{"id":"a72284ac-025b-48b0-9b20-d8579e0db0ef","arxiv_id":"2501.02224","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A 3-10 TeV muon collider could detect a heavy dark Z boson in photon-associated production, reaching kinetic-mixing strengths down to about 10^-3 near the kinematic limit by using resolution-optimized recoil-mass and electron-pair-mass windows.","lead":"This paper calculates how well future multi-TeV muon colliders could detect a heavy dark Z boson produced together with a photon, by measuring the photon's recoil mass. It finds sensitivity to the kinetic-mixing strength down to about 10^-3 for dark Z masses near the collider's maximum energy, beyond the reach of a 100 TeV proton collider.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Near-threshold Δm_recoil values in Tables II/III appear inconsistent with the paper's own Eq. (16) for the stated benchmark masses; until resolved, the quoted significances and ε reach are not reproducible.","rationale":"I read the central claim as a detector-level sensitivity projection whose endpoint is set by Δm_recoil through Eq. (17). The paper's own Eq. (16) and the stated Delphes parameters form an internally testable resolution model. My calculation shows a factor of about 2.3 discrepancy between the quoted widths for the tabulated masses and the widths predicted by Eq. (16). This is more specific than the beam-spread concern because it does not rely on external assumptions about future muon beams: it tests the paper against itself. It is load-bearing because the optimized windows enter directly into the significances in Tables II/III, which are the evidence for the abstract's claim of substantially surpassing a 100 TeV pp collider near threshold. If the fitted widths are actually ~4 GeV, the background rises and the quoted S=12.4 and the derived ε sensitivities weaken by roughly 20–30%. The physics idea and the qualitative conclusion may still survive, which is why the reader's CONDITIONAL verdict remains appropriate; the discrepancy is an additional verification condition rather than a reason to reject the approach. The proposed rerun with the exact card and benchmark masses settles whether this concern lands.","tokens_in":21804,"tokens_out":23350,"duration_ms":238283,"concrete_test":"Rerun the detector-level simulation for M_ZD=2.7 TeV at √s=3 TeV and M_ZD=9.7 TeV at √s=10 TeV with the exact MuonColliderDet.tcl card used in the paper; fit a Gaussian to the m_recoil distribution and compare σ to the analytic prediction from Eq. (16), σ_m=(√s/M)Eγ sqrt((a/√Eγ)^2+b^2), which gives ≈4.3 GeV and ≈4.1 GeV respectively. Repeat for M_ZD=√s−100 GeV to verify the endpoint values in Figure 8. If the fitted widths are ≈1.8 GeV, then the quoted resolution parameters or the benchmark labels are wrong; if the widths are ≈4 GeV, Tables II/III and the ε values derived from them must be recomputed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on the optimized mass windows of Eq. (17), whose widths Δm_recoil are taken from Gaussian fits using the Delphes card and Eq. (16). For the tabulated benchmark points, those widths are not consistent with the paper's own resolution model. At √s=3 TeV and M_ZD=2.7 TeV, the associated photon has Eγ=(s−M^2)/(2√s)=285 GeV. Eq. (16) with a=0.156, b=0.01 gives σ_E/E=1.36%, σ_E=3.9 GeV, and Δm_recoil≈(√s/M)σ_E=4.3 GeV, not the 1.66 GeV quoted in Table II/Figure 4. At √s=10 TeV and M_ZD=9.7 TeV, Eγ=295 GeV gives Δm≈4.1 GeV, not the 1.77 GeV quoted in Table III. The stated 1.7–1.9 GeV values correspond instead to Eγ≈100 GeV, i.e. to M≈√s−100 GeV, not to the masses in the cut-flow tables. If the correct widths for the tabulated masses are ~4 GeV, the m_recoil window in Table III widens from ±3.5 GeV to ±8.2 GeV, the background grows by roughly a factor of 2.3, S falls from 12.4 to about 8, and the ε sensitivity at that point degrades by ~20–30%. The same check applies to the 2.7 TeV row. These cut-flow tables are the only detector-level validation of the optimized-window method, so the headline O(10^-3) reach claims inherit this unresolved discrepancy.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a search for a heavy dark Z boson (Z_D) with mass above 1 TeV at multi-TeV muon colliders, using the associated production process μ+μ− → Z_D γ followed by Z_D → jjX or Z_D → e+e−. The central idea is to replace fixed mass windows by M_ZD-dependent cuts on the photon recoil mass m_recoil and the e+e− invariant mass m_ee, with the widths Δm_recoil and Δm_ee extracted from Gaussian fits to Delphes-simulated signal samples. The authors present cut-flow tables, significance estimates, and 2σ/5σ sensitivity contours in the (M_ZD, ε) plane for 3, 6, and 10 TeV muon colliders, concluding that their optimized recoil-mass technique reaches ε ~ 2–4×10^-3 near the kinematic limit and substantially surpasses the projected reach of a 100 TeV proton-proton collider at large M_ZD.","tokens_in":21923,"tokens_out":11220,"duration_ms":112645,"significance":"The physics motivation and the main mechanism are attractive: the recoil-mass relation m_recoil² = s − 2√s E_γ is exact at Born level, and exploiting the better photon energy resolution for softer photons is a genuinely useful idea. The paper provides a reproducible-looking simulation chain (MadGraph, Pythia, Delphes with the MuonColliderDet card), explicit background lists, cut-flow tables, and sensitivity projections, which are concrete assets. The central claim is nevertheless numerically anchored in the quoted Δm_recoil values and in the absence of machine-level smearing; if those are corrected, the absolute significances and the ε-reach contours would change, even though the qualitative trend of improved sensitivity at high M_ZD is likely to survive.","major_comments":[{"comment":"The quoted Δm_recoil values are not consistent with the paper’s own photon energy resolution model. For √s = 3 TeV and M_ZD = 2.7 TeV, the associated photon has Eγ = (s − M²)/(2√s) = 285 GeV; Eq. (16) with a = 0.156 and b = 0.01 gives σ_E/E ≈ 1.36%, σ_E ≈ 3.9 GeV, and Δm_recoil ≈ (√s/M) σ_E ≈ 4.3 GeV, not the 1.66 GeV quoted in Table II. Similarly, for √s = 10 TeV and M_ZD = 9.7 TeV, Eq. (16) gives Δm_recoil ≈ 4.1 GeV, not the 1.77 GeV quoted in Table III. The quoted values correspond instead to Eγ ≈ 100 GeV, i.e., M_ZD ≈ √s − 100 GeV, rather than to the masses in the cut-flow tables. Because Eq. (17) sets the window as ±2Δm_recoil, the Table II/III significances do not follow from the stated resolution model. Please either reconcile the fitted widths with Eq. (16) by explaining how the Delphes reconstruction achieves a resolution roughly 2.4 times better than the card parameterization, or correct the widths and rerun the cut-flow and sensitivity computations.","section":"§III–IV, Eq. (16), Tables II and III"},{"comment":"The near-threshold reach, which produces the headline ε values of 2–4×10^-3, depends on Δm_recoil values of order 1.7–1.9 GeV at M_ZD close to √s. In the simulations these widths come only from the Delphes photon energy resolution; the analysis does not include the beam energy spread or beamstrahlung of a realistic multi-TeV muon collider. A momentum spread of order σ_p/p ~ 10^-3 smears √s by about 3 GeV at 3 TeV and 10 GeV at 10 TeV, and near threshold dm_recoil/d√s ≈ M/√s ≈ 1, so this machine-level smearing is as large as or larger than the quoted detector widths. Please quantify the effect of the beam energy spread on Δm_recoil and on the optimized mass windows; this is essential for the ε ~ O(10^-3) sensitivity claims in Section IV.D.","section":"§IV, Eq. (17), Fig. 4"},{"comment":"The widths Δm_recoil and Δm_ee are obtained from Gaussian fits to the same detector-level signal samples on which the mass-window cuts are then applied, and the multiplier 2 in Eq. (17) is not justified by an independent scan or by a signal-plus-background fit. This is an in-sample optimization: it does not invalidate the qualitative mechanism, but it makes the absolute significances and the derived ε contours optimistic in a way that is not quantified. Please report how the significance and the reach vary with the window multiplier (for example 1.5Δ, 2Δ, 2.5Δ) and, ideally, set the widths using a procedure that does not reuse the signal sample under test.","section":"§IV.B–IV.D, Eq. (17)"}],"minor_comments":[{"comment":"The comparison with the HL-LHC and 100 TeV pp collider uses the 2σ sensitivity curves from Ref. [30], while the MuC results are shown as both 2σ and 5σ contours; please state explicitly that the hadron-collider curves are the appropriate 2σ or 95% CL limits, so the comparison is apples-to-apples.","section":"§IV.D, Fig. 8"},{"comment":"The background cross sections are given without systematic uncertainties; a short discussion of the dominant theoretical and detector-level uncertainties (scale choices, jet energy scale, lepton veto efficiency) would help assess the robustness of the quoted significances.","section":"§IV.A, Table I"},{"comment":"The figure captions do not identify which curve corresponds to which M_ZD value; please add legends or explicit labels to Figures 2, 5, and 6.","section":"§III, Fig. 2 and §IV.B, Figs. 5–6"},{"comment":"The Gaussian fit to the asymmetric m_ee distribution is restricted to a ±25% window around the true mass; please specify how the fitted width changes with the choice of this window and whether the quoted Δm_ee values are sensitive to it.","section":"§III, Fig. 3 and fit procedure"},{"comment":"The phrase “substantially surpassing the reach of a 100 TeV proton-proton collider” is used for the heavy-mass regime; in the lighter-mass region the MuC does not outperform the hadron colliders, and the abstract could state this qualification more precisely.","section":"Abstract and §I"},{"comment":"There are minor typographical and formatting issues, including inconsistent capitalization of “Delphes”/“DELPHES” and extra spacing in “F ASER/F ASER2”; these should be cleaned up.","section":"Various"}],"recommendation":"major_revision","confidential_remarks":"The paper’s central idea is promising and the simulation setup is mostly transparent, but the numerical discrepancy between Eq. (16) and the Δm_recoil values in Tables II/III, together with the absence of beam energy spread in the resolution model, directly affects the quoted significances and the headline ε reach. The discrepancy is not necessarily fatal—it may be a typo in the benchmark masses or an unstated reconstruction effect—but it must be resolved before the sensitivity projections can be trusted. I would also encourage the authors to make the Delphes card and analysis code available for reproducibility."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a competent, workmanlike sensitivity projection for a heavy dark Z at 3–10 TeV muon colliders. The core idea—make the recoil-mass and m_ee windows M_ZD-dependent because the detector resolution depends on the photon/lepton energy—is correct and is the paper's real contribution. The qualitative trends are right: heavier Z_D means softer photons, better photon resolution, and therefore a tighter m_recoil window; lighter Z_D means lower-energy electrons and a tighter m_ee window. The crossover near sqrt(s)/2 is a nice observation. The simulation pipeline is transparent and the cut-flow tables are easy to follow.\n\nThe soft spots are real. First, the Δm_recoil values in Tables II and III do not match the paper's own resolution model. For M_ZD=2.7 TeV at 3 TeV and M_ZD=9.7 TeV at 10 TeV, the associated photon has Eγ ≈ 285–296 GeV; Eq. (16) with the stated Delphes card then gives σ_E ≈ 4 GeV and Δm_recoil ≈ 4.1–4.3 GeV, not the quoted 1.66/1.77 GeV. The quoted numbers correspond to Eγ ≈ 100 GeV, i.e. M_ZD ≈ sqrt(s)−100 GeV. This is a genuine inconsistency that should be fixed. Having said that, the stress-test's claim that the significance falls to about 8 if you use the correct width is not right: a wider window also captures more signal, and my rough rescaling gives S of order 13 at the 10 TeV point rather than 8. So the discrepancy is a reproducibility problem, not necessarily a fatal one for the qualitative reach.\n\nThe more serious concern is beam energy spread. A 0.1% spread smears sqrt(s) by 3–10 GeV at these energies, which is larger than the 1–2 GeV Δm_recoil that drives the endpoint sensitivity. The paper assumes mono-energetic beams. Unless beam-energy spread is actually much smaller, the near-threshold O(10^-3) reach is optimistic. BIB is also not modeled, and the significances are pure statistics with no systematics; these are common limitations in fast-sim studies, but they should be stated more prominently.\n\nWho is this for? People building the physics case for a multi-TeV muon collider and dark-sector hunters. I'd send it to a serious referee, but the authors need to fix the resolution tables and add a quantitative discussion of beam-energy spread before I'd trust the endpoint numbers.","headline":"A sensible and useful MuC sensitivity study whose optimized recoil-mass windows are a legitimate refinement, but the benchmark tables are internally inconsistent and the near-threshold reach neglects beam energy spread, so the headline O(10^-3) numbers are optimistic.","tokens_in":22746,"tokens_out":10402,"would_cite":false,"duration_ms":102163,"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 multi-TeV muon collider could spot heavy dark $Z$ bosons down to kinetic mixing of a few parts in a thousand.","keywords":["dark Z boson","kinetic mixing","muon collider","recoil mass","associated production","dark sector","beam-induced background","beyond the Standard Model"],"falsifier":"Measure the photon recoil-mass resolution on a known standard candle at a 3 TeV muon collider, for example $\\mu^+\\mu^- \\to Z\\gamma$ with the $Z$ produced against a photon of about 300 GeV, and compare the fitted Gaussian width of $m_{\\rm recoil}$ with the $\\Delta m_{\\rm recoil} \\approx 1$--$2$ GeV used for $M_{Z_{\\rm D}}$ near 2.7 TeV; a width several times larger would rule out the claimed $\\varepsilon \\approx 3.9\\times 10^{-3}$, while a width at or below that level would support it.","tokens_in":21357,"feed_emoji":"⚛️","tokens_out":13978,"duration_ms":122151,"temperature":0.7,"pith_summary":"This paper argues that a multi-TeV muon collider is the right machine to search for a heavy dark $Z$ boson, the gauge boson of a dark U(1) sector that couples to Standard Model particles through kinetic mixing. The production mode that carries the argument is associated radiation, $\\mu^+\\mu^- \\to Z_{\\rm D}\\gamma$, because the dark $Z$ mass is then fixed by the collision energy and the measured photon energy through the recoil relation $m_{\\rm recoil}^2 = s - 2\\sqrt{s}\\, E_\\gamma$. Rather than applying one fixed mass window, the authors tune the $m_{\\rm recoil}$ and $m_{ee}$ windows to the detector resolution expected at each assumed mass, which lets them exploit the better photon resolution that comes with the softer photons radiated by heavier $Z_{\\rm D}$ bosons. Combining the dijet and $e^+e^-$ decay channels, they claim kinetic-mixing sensitivities $\\varepsilon \\approx 2\\text{--}4\\times 10^{-3}$ for masses within about 100 GeV of the beam energy at 3, 6, and 10 TeV muon colliders. If correct, that reach would substantially exceed the projected sensitivity of a 100 TeV proton-proton collider for heavy $Z_{\\rm D}$ masses, and the same recoil technique would still work if the $Z_{\\rm D}$ decays invisibly into dark-sector states.","feed_headline":"Probe heavy dark Z bosons down to 0.002 mixing at a muon collider","feed_subtitle":"Optimized recoil-mass cuts at 3–10 TeV would reach beyond a 100 TeV proton collider for heavy masses.","key_machinery":"The load-bearing object is the recoil-mass identity $m_{\\rm recoil}^2 = s - 2\\sqrt{s}\\, E_\\gamma$, which converts a single photon energy measurement into a dark-$Z$ mass without any assumption about how the $Z_{\\rm D}$ decays. The machinery that carries the analysis is the pair of optimized mass windows in Eq. (17), with widths $\\Delta m_{\\rm recoil}$ and $\\Delta m_{ee}$ extracted from Gaussian fits to detector-level distributions for each $M_{Z_{\\rm D}}$ and $\\sqrt{s}$. These widths encode the energy-dependent photon and electron resolutions: soft photons from heavy $Z_{\\rm D}$ bosons are measured with better relative precision, so the $m_{\\rm recoil}$ window can be made very narrow exactly where the cross section is largest, while the $m_{ee}$ window is tightest for lighter $Z_{\\rm D}$ bosons whose electron pairs are less energetic. The complementarity of the two windows is what lets a single analysis stay sensitive across the full kinematically allowed mass range.","core_discovery":"The central claim is that optimized, mass-dependent selection turns near-threshold production into a discovery channel for a heavy dark $Z$. The signal is a photon recoiling against the $Z_{\\rm D}$; the standard kinematic identity $m_{\\rm recoil}^2 = s - 2\\sqrt{s}\\, E_\\gamma$ fixes $M_{Z_{\\rm D}}$ from the photon energy alone. The optimized selections are $|m_{\\rm recoil} - M_{Z_{\\rm D}}| < 2\\Delta m_{\\rm recoil}$ and $|m_{ee} - M_{Z_{\\rm D}}| < 2\\Delta m_{ee}$, where $\\Delta m_{\\rm recoil}$ and $\\Delta m_{ee}$ are the standard deviations of Gaussian fits to detector-level $m_{\\rm recoil}$ and $m_{ee}$ distributions for each mass hypothesis. Because a heavier $Z_{\\rm D}$ leaves less energy to the photon, the photon energy resolution is better and $\\Delta m_{\\rm recoil}$ narrows to roughly a few GeV near the kinematic limit; because a lighter $Z_{\\rm D}$ produces a lower-energy electron pair, the $m_{ee}$ window is tighter in the low-mass regime. Using the $m_{ee}$ selection below about $\\sqrt{s}/2$ and the $m_{\\rm recoil}$ selection above it, and combining the $jjX$ and $e^+e^-$ channels, the authors obtain $2\\sigma$ sensitivities $\\varepsilon = 3.9\\times 10^{-3}$ at $\\sqrt{s} = 3$ TeV, $\\varepsilon = 2.7\\times 10^{-3}$ at 6 TeV, and $\\varepsilon = 2.1\\times 10^{-3}$ at 10 TeV for $M_{Z_{\\rm D}} = \\sqrt{s} - 100$ GeV, and they argue this substantially surpasses the reach of a 100 TeV proton-proton collider at such masses.","pith_inferences":["The near-threshold $\\varepsilon$ values depend on the calorimeter resolution assumed in the simulation and on neglecting beam-energy spread; if a real muon beam smears $\\sqrt{s}$ by several GeV at multi-TeV energies, the recoil peaks broaden and the quoted $O(10^{-3})$ sensitivities would degrade. This is an inference from the simulation setup, not a claim tested in the paper.","The optimized-window logic ought to transfer to any resonance produced with an associated photon at a lepton collider, such as a heavy Higgs boson or a generic $Z'$, so the same tuning procedure could be applied to other searches.","The cut-based analysis likely underestimates what a full spectral fit could do: using the whole $m_{\\rm recoil}$ shape rather than a single $\\pm 2\\Delta$ window would extract more information from the same events, and a binned-likelihood version is a natural next step.","Because beam-induced backgrounds are handled only through pseudorapidity and $p_T$ cuts, an experimental study with full background overlay would be needed to confirm the quoted acceptance, especially in the forward region."],"forward_implications":["For a $Z_{\\rm D}$ within about 100 GeV of the beam energy, a 3, 6, or 10 TeV muon collider could exclude or discover kinetic mixing at the level of a few parts in a thousand, beyond the projected 100 TeV proton-proton reach at such masses.","The photon-recoil measurement does not require knowing the $Z_{\\rm D}$ decay products, so the same search strategy remains valid if the dark $Z$ decays into invisible dark-sector states.","The optimal selection switches from $m_{ee}$ to $m_{\\rm recoil}$ near $M_{Z_{\\rm D}} \\approx \\sqrt{s}/2$, so a single fixed mass window would sacrifice sensitivity on one side or the other.","Each collision energy is most sensitive near its own kinematic limit, so a sequence of muon colliders at 3, 6, and 10 TeV would extend heavy dark-$Z$ coverage in stages, with the highest energy giving the smallest $\\varepsilon$."],"supporting_citations":[{"why":"Supplies the kinetic-mixing Lagrangian and the definition of the dark Z boson and its coupling $\\varepsilon$ to Standard Model fermions.","marker":"[6]"},{"why":"Provides the HL-LHC and 100 TeV pp luminosity and sensitivity limits used as the comparison baseline that the muon-collider reach is claimed to surpass.","marker":"[30]"},{"why":"Introduces the radiative-return/recoil-mass approach for associated production at a muon collider that the analysis builds on.","marker":"[44]"},{"why":"Defines the 3, 6, and 10 TeV muon-collider configurations with integrated luminosities and the detector assumptions behind the analysis.","marker":"[46]"},{"why":"Used to generate the parton-level signal and background cross sections and event samples.","marker":"[72]"},{"why":"Provides the fast detector simulation used to obtain the detector-level $m_{\\rm recoil}$ and $m_{ee}$ distributions and the resolution widths.","marker":"[74]"},{"why":"Supplies the parton shower and hadronization step in the event-generation chain.","marker":"[75]"}],"fun_headline_variants":["Heavy dark Z probed to 0.002 mixing at TeV muon colliders","Recoil-mass cuts reveal heavy dark Z at muon colliders","Muon collider recoil method beats 100 TeV pp for dark Z","Optimized recoil mass finds dark Z near threshold at muon colliders","TeV muon colliders target heavy dark Z with recoil trick"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quoted near-threshold sensitivities rest on the assumed energy resolution for soft photons, including a calorimeter constant term of about 1%, and on neglecting beam-energy spread and beamstrahlung; if the real resolution is worse or the beam smears $\\sqrt{s}$ by several GeV, the optimized windows must widen and the $O(10^{-3})$ reach shrinks.","fun_headline_variants_meta":{"raw":{"variants":["Heavy dark Z probed to 0.002 mixing at TeV muon colliders","Recoil-mass cuts reveal heavy dark Z at muon colliders","Muon collider recoil method beats 100 TeV pp for dark Z","Optimized recoil mass finds dark Z near threshold at muon colliders","TeV muon colliders target heavy dark Z with recoil trick"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000311,"raw_usage":{"total_tokens":2001,"prompt_tokens":1403,"completion_tokens":598,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":1019,"completion_tokens_details":{"reasoning_tokens":495}},"tokens_in":1019,"tokens_out":598,"duration_ms":5887,"temperature":1.0,"reasoning_tokens":495,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:16:23.704420+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the photon recoil-mass resolution on a known standard candle at a 3 TeV muon collider, for example $\\mu^+\\mu^- \\to Z\\gamma$ with the $Z$ produced against a photon of about 300 GeV, and compare the fitted Gaussian width of $m_{\\rm recoil}$ with the $\\Delta m_{\\rm recoil} \\approx 1$--$2$ GeV used for $M_{Z_{\\rm D}}$ near 2.7 TeV; a width several times larger would rule out the claimed $\\varepsilon \\approx 3.9\\times 10^{-3}$, while a width at or below that level would support it.","supporting_citations":[],"review_version":1}