, the same state should appear in the $b\\bar b\\tau^+\\tau^-$ channel, and the existing $\\tau\\tau$ limits already restrict which $\\tan\\beta$ values remain viable.","The benchmark point with $m_A\\simeq28$ GeV and $m_{h_1}\\simeq125$ GeV predicts the heavier CP-even and charged Higgs bosons around 180 GeV and 174 GeV, well within reach of direct LHC searches.","At a 100 TeV collider the $A$ production cross section reaches roughly 5000 pb and $A\\to\\mu^+\\mu^-$ roughly 10 pb, so the scenario is either confirmed or excluded early in the future program.","The parameter region $m_A\\sim28$ GeV with $h_1$ SM-like forces $m_{h_2}\\lesssim350$ GeV, making the model's extra scalars testable rather than decoupled."],"supporting_citations":[{"why":"the LHC dimuon search that reports the 28 GeV excess and supplies the analysis cuts the paper follows.","marker":"[9]"},{"why":"shows that light CP-odd Higgs states can survive current experimental results, a starting point for the spectrum scan.","marker":"[14]"},{"why":"provides the $(\\tan\\beta)^6/m_A^4$ scaling of the $B_s\\to\\mu^+\\mu^-$ contribution that constrains a light $A$.","marker":"[22]"},{"why":"supplies the Higgs-sector mass relations and scalar potential parametrisation used to scan the spectrum.","marker":"[24]"},{"why":"parton-level event generation used to simulate the signal and background processes.","marker":"[31]"},{"why":"standard-model predictions for the rare B decays that set the target values for $B_s\\to\\mu^+\\mu^-$ and $b\\to s\\gamma$.","marker":"[16]"},{"why":"experimental measurements of $B_s\\to\\mu^+\\mu^-$ that provide the 2 sigma window in the constraints.","marker":"[17]"},{"why":"the charged-Higgs mass bound from LEP applied to $m_{H^\\pm}$.","marker":"[26]"}],"fun_headline_variants":["Lepton-specific 2HDM yields 28 GeV dimuon bump","A 28 GeV pseudoscalar produces dimuon excess at 8 and 13 TeV","LS-2HDM predicts 28 GeV dimuon bump with tan beta ~ 12","Lepton-specific Higgs model fits 28 GeV dimuon signal"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central excess claim rests on the assumption that the simulated collision events and the selection cuts match the real LHC dimuon search closely enough to produce 1.5 and 2 sigma significances, but the paper does not show the data points or the background uncertainties.","fun_headline_variants_meta":{"raw":{"variants":["Lepton-specific 2HDM yields 28 GeV dimuon bump","A 28 GeV pseudoscalar produces dimuon excess at 8 and 13 TeV","LS-2HDM predicts 28 GeV dimuon bump with tan beta ~ 12","Lepton-specific Higgs model fits 28 GeV dimuon signal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000819,"raw_usage":{"total_tokens":3654,"prompt_tokens":1085,"completion_tokens":2569,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":2478}},"tokens_in":701,"tokens_out":2569,"duration_ms":19667,"temperature":1.0,"reasoning_tokens":2478,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:46:50.335744+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Plot the measured $b\\bar b\\mu^+\\mu^-$ invariant-mass spectrum around 28 GeV from the 13 TeV dataset with the same cuts; if the data show no excess above the simulated background at the level the model predicts, the claim is falsified.","supporting_citations":[{"cited_title":"Search for resonances in the mass spectrum of muon pairs produced in association with b quark jets in proton-proton collisions at $\\sqrt{s} =$ 8 and 13 TeV","cited_arxiv_id":"1808.01890","evidence_quote":"the LHC dimuon search that reports the 28 GeV excess and supplies the analysis cuts the paper follows."},{"cited_title":"Reconstructing heavy Higgs boson masses in Type X Two-Higgs Doublet Model with a light pseudoscalar","cited_arxiv_id":"1807.05379","evidence_quote":"shows that light CP-odd Higgs states can survive current experimental results, a starting point for the spectrum scan."}],"review_version":1}