{"id":"b3c6aab7-0324-489e-aac8-848a1536451a","arxiv_id":"2412.15138","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The paper shows that quantum interference between a 365 GeV pseudoscalar Higgs and the QCD continuum significantly changes its measured ttbar rate, offering observables to distinguish such a Higgs from toponium.","lead":"This paper studies how to tell apart two explanations for an excess of top-quark pair events seen by CMS at the LHC: a new pseudoscalar Higgs particle or a 'toponium' bound state of top and anti-top quarks. It shows that interference between the Higgs signal and ordinary QCD background changes the measured rate in a way that could be used to discriminate between the two explanations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The wide-window interference ratio and energy-dependence test rest on an unquantified constant K_i=1.6 applied to the off-shell tail, where the Γ_HO/Γ_LO≈1.6 width-compensation argument fails; a dedicated NLO calculation is needed.","rationale":"The paper's stated goal is to show that interference between gg→A→ttbar and the QCD continuum produces observable integrated-rate and energy-dependence effects that can distinguish A from toponium. I focused on the two headline numbers (ratios ≈1.2 and ≈4, and the 13→30 TeV growth) and traced them back to Section 3. They all depend on the unknown NNLO interference K-factor, which is set to the geometric mean 1.6. The authors' consistency argument—that Γ_HO/Γ_LO≈1.6 cancels the K-factor—works only at the pole; the wide-window observable is dominated by the off-shell tail where the width is irrelevant, so the K-factor directly inflates the tail. This is the most load-bearing point: without an uncertainty estimate or a dedicated NLO calculation, the quantitative discrimination claim is not yet robust. The paper is transparent and honest about the approximation, and the reader's CONDITIONAL verdict is appropriate; a change of verdict is not needed, only the specified test to validate the key numbers.","tokens_in":13465,"tokens_out":13046,"duration_ms":116795,"concrete_test":"Recompute the interference contribution to the m_tt distribution using the higher-order width Γ_HO=6.5 GeV in the A propagator and K_i=1.0, as a proxy for the resummed NLO propagator, and compare the integrated signal-only/(signal+interference) ratios in the [345,375] and [345,600] GeV windows with the paper's values (≈1.2 and ≈4). If the wide-window ratio decreases by more than ≈30% (e.g., from 4 to below 3), the assumed K_i=1.6 shortcut is not reliable and the discrimination claim needs a full NLO calculation (e.g., MG5_aMC@NLO with a complex-mass A) to be quantitative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3's central quantitative results—the signal-only/(signal+interference) ratios of ≈1.2 over m_tt∈[345,375] GeV and ≈4 over [345,600] GeV, and the ~20% growth of this ratio between 13 and 30 TeV—are computed with the unknown NNLO interference approximated by a constant K-factor K_i=1.6 multiplying the leading-order interference, which uses the leading-order width Γ_LO≈4.2 GeV. The authors justify this by noting Γ_HO/Γ_LO≈1.6, so the width ratio 'compensates' the K-factor. This compensation is only operative near the resonance pole: for |m_tt^2−M_A^2|≫M_A Γ, the Breit-Wigner denominator is dominated by the real part and is essentially independent of the width. A constant K_i=1.6 therefore enhances the off-shell tail (where the real part of the interference extends to m_tt≈600 GeV) by ~60% relative to a 'resummed-width, no-extra-K' treatment. The wide-window ratio and the energy dependence are governed precisely by this tail, so the headline numbers scale almost linearly with the guessed K_i. No uncertainty band or scan over K_i is given; if the true K_i is 1.2 instead of 1.6, the factor-of-4 ratio could decrease to ~2, substantially weakening the claim that the wide-window integrated rate is a clean discriminator.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies how to discriminate a pseudoscalar Higgs boson A from a toponium quasi-bound state as explanations of the CMS ttbar threshold excess. It works in an effective model with only a new CP-odd singlet A coupling to tops, fixes M_A=365 GeV and g_Att=0.78 to reproduce the CMS rate of 7.1 pb, and computes the line shape of gg->A->ttbar including its interference with the gg->ttbar QCD continuum, using K-factors K_s~2, K_b~1.3, K_i~1.6. It claims integrated-rate ratios signal-only over signal-plus-interference of ~1.2 in [345,375] GeV and ~4 in [345,600] GeV, with a ~20% increase of this ratio between 13 and 30 TeV, and argues these are discriminating observables absent for toponium. It also estimates associated hA and ttbar A production to be small at the LHC but observable at a 100 TeV collider.","tokens_in":13777,"tokens_out":9055,"duration_ms":84356,"significance":"The qualitative observation is useful: for a narrow pseudoscalar near the ttbar threshold the peak-dip interference may be unresolved, but integrated-rate and energy-dependence tests could in principle distinguish A from toponium. The paper is transparent about using SUSHI, HDECAY, HPAIR, and HQQ with stated inputs, and it credits the weakness of the CMS 20% mass resolution explicitly. However, the quantitative discrimination claims are not yet robust: they rest on an uncalculated NNLO K-factor for the interference, on a coupling fitted to the CMS central value, and on toponium inputs from another paper. The central idea is defensible, but the numerical headline numbers need a dedicated higher-order treatment or an uncertainty scan before they can support the stated conclusions.","major_comments":[{"comment":"The values 1.2, 4, and the ~20% energy growth in Fig. 3 and the surrounding text are computed by multiplying the leading-order gg->A interference by an assumed constant NNLO K-factor K_i=1.6. The argument that the width ratio Gamma_HO/Gamma_LO ~ 1.6 compensates this K-factor is not valid in the off-shell region: for m_tt^2 - M_A^2 much larger than M_A Gamma_A the Breit-Wigner denominator is dominated by the real part, so the choice of width is irrelevant and the interference tail out to m_tt ~ 600 GeV is simply rescaled by the guessed K_i. Because the wide-window ratio and the energy dependence are controlled by exactly this tail, the quoted numbers scale with K_i and no uncertainty band is given. Please provide either a dedicated NLO interference calculation, a scan over K_i (e.g., 1.0-2.0), or an explicit estimate of the resulting uncertainty; as written, the quantitative discrimination claim is not supported.","section":"Section 3, paragraph beginning 'For consistency'"},{"comment":"As written this claim is circular. The parameter g_Att=0.78 has just been chosen so that the signal rate matches the CMS value 7.1 pb (with Gamma_LO=4.2 GeV following from that choice), so the statement that interference is needed to obtain 7.1 pb is true by construction rather than a predictive result. Please clarify what is fixed by CMS (signal-only cross section, signal-plus-interference cross section, or the coupling itself) and propagate the experimental uncertainty on the 7.1 pb normalisation through the quoted ratios.","section":"Section 3, sentence 'In fact, the interference is crucial ...' and Fig. 3"},{"comment":"The comparison of the energy dependence of A production with the toponium curve from Ref. [31] is presented without stating the common inputs (PDF set, m_t value, QCD scales, K-factors) used for both curves. To support the discrimination claim the two calculations should be matched in these inputs, or the sensitivity of the ~20% difference to the toponium calculation should be quantified.","section":"Section 3, last paragraph and Fig. 3 (right)"}],"minor_comments":[{"comment":"The sentence introducing associated A production contains a stray 'A' after 't tbar'; it should read 'gg/q qbar -> t tbar A.'","section":"Abstract"},{"comment":"The caption quotes Gamma_A = 4.3 GeV while the text quotes Gamma_LO_A ~ 4.2 GeV for M_A=365 GeV and g_Att=0.78; please make these values consistent.","section":"Caption of Fig. 1 and Section 3 text"},{"comment":"The sentence 'the total width Gamma_A is completely fixed by this normalisation' should be qualified: CMS considered several ad hoc width choices, and here the width is fixed only after a specific interpretation of the CMS signal rate is adopted.","section":"Section 2, text after Eq. (4)"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about the K_i approximation and about the toponium interpretation being preferred by Occam's razor. The main issue is that the central quantitative observables are more sensitive to the guessed interference K-factor than the presentation suggests, and the fit-to-CMS normalisation makes the 'interference is crucial' statement partly tautological. I therefore recommend major revision rather than rejection: a K_i scan or a matched NLO treatment, plus a clear statement of what is fitted versus predicted, would make the claims quantitative. If the CMS excess is not confirmed by ATLAS, the numerical anchor is lost, but the general formalism for pseudoscalar-toponium discrimination would remain useful."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"John,\n\nHere's my take on Djouadi, Ellis, Quevillon. It's a solid phenomenological study that applies existing interference calculations to the CMS 365 GeV ttbar excess. What's actually new is not the interference formalism itself—that was in their earlier papers and CMS discussions—but the specific numbers: the integrated-window ratios of 1.2 and 4, the energy-dependence change of ~20% when going from 13 to 30 TeV, and the hA and ttbarA cross sections for the benchmark point that matches the CMS signal rate.\n\nThe paper gets credit for being transparent. They state their approximations, list the K-factors for signal and background, mention the missing NLO for the interference, and use established tools (SUSHI, HDECAY, HPAIR, HQQ). They're also honest that toponium is the more plausible explanation, so the paper reads as a careful \"if it's a pseudoscalar, here's how to tell\" rather than a claim about the true nature of the excess.\n\nNow the soft spot, and it's the same one flagged in the stress-test note, which I agree with. The central quantitative claims—the wide-window ratio near 4 and the energy-dependence test—depend on an unquantified constant K_i=1.6 for the interference, taken as the geometric mean of the signal and background K-factors. The authors try to justify this by noting that the higher-order width ratio Γ_HO/Γ_LO ≈ 1.6 compensates. That argument works near the pole, but the wide-window ratio is dominated by the off-shell tail, where the Breit-Wigner denominator is essentially width-independent. So the compensation fails exactly where the observable lives. If the true K_i is 1.2 instead of 1.6, the factor-of-4 ratio drops to roughly 2, which would substantially weaken the claim. The paper shows no scan over K_i and gives no uncertainty estimate.\n\nThat said, the qualitative conclusion is likely robust: interference matters, integrated windows can capture it, and the energy dependence differs from toponium. What's uncertain is the size of the effect. The coupling g_Att is fitted to the CMS excess, so this is a consistent scenario rather than a prediction, which the authors recognize.\n\nBottom line: this deserves a careful referee, not a desk reject. The referee should ask for an explicit systematic uncertainty on K_i, or at least a scan. With that, it's publishable as a useful addition to the conversation about the CMS excess. I'd bring it to our reading group, and I'd cite it if I wrote on this topic, though I'd note the K_i caveat.\n\nBest,\n[Your name]","headline":"A transparent, well-crafted phenomenological study of pseudoscalar-vs-toponium discrimination for the CMS 365 GeV excess, but the central discrimination ratios come with an unquantified interference K-factor that deserves a systematic uncertainty estimate.","tokens_in":14376,"tokens_out":2678,"would_cite":true,"duration_ms":21048,"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":"Interference between a 365 GeV pseudoscalar Higgs and the top-pair QCD continuum is large, measurable through mass-window and energy ratios, and absent for toponium.","keywords":["pseudoscalar Higgs","toponium","top quark pair production","signal-background interference","ttbar threshold excess","gluon fusion","associated Higgs production","two-Higgs-doublet model"],"falsifier":"Measure the $t\\bar t$ invariant-mass distribution with resolution improved toward 10% and integrate the cross section over $[345,375]$ GeV and over $[345,600]$ GeV. If the ratio of signal-only to signal-plus-interference is close to 1.2 in the narrow window and close to 4 in the wide window, the interference picture is supported; if the ratios come out near 1, or the wide-window ratio does not change with collision energy, the claimed discriminator is ruled out.","tokens_in":13183,"feed_emoji":"⚛️","tokens_out":9399,"duration_ms":58110,"temperature":0.7,"pith_summary":"The paper aims to settle how a future experiment can tell apart two explanations of an excess of top-quark-pair events near the pair-production threshold: a new pseudoscalar Higgs boson $A$ decaying to $t\\bar t$, or a toponium quasi-bound state predicted by QCD. Working in an effective theory with $M_A=365$ GeV and a reduced top-quark coupling $g_{Att}=0.78$ chosen to match the observed rate, it shows that the interference between $gg\\to A\\to t\\bar t$ and the QCD continuum $gg\\to t\\bar t$ is large, so ignoring it badly miscalculates the yield. The key discriminating observables are ratios of integrated cross sections: about 1.2 over a narrow mass window and about 4 over a wide window, with an additional ~20% growth when the collider energy rises from 13 to 30 TeV. Toponium, where the QCD continuum is itself the signal, has no equivalent interference. The paper also estimates that associated $hA$ and $t\\bar t A$ production, although tiny at 13 TeV, would become observable at much higher energies, offering another route to distinguish the two scenarios.","feed_headline":"A 365 GeV pseudoscalar Higgs leaves a fourfold interference signature","feed_subtitle":"Mass-window and energy-scaling ratios separate a heavy pseudoscalar from toponium.","key_machinery":"The machinery is the interference term between the $s$-channel pseudoscalar amplitude $gg\\to A\\to t\\bar t$ and the QCD continuum amplitude $gg\\to t\\bar t$, computed following earlier work at leading order with full top-quark mass effects and with higher-order QCD corrections inserted as K-factors: about 2 for the signal, 1.3 for the background, and an assumed 1.6 for the interference (the geometric mean of the two), with the leading-order width $\\Gamma_{\\rm LO}^{A}\\simeq 4.2$ GeV used in the interference term because the ratio $\\Gamma_{\\rm HO}^{A}/\\Gamma_{\\rm LO}^{A}\\simeq 1.6$ compensates the K-factor. This interference term is what produces the peak-dip line shape and the large dependence of total rates on the $m_{t\\bar t}$ integration window and on $\\sqrt{s}$.","core_discovery":"The central claim is that for a pseudoscalar $A$ with mass just above the $2m_t=345$ GeV threshold, the resonant electroweak process $gg\\to A\\to t\\bar t$ and the QCD continuum process $gg\\to t\\bar t$ interfere coherently, and this interference is an essential part of the production rate rather than a small correction. For the parameter point $M_A=365$ GeV, $g_{Att}=0.78$, the total width is fixed to about 4.2 GeV at leading order, and the real part of the interference is negative for invariant masses above $M_A$ and extends far beyond the resonance width, while the imaginary part is always negative. The result is a strong suppression of the integrated yield: the signal-only to signal-plus-interference ratio is roughly 1.2 when $m_{t\\bar t}$ is integrated over $[345,375]$ GeV and roughly 4 over $[345,600]$ GeV, and this ratio grows about 20% as the collision energy increases from 13 to 30 TeV. Because toponium production is itself the QCD continuum, it displays no such interference, so these rate ratios and their energy dependence offer a concrete experimental discriminator even though the peak/dip structure itself is smeared by the invariant-mass resolution.","pith_inferences":["A direct measurement of the narrow-to-wide integrated cross-section ratio at a hadron collider would act as a model-independent interference test, without needing to resolve the line shape; the paper's numbers imply a clean target.","The same mass-window ratio technique could be applied to any narrow spin-0 resonance decaying to $t\\bar t$ well above threshold, not just the specific 365 GeV point, since the real-part interference always extends far beyond the width.","The energy-dependence test could be sharpened by measuring the same windowed ratio at two collider energies, which reduces dependence on absolute luminosity and acceptance uncertainties.","If a future two-loop calculation replaces the geometric-mean K-factor, the main qualitative conclusion (interference suppresses wide-window yields and grows with energy) is likely to survive, but the exact 1.2 and 4 numbers would shift."],"forward_implications":["If the excess is due to an $A$ boson, the observed $t\\bar t$ yield in a wide invariant-mass window (345 to 600 GeV) is about a factor of 4 smaller than the signal-only estimate, while a narrow window (345 to 375 GeV) differs by only about 20%.","The signal-plus-interference rate has a distinctive energy scaling: the ratio of signal-only to signal-plus-interference increases by about 20% between 13 and 30 TeV, and similarly between 30 and 100 TeV, unlike toponium production.","With current invariant-mass resolution around 20%, the peak/dip structure is hidden, but comparing integrated yields in different $m_{t\\bar t}$ windows is a practical substitute.","Associated production rates $gg\\to hA$ and $gg/q\\bar q\\to t\\bar t A$ are about 25 fb and 15 fb at 13 TeV, respectively; at a 100 TeV machine they rise by factors of roughly 50 and 200, making them detectable with high luminosity.","Subdominant $A$ decay channels such as $A\\to\\gamma\\gamma, Z\\gamma, ZZ, WW$ have small branching fractions in this scenario and would mimic toponium, so they are not a clean discriminator."],"supporting_citations":[{"why":"Reports the observed $t\\bar t$ threshold excess whose overall rate fixes the benchmark values $M_A=365$ GeV and $g_{Att}=0.78$.","marker":"[1]"},{"why":"Supplies the leading-order calculation of the $gg\\to A\\to t\\bar t$ signal, the $gg\\to t\\bar t$ background, and their real and imaginary interference terms used in Section 3.","marker":"[12]"},{"why":"Supplies the higher-order $gg\\to A$ cross-section calculation used to normalize the signal production rate.","marker":"[22]"},{"why":"Supplies the higher-order $A\\to t\\bar t$ total width ($\\Gamma_{\\rm HO}^A\\simeq 6.5$ GeV) used alongside the leading-order width in the interference.","marker":"[26]"},{"why":"Provide the NNLO QCD corrections that give the signal K-factor of about 2.","marker":"[23, 24]"},{"why":"Provide the NNLO QCD corrections that give the $t\\bar t$ background K-factor of about 1.3.","marker":"[28]"},{"why":"Provides the energy dependence of toponium production used as the comparison in the $\\sqrt{s}$ scaling plot.","marker":"[31]"},{"why":"Earlier study of the same effective pseudoscalar scenario that fixes the effective-theory setup and the treatment of $A$ couplings.","marker":"[2]"}],"fun_headline_variants":["Interference ratios tell Higgs from toponium","Mass-window ratios separate heavy A from toponium","Pseudoscalar vs toponium: rate ratios discriminate","Interference suppresses yield, marking pseudoscalar Higgs","Energy scaling of t tbar yield reveals pseudoscalar"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The size of the predicted interference effects assumes that the unknown higher-order QCD correction to the interference is a factor of 1.6 (the geometric mean of the signal and background corrections), and that using the leading-order width in the interference term with a compensating width ratio is valid; if the true correction differs, the quoted 1.2 and 4 ratios and the 20% energy dependence would shift.","fun_headline_variants_meta":{"raw":{"variants":["Interference ratios tell Higgs from toponium","Mass-window ratios separate heavy A from toponium","Pseudoscalar vs toponium: rate ratios discriminate","Interference suppresses yield, marking pseudoscalar Higgs","Energy scaling of t tbar yield reveals pseudoscalar"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000168,"raw_usage":{"total_tokens":1344,"prompt_tokens":1114,"completion_tokens":230,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":730,"completion_tokens_details":{"reasoning_tokens":152}},"tokens_in":730,"tokens_out":230,"duration_ms":2917,"temperature":1.0,"reasoning_tokens":152,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T11:35:24.084590+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the $t\\bar t$ invariant-mass distribution with resolution improved toward 10% and integrate the cross section over $[345,375]$ GeV and over $[345,600]$ GeV. If the ratio of signal-only to signal-plus-interference is close to 1.2 in the narrow window and close to 4 in the wide window, the interference picture is supported; if the ratios come out near 1, or the wide-window ratio does not change with collision energy, the claimed discriminator is ruled out.","supporting_citations":[{"cited_title":"Djouadi, J","cited_arxiv_id":null,"evidence_quote":"Supplies the leading-order calculation of the $gg\\to A\\to t\\bar t$ signal, the $gg\\to t\\bar t$ background, and their real and imaginary interference terms used in Section 3."},{"cited_title":"Harlander, S","cited_arxiv_id":null,"evidence_quote":"Supplies the higher-order $gg\\to A$ cross-section calculation used to normalize the signal production rate."},{"cited_title":"Djouadi, J","cited_arxiv_id":null,"evidence_quote":"Supplies the higher-order $A\\to t\\bar t$ total width ($\\Gamma_{\\rm HO}^A\\simeq 6.5$ GeV) used alongside the leading-order width in the interference."},{"cited_title":"Nason, S","cited_arxiv_id":null,"evidence_quote":"Provide the NNLO QCD corrections that give the $t\\bar t$ background K-factor of about 1.3."},{"cited_title":"Djouadi, R","cited_arxiv_id":null,"evidence_quote":"Earlier study of the same effective pseudoscalar scenario that fixes the effective-theory setup and the treatment of $A$ couplings."}],"review_version":1}