{"id":"62f4e461-df6d-484b-9466-12e1be4e2ccc","arxiv_id":"2412.08212","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The mono-Higgs plus missing energy signature sets 95% CL bounds on the ALP-Higgs coupling CaH/Lambda^2 and the sterile neutrino-Higgs coupling lambda3/M*, but these bounds are weaker than existing limits from invisible Higgs decays.","lead":"This paper reuses an ATLAS mono-Higgs plus missing-energy search to constrain two new physics models: an axion-like particle (ALP) coupled to the Higgs, and a sterile neutrino coupled to the Higgs. It reports 95% confidence level bounds on the two couplings at the LHC and projected sensitivities at the HL-LHC, but the new bounds are weaker than existing constraints from invisible Higgs decays.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unvalidated background rescaling in Eq. (11) underpins all HL-LHC projections; per-bin shape and acceptance changes with center-of-mass energy are not tested.","rationale":"The reader's weakest_assumption is exactly the rescaling in Eq. (11), and I independently identify it as the most load-bearing concern. It affects all HL-LHC projections, which are a substantial part of the paper's quantitative claims, and the paper provides no validation that the per-bin background shapes and cut efficiencies are invariant under the 13-to-14 TeV change. The concrete simulation test would settle whether the assumption holds; if it does, the HL-LHC projections stand, and if it does not, the curves would shift. I considered the invisible-Higgs decay comparison raised in the paper's conclusion: the paper acknowledges this limitation, and the quoted mono-Higgs bound at ma=1 GeV (about 3.8e-6 GeV^-2) is only a factor of a few above the invisible-decay bound (8.1e-7 GeV^-2), so it is a real but secondary caveat about framing rather than a threat to the numerical validity of the 13 TeV limits. The verdict remains CONDITIONAL because the paper should either validate the rescaling or recompute the projections; no change to the reader's verdict is needed.","tokens_in":19265,"tokens_out":8874,"duration_ms":94271,"concrete_test":"Simulate the dominant backgrounds (t-tbar, W+HF, Z+HF) at 14 TeV with the same MadGraph5+Pythia8+Delphes setup and ATLAS selection cuts used for the signal, at 139/fb normalization, and compare the per-MET-bin event yields to the N_b^rescaled entries in Table II. If the per-bin yields deviate by more than ~20% (or by more than the simulation's statistical uncertainty), the rescaling assumption in Eq. (11) is invalid and the HL-LHC curves in Figs. 7-10 must be recomputed. A simpler partial validation: first reproduce the 13 TeV background counts in Table II from ATLAS with the same simulation chain before extrapolating to 14 TeV.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's HL-LHC sensitivity projections (Figs. 7-10, dotted and dot-dashed curves) rest entirely on Eq. (11), which rescales each 13 TeV background event count by the ratio of total 13/14 TeV production cross sections for t-tbar, W+HF, and Z+HF. This assumes that the MET distribution shape, the acceptance of the resolved-region selection (lepton veto, Delta-phi>20deg, PTh thresholds, mTb cuts, Njets, mbb window), and the per-bin migration are unchanged between 13 and 14 TeV. No 14 TeV background simulation is performed to validate this assumption; only the total cross-section ratios (about 1.1-1.2) are used. The signal is fully simulated at 14 TeV, so the projection mixes a full simulation of the signal with a scalar-rescaled background. Because the ATLAS analysis bins exclusively in MET, and the quoted HL-LHC sensitivity changes ordering between the 200-350 and 350-500 GeV bins depending on systematic uncertainties, the result is sensitive to exactly the quantity that is being linearly rescaled. If the MET spectrum hardens or migrates at 14 TeV, the per-bin background yields could shift by more than the quoted 20% systematic, directly altering the projected exclusion curves for CaH/Lambda^2 and lambda3/M*. This does not affect the 13 TeV limits, which use ATLAS data directly, but it is an unvalidated assumption at the core of the paper's forward-looking projections.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper recasts the ATLAS mono-Higgs plus missing transverse energy search (Ref. [8]) to constrain two effective operators: the dimension-six ALP-Higgs interaction CaH/Λ^2 (∂a)^2 H†H, probed via pp→haa with h→bb, and the dimension-five sterile-neutrino-Higgs interaction λ3/M* H†H NN, probed via pp→hNN with h→bb. For the 13 TeV, 139 fb^-1 LHC dataset, the authors follow the ATLAS resolved-region selection and apply model-independent upper limits on the visible cross section to derive 95% C.L. excluded regions on the two couplings as functions of the invisible-particle mass (1–60 GeV) in three MET bins. For the HL-LHC at 14 TeV and 3000 fb^-1, they use Eq. (11) to rescale the ATLAS background event counts by inclusive cross-section ratios, and Eq. (13) to define expected sensitivities with and without a 20% systematic uncertainty. The main results are the exclusion and projection curves in Figs. 7–10, with the strongest 13 TeV constraint reported in the 200<MET≤350 GeV bin.","tokens_in":19618,"tokens_out":9742,"duration_ms":99669,"significance":"If the 13 TeV recasting is correct, the paper provides new, but not record-setting, constraints on two specific effective operators from a public ATLAS search. The cut-flow tables and explicit use of ATLAS selection criteria make the 13 TeV part reasonably reproducible, which is a strength. However, two factors limit the significance: the HL-LHC projections rely on an unvalidated background-rescaling assumption that is central to a large fraction of the presented results, and the authors themselves acknowledge in Sec. V that the simpler observable h→invisible already gives stronger bounds on both couplings (CaH/Λ^2<8.1e-7 GeV^-2 and λ3/M*<3.6e-5 GeV^-1 at 1 GeV). The mono-Higgs channel is therefore complementary rather than competitive, and the paper's claim that it is a 'robust probe' should be tempered. The 13 TeV limits alone are a modest but useful recasting contribution; the HL-LHC projections, as currently derived, are not a reliable basis for quantitative conclusions.","major_comments":[{"comment":"The HL-LHC sensitivity projections are built on Eq. (11), which rescales the ATLAS 13 TeV background event counts in each MET bin by the ratio of inclusive 14 TeV to 13 TeV production cross sections for t-tbar, W+HF, and Z+HF. This assumes that the MET shape, the acceptance of the resolved-region selection (lepton vetoes, Delta-phi>20 deg, Higgs pT thresholds, mTb cuts, Njets, and the mbb window), and the bin-to-bin migration are identical at 13 and 14 TeV. No 14 TeV background simulation is presented to test this assumption, and the inclusive cross-section ratios (about 1.1-1.2) do not by themselves validate per-bin yields. Because the signal is fully simulated at 14 TeV while the backgrounds are only rescaled, the projection is asymmetric; any change in the MET spectrum with sqrt(s) would directly shift the per-bin background counts and hence the excluded values of CaH/Lambda^2 and lambda3/M*. Since all HL-LHC curves in Figs. 7-10 and the corresponding conclusions depend on this rescaling, the assumption is load-bearing and should be validated with a 14 TeV background simulation or at least a generator-level shape comparison.","section":"III.A.2, Eq. (11), Table II, Figs. 7-10"},{"comment":"The authors state that the existing bound B(H->invisible)<0.16 yields CaH/Lambda^2<8.1e-7 GeV^-2 (ma=1 GeV) and lambda3/M*<3.6e-5 GeV^-1 (mN=1 GeV), which are stronger than the mono-Higgs limits obtained in this work (3.8e-6 GeV^-2 and 1.5e-4 GeV^-1, respectively). Given this acknowledged hierarchy, the abstract and conclusions overstate the significance by calling the mono-Higgs signature 'a robust probe' and saying the results 'establish bounds' without noting that the same operators are already more strongly constrained by a simpler observable. The paper should either present the mono-Higgs results as complementary probes of the MET distribution or explicitly frame the constraints as weaker than existing bounds.","section":"V (Conclusions)"}],"minor_comments":[{"comment":"The heading 'Axion-like Particke' should read 'Axion-like Particle'.","section":"II.A (heading)"},{"comment":"The sentence 'The signal cross section sigma(pp -> haa, h->bb) is shown in FIG. 9' refers to Fig. 9 in the appendix; the cross section is shown in Fig. 2.","section":"III.A"},{"comment":"The text uses 'M_ET' for missing transverse energy; it should use MET or E_T^miss consistently.","section":"Abstract"},{"comment":"The caption says 'but in the resolved region with MET > 500 GeV'; this table is for the merged region and should be labeled accordingly.","section":"Table IX caption"},{"comment":"There are several typos: 'invisble', 'sensitivites', 'across across', 'sing' (for 'using'), and 'Cah' (for 'CaH') in the conclusions.","section":"Throughout"},{"comment":"The text refers to the 'coupling lambda3/M2*' but the parameter is lambda3/M*; the superscript appears to be a typo.","section":"II.B.1"},{"comment":"For reproducibility, the paper should list the ATLAS model-independent upper-limit values used in the 13 TeV exclusions, for example by citing the specific tables of Ref. [8], since the conversion from these limits to coupling exclusions is otherwise opaque.","section":"IV.A"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a straightforward recast of a public ATLAS search. The 13 TeV limits appear defensible, although they are not competitive with existing H->invisible constraints on the same operators. The HL-LHC projections are the main weakness: Eq. (11) is an unvalidated assumption that underpins all of the forward-looking curves. I would ask the authors to either replace the rescaling with an actual 14 TeV background simulation or to explicitly label the projections as illustrative and remove the quantitative claims that depend on the assumed per-bin scaling. With that change and a more measured set of conclusions, the paper could be acceptable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the useful core: the paper recasts ATLAS's 139 fb^-1 mono-Higgs search to set 95% CL limits on the dimension-six ALP-Higgs operator (da)^2 phi^dagger phi and the dimension-five sterile neutrino-Higgs operator H^dagger H N N. As far as I know, no one has applied this particular search to these exact operators before, so the 13 TeV exclusion curves in Figs. 7 and 8 are a new, small addition to the literature. The recasting procedure is standard: generate signal with MadGraph/Delphes, apply the ATLAS resolved-region cuts, and read off the model-independent upper limits. I don't see a circularity problem or a gross error in the scaling of the signal cross sections.\n\nThe paper also deserves credit for being honest about where the bounds sit. In Section V it explicitly notes that the same operators induce H -> invisible, and that the CMS bound Br(H -> inv) < 0.16 yields CaH/Lambda^2 < 8.1e-7 GeV^-2 and lambda3/M* < 3.6e-5 GeV^-1 at 1 GeV, which are 'somewhat better' than the mono-Higgs limits. That admission is more candid than many recasting papers.\n\nThe weak spot is the HL-LHC projection. Eq. (11) rescales each 13 TeV background bin by the ratio of total production cross-sections at 14 and 13 TeV for t-tbar, W+HF, Z+HF. That assumes the MET shape, acceptance, and cut efficiencies are unchanged between the two energies. No 14 TeV background simulation is done to check this. Since the analysis bins purely in MET, and the sensitivity ordering between the 200-350 and 350-500 GeV bins changes with the systematic assumption, the projected limits in Figs. 7-10 could shift if the MET spectrum hardens at 14 TeV. This is not a fatal flaw—the paper labels it a 'simple rescaling' and the projections are expressly estimates—but it is the least solid part of the paper.\n\nMinor issues: a few typos ('Particke', 'sensitivites', 'invisble'), and the merged-region results are arguably underpowered, but they are consigned to appendices.\n\nOverall, the central 13 TeV result is defensible and reproducible. The HL-LHC numbers should be read as indicative, not definitive. I would send it to a serious referee; a good referee will ask for a more careful treatment or caveat on the background rescaling, and for a more prominent comparison with invisible decay limits.\n\nFor a reading group, it's a decent example of a clean recast, but not a must-read. I'd cite it if I were working on these couplings.","headline":"Solid 13 TeV recast of ATLAS mono-Higgs data to ALP and sterile neutrino couplings; HL-LHC projections rest on an unvalidated background rescaling, and the new limits are weaker than existing invisible Higgs bounds.","tokens_in":20083,"tokens_out":3898,"would_cite":true,"duration_ms":36852,"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 argues that the LHC's mono-Higgs plus large missing-energy search can place 95% confidence-level bounds on both the ALP-Higgs coupling $C_{aH}/\\Lambda^2$ and the sterile neutrino-Higgs coupling $\\lambda_3/M_*$ for particle…","keywords":["axion-like particle","sterile neutrino","mono-Higgs","missing transverse energy","effective field theory","LHC","HL-LHC","Higgs portal"],"falsifier":"Compute the $t\\bar t$, $W$+heavy-flavor, and $Z$+heavy-flavor backgrounds at 14 TeV with the full parton-level and detector simulation of the resolved-region selection, and compare the resulting per-bin event counts with the rescaled values in Table II; if the missing-energy distributions shift with center-of-mass energy, the projected coupling limits would move accordingly.","tokens_in":19064,"feed_emoji":"🔍","tokens_out":8624,"duration_ms":78694,"temperature":0.7,"pith_summary":"The paper argues that the mono-Higgs plus large missing transverse energy signature, where the Higgs decays to $b\\bar b$, can be used to constrain two beyond-Standard-Model couplings: the ALP-Higgs coupling $C_{aH}/\\Lambda^2$ from a dimension-six operator and the sterile neutrino-Higgs coupling $\\lambda_3/M_*$ from a dimension-five operator. Using 139 fb$^{-1}$ of 13 TeV LHC data on mono-Higgs events, it derives 95% confidence-level exclusion bounds for particle masses from 1 to 60 GeV, with the tightest constraints in the missing transverse energy range $200 < M_{ET} \\le 350$ GeV. If correct, the mono-Higgs search is a valid probe of both operators, and the quoted exclusion curves represent actual limits from the LHC data. The paper also projects what sensitivities the HL-LHC with 3000 fb$^{-1}$ at 14 TeV would reach.","feed_headline":"Mono-Higgs data sets limits on ALP and sterile-neutrino couplings","feed_subtitle":"A single missing-energy search constrains ALP-Higgs and sterile neutrino-Higgs couplings, strongest at 200-350 GeV.","key_machinery":"The central objects are two effective operators: the dimension-six ALP-Higgs interaction $(\\partial_\\mu a)(\\partial^\\mu a)\\phi^\\dagger\\phi$ with coefficient $C_{aH}/\\Lambda^2$, and the dimension-five sterile neutrino-Higgs interaction $\\phi^\\dagger\\phi N N$ with coefficient $\\lambda_3/M_*$, where the ALP $a$ and sterile neutrino $N$ are assumed to escape the detector as missing energy. The search topology is $pp\\to h a a$ or $pp\\to h N N$ with $h\\to b\\bar b$, selected by the resolved-region cuts of the 13 TeV mono-Higgs search: $M_{ET}>150$ GeV, lepton veto, $\\Delta\\phi(\\mathrm{jet}_{123},M_{ET})>20^\\circ$, at least two $b$-tagged jets, a Higgs transverse-momentum threshold that depends on $M_{ET}$, $b$-jet transverse mass cuts, a jet multiplicity restriction, and a $50 < m_{b\\bar b} < 280$ GeV invariant-mass window. The argument then uses the scaling of signal event counts as $(C_{aH}/\\Lambda^2)^2$ and $(\\lambda_3/M_*)^2$ to convert the measured cross-section limits, or the projected significance from the background-rescaling formula, into coupling exclusions.","core_discovery":"The central claim is that the existing mono-Higgs search data can be reinterpreted to constrain two hidden-sector models with a single event topology: $pp \\to h a a$ for an axion-like particle and $pp \\to h N N$ for a sterile neutrino, each followed by $h \\to b\\bar b$ while the invisible particles escape as missing transverse energy. The strongest 95% C.L. bound on both couplings appears in the $200 < M_{ET} \\le 350$ GeV bin of the resolved region, giving for the ALP coupling values around $3.8\\times10^{-6}$ GeV$^{-2}$ at $m_a = 1$ GeV rising to about $1\\times10^{-5}$ GeV$^{-2}$ at 60 GeV, and for the sterile neutrino coupling $1.5\\times10^{-4}$ GeV$^{-1}$ at $m_N = 1$ GeV rising to about $10^{-3}$ GeV$^{-1}$ at 60 GeV. At the HL-LHC, the projected sensitivities improve across all missing-energy bins, with the ALP model remaining more sensitive than the sterile neutrino model at equivalent masses, especially below about 10 GeV.","pith_inferences":["A natural extension the paper does not develop is to combine the mono-Higgs bin information with the invisible-Higgs branching-fraction bound in a global fit, which would likely sharpen the low-mass region of both exclusion plots.","Because the production cross section drops sharply once the invisible particle mass exceeds $m_H/2$, the bounds are limited to the sub-60 GeV range; a dedicated analysis of the off-shell Higgs region could extend the reach to higher masses.","If the ALP's dimension-five couplings are switched on even at small strength, the final state becomes $b\\bar b$ plus two photons, gluons, or fermion pairs rather than missing energy, so a search for $b\\bar b+\\gamma\\gamma$ would be a directly testable way to probe the same $C_{aH}$ coupling.","The background rescaling used for the 14 TeV projection could be validated before 3000 fb$^{-1}$ of data exist by comparing the rescaled 13 TeV predictions against observed 14 TeV event counts in early HL-LHC running, providing an early check on the projected limits."],"forward_implications":["The 13 TeV mono-Higgs data already exclude ALP-Higgs couplings down to about $3.8\\times10^{-6}$ GeV$^{-2}$ at $m_a=1$ GeV and sterile neutrino-Higgs couplings down to about $1.5\\times10^{-4}$ GeV$^{-1}$ at $m_N=1$ GeV in the $200 < M_{ET} \\le 350$ GeV bin.","At the HL-LHC the same search improves the reach in every missing-energy bin; without systematic uncertainty the best projected ALP bound is about $1.8\\times10^{-6}$ GeV$^{-2}$ in the $200$--$350$ GeV bin, while with a 20% systematic the $350$--$500$ GeV bin becomes the strongest.","The ALP model is probed more strongly than the sterile neutrino model at all equivalent masses, and the gap widens at low mass, so a null result in mono-Higgs would translate into a stronger exclusion for ALP-Higgs couplings than for sterile neutrino-Higgs couplings.","Because both operators also induce invisible Higgs decay, the existing bound $B(H\\to \\text{invisible})<0.16$ gives comparable limits at $m=1$ GeV, meaning the mono-Higgs search complements rather than replaces the invisible-width constraint."],"supporting_citations":[{"why":"Supplies the 13 TeV mono-Higgs search data, the resolved-region selection cuts, and the model-independent upper limits on visible cross sections that the exclusion bounds are derived from.","marker":"[8]"},{"why":"Provides the MadGraph5 aMC@NLO event generator used to compute the $pp\\to h a a$ and $pp\\to h N N$ production cross sections and selection efficiencies.","marker":"[11]"},{"why":"Provides the ALP effective field theory UFO model file that is modified with FeynRules to simulate the dimension-six ALP-Higgs interaction.","marker":"[12]"},{"why":"Provides the FeynRules package that generates the UFO model files for both the ALP and sterile neutrino signal simulations.","marker":"[13]"},{"why":"Provides the generic sterile neutrino UFO model that is modified for the $\\lambda_3/M_*$ operator used in the $pp\\to h N N$ simulation.","marker":"[18–22]"},{"why":"Supplies the significance formula used for the HL-LHC sensitivity projections with and without systematic uncertainty.","marker":"[30]"}],"fun_headline_variants":["Mono-Higgs probe sets fresh limits on ALPs and sterile neutrinos","ALP and sterile neutrino couplings squeezed by mono-Higgs data","One missing-energy channel constrains two hidden sectors at LHC","Mono-Higgs + MET: new bounds for ALPs and sterile neutrinos","HL-LHC eyes sharper ALP and sterile neutrino sensitivity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The HL-LHC projections assume that the numbers of background events in each missing-energy bin scale from 13 to 14 TeV by the ratios of the total $t\\bar t$, $W$+heavy-flavor, and $Z$+heavy-flavor production cross sections, which presumes the missing-energy shapes and selection efficiencies are unchanged between the two energies.","fun_headline_variants_meta":{"raw":{"variants":["Mono-Higgs probe sets fresh limits on ALPs and sterile neutrinos","ALP and sterile neutrino couplings squeezed by mono-Higgs data","One missing-energy channel constrains two hidden sectors at LHC","Mono-Higgs + MET: new bounds for ALPs and sterile neutrinos","HL-LHC eyes sharper ALP and sterile neutrino sensitivity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000186,"raw_usage":{"total_tokens":1432,"prompt_tokens":1158,"completion_tokens":274,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":774,"completion_tokens_details":{"reasoning_tokens":179}},"tokens_in":774,"tokens_out":274,"duration_ms":3015,"temperature":1.0,"reasoning_tokens":179,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:04:39.661516+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the $t\\bar t$, $W$+heavy-flavor, and $Z$+heavy-flavor backgrounds at 14 TeV with the full parton-level and detector simulation of the resolved-region selection, and compare the resulting per-bin event counts with the rescaled values in Table II; if the missing-energy distributions shift with center-of-mass energy, the projected coupling limits would move accordingly.","supporting_citations":[{"cited_title":"As in the earlier ALP analysis, we adopt the signal event analysis framework from the ATLAS paper","cited_arxiv_id":null,"evidence_quote":"Supplies the 13 TeV mono-Higgs search data, the resolved-region selection cuts, and the model-independent upper limits on visible cross sections that the exclusion bounds are derived from."}],"review_version":1}