{"id":"4c7338fa-7cef-401d-b649-2018987dd237","arxiv_id":"1908.01698","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Beam-beam deflection biases the FCC-ee small-angle Bhabha luminosity measurement by about 0.19% at the Z pole, and the proposed in-situ acollinearity and dimuon-based calibrations can reduce the residual to below the 10^-4 target.","lead":"Beam-beam electromagnetic fields at the proposed FCC-ee collider deflect the final-state electrons used to measure luminosity, creating a bias near 0.2%, roughly twenty times larger than the target precision. The paper quantifies this effect with two independent simulation approaches and proposes two in-situ methods, an acollinearity measurement in the luminometer and a dimuon-based method in the central detector, that determine the correction to the required accuracy.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Acol calibration in Sec. 5 is computed before LumiCal acceptance and clustering; since the focusing changes which events are selected, the measured Acol may not be the calibrated quantity.","rationale":"The paper's quantitative core is carefully built: Eq. 3.2 reproduces the quoted ~0.19% bias, the kick and angle relations are internally consistent, and the independent Guinea-Pig and Bassetti-Erskine numerical routes agree to about 10%. I am not objecting to the physics of the focusing, nor to the idea of measuring and correcting it. The soft spot is the chain that turns a measured Acol value into a luminosity correction. The in-situ method is proposed as self-contained using only the luminometer; if its calibration is computed without the acceptance and reconstruction effects that the measurement inevitably includes, the claimed 1-2% accuracy is unsupported. This concern is nested inside the reader's broader statement that the calibration curves are computed entirely inside the simulation framework, but it is more specific: even for leading-order Bhabha events, the acceptance selection can change the meaning of Acol. The proposed test is a well-scoped simulation that the paper itself says is beyond its current scope (Sec. 3.3.3), and Sec. 5.1 acknowledges that the true Acol relation requires dedicated simulations. Therefore the correct verdict remains CONDITIONAL, as the reader concluded, with the condition made more explicit. I would not reject: there is no demonstrated inconsistency, and the paper honestly flags the missing full-simulation step. I would not accept unconditionally until the test is performed.","tokens_in":18804,"tokens_out":9952,"duration_ms":112820,"concrete_test":"One decisive check: run BHWIDE radiative Bhabha events through Guinea-Pig, then through a GEANT4 LumiCal simulation with the FCAL clustering algorithm and the exact asymmetric-acceptance selection (64-86 mrad narrow, wide on the opposite side), for the nominal CDR parameters and for several of the parameter variations in Fig. 10/12. Recompute (a) the reconstructed Acol from the selected clusters and (b) the luminosity bias in the selected sample, then overlay these points on Fig. 12. If the reconstructed Acol-to-bias points stay within the +/-1e-4 band around the same linear fit, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the beam-beam luminosity bias can be corrected to a residual below 1e-4 rests on two in-model calibrations: the kick-to-bias relation of Fig. 10 and the Acol-to-bias relation of Fig. 12. The weaker link is the Acol method of Sec. 5. Eq. 5.1 defines Acol from generated-level (or pre-selection) Bhabha events: Sec. 3.3.3 states that the observable is computed from the px kick and final-state focusing for leading-order events, and the BHWIDE points in Fig. 12 use final-state charged leptons with only a 5 GeV cut, with LumiCal clustering and reconstruction explicitly deferred. But the actual Acol is measured on events that survive the LumiCal asymmetric acceptance, and the focusing itself changes which events survive: at the lower edge theta_min = 64 mrad and phi = 0, the total deflection is about 150 micro rad (Fig. 11, left), so a sizable fraction of lower-edge events with the largest positive Delta theta_+- is removed from that hemisphere, while at phi = pi the defocusing changes upper-edge admission differently. The mean of Delta theta_+- over the selected sample is therefore not the mean over the generated theta* distribution used to build Fig. 12. A percent-level distortion of Acol maps to roughly a 2e-5 change in the 0.19% bias; that is tolerable only if the distortion is absent or is included in the calibration. Similarly, radiative Bhabha events with cluster merging can shift Acol by an amount the paper does not quantify (Sec. 5.1). The load-bearing issue is not that the simulations are wrong, but that the experimentally measurable Acol has not been shown to be the same quantity as the calibrated Acol; without a full simulation through LumiCal selection and clustering, the claimed better-than-5% knowledge of the correction factor is not established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper quantifies the effect of the electromagnetic fields of the counter-rotating bunches on the low-angle Bhabha cross section used for the FCC-ee luminosity measurement. The authors derive Eq. (3.2), which converts the simulated focusing deflection at the acceptance edges into a luminosity bias; with the nominal Z-pole parameters this bias is about -0.19%, roughly twenty times the target precision. They validate the deflection calculation with two independent tools, Guinea-Pig and a Bassetti-Erskine numerical integration, study convergence with slice number and integration time, and show a linear correlation between the bias and both the initial-state px kick (Fig. 10) and an acollinearity variable Acol (Fig. 12). They propose two correction schemes: measuring the crossing-angle increase with dimuon events (Section 4) and measuring Acol with the LumiCal during the machine ramp-up or with pilot bunches (Section 5). The claimed outcome is a residual absolute luminosity uncertainty below 10^-4.","tokens_in":19115,"tokens_out":10137,"duration_ms":101616,"significance":"If the claimed residual precision is met, the paper solves a critical systematic for the FCC-ee Z-pole programme. The central calculation is credible: Eq. (3.2) follows from the 1/theta^3 counting integral and reproduces the quoted 0.19% bias with the stated deflections, and the agreement between two independent field calculations to better than about 10%, with documented convergence checks, gives confidence in the size and sign of the effect. The proposal of two experimentally measurable observables, the initial-state px kick and Acol, is a genuine step beyond a purely simulation-based correction. However, the headline claim that the correction can be determined with 1-2% relative accuracy, and hence with residual below 10^-4, is currently supported only by calibrations computed inside the same simulation framework and without a full LumiCal acceptance and clustering study for radiative Bhabha events. The required additional work is specific and, in my view, feasible; it does not undermine the main physics message but does affect the strength of the final claim.","major_comments":[{"comment":"The Acol calibration is built from generated-level, pre-LumiCal quantities: Section 3.3.3 states that the BHWIDE-based points are obtained from final-state charged leptons with only a 5 GeV cut, ignoring LumiCal clustering. The actual Acol is measured from events that pass the LumiCal asymmetric acceptance and cluster reconstruction. Because the total beam-induced deflection is strongly azimuth-dependent (about +150 microrad at phi=0 and -50 microrad at phi=pi for theta*=64 mrad, as shown in Fig. 11), the acceptance cut removes a phi-dependent subset of the event sample, so the mean Delta(theta+-) over the selected events is not the same as the mean over the generated theta* distribution used for Fig. 12. This selection distortion is not quantified. Since a 1% shift in Acol corresponds to roughly 2e-5 in the luminosity bias through the Fig. 12 mapping, the residual below 10^-4 claim requires either demonstrating that the distortion is negligible or including it in the calibration.","section":"Section 5.1, Eq. (5.1), Fig. 12"},{"comment":"The bracketing assumption that the true luminosity bias and Acol for radiative Bhabha events lie between the leading-order and the loose-cut BHWIDE determinations is stated as an expectation, not derived or checked. Cluster merging may compensate for collinear radiation, but clustering-dependent losses near the LumiCal inner and outer edges need not fall between those two extremes. Because both proposed correction methods use this assumption when converting a measured observable into a bias, the paper should either provide the LumiCal-level simulation with cluster reconstruction or explicitly include the unquantified bracket as a systematic uncertainty in the final residual budget.","section":"Section 3.3.3"},{"comment":"The ramp-up demonstration is a closure test rather than a validation: the text states that the fitted slope and intercept are equal to the input values by construction. Likewise, the +/-10^-4 bands shown in Figs. 10 and 12 are envelopes of simulated parameter variations, not estimates of the absolute accuracy of the Guinea-Pig/Bassetti-Erskine model. The claimed 1-2% accuracy of the correction therefore still depends on the unvalidated fidelity of the bunch-field model for the real machine. An externally anchored cross-check, or a quantitative model-dependence study, is needed before the residual below 10^-4 statement can be accepted.","section":"Section 5.2.1, Fig. 15"}],"minor_comments":[{"comment":"The fit annotation in Fig. 13 is garbled in the printed caption and axis label; please restate the fitted function explicitly and label the normalization and the exponent separately.","section":"Fig. 13"},{"comment":"The statement that a few hundreds of events suffice for a 5% relative Acol measurement should quote the assumed Bhabha rate and the polar-angle resolution explicitly, so the reader can check the statistical scaling.","section":"Section 5.1"},{"comment":"Please define Ebeam immediately before Eq. (4.1) and state whether it is the nominal beam energy or the energy including the kick; the sign convention for kx is also not specified.","section":"Section 4, Eq. (4.1)"},{"comment":"The description of the default Guinea-Pig settings is ambiguous: after describing one grid with 750 slices and seven grids with 300 slices, the phrase 'the latter setting' should be rephrased to identify the intended configuration explicitly.","section":"Section 3.1.1"},{"comment":"The fiducial angular range is given as 62-88 mrad and the narrow acceptance as 64-86 mrad; please state explicitly that these are the wide and narrow acceptances, respectively, to avoid confusion.","section":"Section 2"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the paper is a solid simulation study with a clear and correctly derived central bias formula, and the proposed correction methods are well motivated. My reservation is specifically that the headline precision claim (residual below 10^-4) is carried by in-model calibrations for an observable, Acol, that is not yet evaluated after LumiCal acceptance and clustering. This is a missing calculation rather than a wrong one, so I recommend major revision rather than rejection. The revised version should either supply the missing simulation or soften the final precision claim accordingly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, careful simulation study that quantifies a 0.19% beam-beam bias on FCC-ee luminosity at the Z pole and proposes two in-situ correction methods. The internal math holds up, the simulations are properly converged, and two independent field calculations agree to about 10%. The biggest open question is whether the Acol observable as reconstructed in the LumiCal matches the generator-level Acol used in the calibration curves.\n\nWhat's genuinely new: the crab-waist 30 mrad crossing angle introduces a strongly azimuthal final-state focusing, not present in the head-on ILC case. The paper works out the counting-rate bias from the 1/theta^3 integral, derives it cleanly, and shows the bias depends linearly on the initial-state px kick over a broad range of beam parameters. That correlation is the core of the proposed correction methods. The Acol variable is a measurable proxy for the kick, and the ramp-up extrapolation is a clever way to separate beam-induced and misalignment-induced contributions. The authors are honest about what they did not simulate.\n\nSoft spots: the Acol-to-bias calibration (Fig. 12) is built from leading-order and BHWIDE events with only a 5 GeV lepton cut, with LumiCal acceptance and clustering explicitly deferred. But the actual Acol is measured on events that survive the asymmetric acceptance, and the focusing itself changes which events survive, especially near theta_min where the phi=0 deflection is 150 urad. So the measured Acol may be a slightly different quantity than the calibrated one. A few percent distortion would still be within the 5% tolerance on the correction factor, but the paper does not quantify it. This is the main reason I can't call the residual <1e-4 claim established. The dimuon method relies on the 2% accuracy of delta_alpha from a companion paper by overlapping authors; that should be independently checked. These are fixable items, not fundamental flaws.\n\nWho it's for: anyone working on FCC-ee luminosity, forward Bhabha physics, or beam-beam effects at future e+e- machines. It deserves a serious referee. My recommendation: accept with requests for a full LumiCal simulation with clustering and acceptance selection, or an explicit argument that these effects are below the tolerance.","headline":"A careful, internally consistent simulation study that quantifies the beam-beam luminosity bias at FCC-ee and proposes two plausible correction methods, but the claimed sub-1e-4 residual rests on an Acol calibration that has not yet been shown to survive LumiCal acceptance and clustering.","tokens_in":19804,"tokens_out":3095,"would_cite":true,"duration_ms":32938,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"FCC-ee beam-beam fields will bias the luminosity measurement by about 0.19% at the Z pole, and the paper shows how two in-situ observables can bring the residual below 10^-4.","keywords":["beam-beam effects","luminosity measurement","Bhabha scattering","FCC-ee","LumiCal","acollinearity","Z pole","systematic uncertainty"],"falsifier":"A concrete check would be to measure $A_{\\rm col}$ and the Bhabha counting rate during the ramp-up at several bunch intensities and test whether the bias extracted from the simulated $A_{\\rm col}$-to-bias mapping agrees with the measured rate changes; a disagreement larger than the quoted uncertainties, or a visible curvature in the bias-versus-$p_x$ relation, would falsify the linear-calibration claim. A complementary simulation check would be to run the radiative Bhabha sample through a full LumiCal cluster reconstruction and verify that the resulting bias falls between the leading-order and loose-cut brackets used in the paper.","tokens_in":18459,"feed_emoji":"🎯","tokens_out":11122,"duration_ms":96332,"temperature":0.7,"pith_summary":"The paper argues that at FCC-ee the strong electromagnetic fields of the counter-rotating bunches focus the outgoing electrons and positrons of low-angle Bhabha scattering toward the beam axis, so fewer leptons enter the LumiCal acceptance than the Bhabha cross-section predicts. At the Z pole this acceptance loss biases the measured luminosity by about $0.19\\%$, roughly twenty times the $10^{-4}$ target precision, and the bias shifts when bunch parameters change during a fill. The paper claims that the bias is tightly and linearly correlated with two measurable observables: the transverse-momentum kick the initial-state particles receive from the opposite bunch (which effectively enlarges the crossing angle) and an acollinearity variable $A_{\\rm col}$ defined by the azimuthal modulation of the final-state lepton directions. It then proposes two ways to turn one of those observables into a correction: a dimuon-based measurement of the crossing-angle increase in the central detector, and an in-situ measurement of $A_{\\rm col}$ with the luminometer alone, extrapolated during the machine ramp-up. Each method determines the correction factor with $1$--$2\\%$ relative accuracy, leaving a residual absolute luminosity uncertainty below $10^{-4}$, which would remove a dominant systematic obstacle to the FCC-ee Z-line-shape programme.","feed_headline":"Beam fields skew FCC-ee luminosity by 0.19 percent","feed_subtitle":"Dimuon and acollinearity measurements can correct the bias down to below 10^-4 precision.","key_machinery":"The load-bearing object is the angular focusing $\\Delta\\theta_{\\rm FS}$ of a final-state Bhabha lepton in the field of the opposite-charge bunch, computed with a particle-tracking simulation and cross-checked by a numerical integration of the averaged Lorentz force from a Gaussian bunch. Equation (3.2) converts the deflection at the acceptance edges into the fractional luminosity bias, so the entire correction ultimately rests on how accurately $\\Delta\\theta_{\\rm FS}$ is known as a function of polar angle, azimuth, energy, and bunch parameters. The measurement side is carried by the acollinearity variable $A_{\\rm col}$, defined as the difference in the mean electron--positron acollinearity between the two azimuthal hemispheres; it is dominated by the initial-state $p_x$ kick and is mapped to the luminosity bias through calibration curves computed with the same simulation chain. A supporting relation is the power law $A_{\\rm col}\\propto N_m/\\sigma_m^a$ with $a\\simeq 0.72$, which lets ramp-up data be extrapolated to nominal conditions while separating the beam-induced component from a constant misalignment component.","core_discovery":"Stated on the paper's own terms, the central discovery is that beam-beam focusing reduces the LumiCal counting rate for Bhabha events at the Z pole by about $0.19\\%$, twenty times the $10^{-4}$ precision target, and that this bias can be corrected rather than merely estimated. The reduction is governed by Eq. (3.2), which translates the simulated angular deflection of final-state leptons at the acceptance edges into a relative rate change. The paper shows that the bias is a single-valued, essentially linear function of the initial-state $p_x$ kick and of the acollinearity observable $A_{\\rm col}$ across wide variations of bunch intensity, bunch length, transverse sizes, beam offsets, and vertical crossing-angle errors, with all simulated points inside a $\\pm 10^{-4}$ band around a straight line. It therefore concludes that a measurement of either observable---the dimuon crossing-angle increase in one case, $A_{\\rm col}$ in the other---determines the luminosity correction factor with a relative accuracy of $1$--$2\\%$, and that combining the methods leaves the absolute luminosity uncertainty below $10^{-4}$ and the point-to-point uncertainty on the Z-energy scan near $10^{-5}$.","pith_inferences":["Beyond the paper itself, the same $A_{\\rm col}$-versus-bias calibration could be validated early in commissioning by comparing the corrected luminosity from two independent acceptance settings or from a movable luminometer edge, giving a data-driven check before precision running starts.","The linear bias-versus-$p_x$ mapping suggests that the method transfers to the higher-energy FCC-ee stages, where beam parameters differ but the same ramp-up extrapolation can be repeated; the radiative-bracket uncertainty should be re-evaluated there.","An implication the paper leaves implicit is that $A_{\\rm col}$ measurements can double as a continuous alignment monitor: any measured component that does not scale with bunch intensity can be converted into an equivalent horizontal displacement of the luminometer, with a $5\\,\\mu$m scale at which the misalignment contribution becomes negligible.","A full simulation of radiative Bhabha events with LumiCal cluster reconstruction is the most direct way to close the gap between the leading-order and loose-cut brackets; the paper's residual-uncertainty claim depends on the true value lying inside that band."],"forward_implications":["The $0.19\\%$ acceptance bias would dominate the luminosity error if left uncorrected, so a working correction is a prerequisite for the Z-line-shape physics programme at FCC-ee.","A measurement of the crossing-angle increase from dimuon events, already needed for the centre-of-mass energy calibration, gives the luminosity correction factor to $2\\%$ relative accuracy.","The $A_{\\rm col}$ observable can be measured with the luminometer alone; a few hundred Bhabha events per azimuthal hemisphere already provide $5\\%$ statistical precision, allowing per-fill in-situ monitoring.","During the machine ramp-up, $A_{\\rm col}$ scales as $N_m/\\sigma_m^{0.72}$, so a linear fit separates the beam-induced component from a constant misalignment component and determines the beam-induced slope with $1$--$2\\%$ precision.","Pilot bunches at reduced intensity can perform the same separation in about ten minutes while losing less than $0.5\\%$ of the luminosity."],"supporting_citations":[{"why":"Supplies the FCC-ee CDR beam parameters, LumiCal geometry, and the 10^-4 luminosity precision targets used throughout.","marker":"[2]"},{"why":"Extends the Guinea-Pig code to track Bhabha events in the beam fields, the method adopted and adapted here.","marker":"[5]"},{"why":"Provides the Guinea-Pig beam-beam simulation used for all numerical event tracking.","marker":"[6]"},{"why":"Generates the radiative Bhabha events that bracket the luminosity bias when photon radiation and clustering effects are included.","marker":"[10]"},{"why":"Gives the closed expression for the field of a two-dimensional Gaussian bunch used in the numerical cross-check.","marker":"[11]"},{"why":"Establishes the px-kick-induced crossing-angle increase and the ramp-up extrapolation method that the dimuon correction relies on.","marker":"[15]"},{"why":"Simulates the LumiCal response and provides the polar-angle resolution used to estimate the statistical precision of Acol.","marker":"[16]"},{"why":"Supplies the clustering algorithm whose merging of electron and photon is the reason the paper brackets rather than fully simulates the radiative bias.","marker":"[17]"}],"fun_headline_variants":["Beam-beam skews FCC-ee luminosity 0.19%, corrected to <1e-4","FCC-ee luminosity beam-beam bias tamed with dimuon and acollinearity","Correcting 0.19% beam-beam luminosity shift at Z pole","Beam-beam luminosity error at FCC-ee reduced below 10^-4"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The correction strategy assumes that the simulated calibration curves---computed with a Gaussian-bunch model and nominal beam parameters, using radiative events that are bracketed but not fully cluster-simulated---correctly transfer to the real machine for every fill, and if that transfer fails the claimed residual below $10^{-4}$ would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Beam-beam skews FCC-ee luminosity 0.19%, corrected to <1e-4","FCC-ee luminosity beam-beam bias tamed with dimuon and acollinearity","Correcting 0.19% beam-beam luminosity shift at Z pole","Beam-beam luminosity error at FCC-ee reduced below 10^-4"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000315,"raw_usage":{"total_tokens":1821,"prompt_tokens":1014,"completion_tokens":807,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":630,"completion_tokens_details":{"reasoning_tokens":715}},"tokens_in":630,"tokens_out":807,"duration_ms":7005,"temperature":1.0,"reasoning_tokens":715,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:08:09.637899+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check would be to measure $A_{\\rm col}$ and the Bhabha counting rate during the ramp-up at several bunch intensities and test whether the bias extracted from the simulated $A_{\\rm col}$-to-bias mapping agrees with the measured rate changes; a disagreement larger than the quoted uncertainties, or a visible curvature in the bias-versus-$p_x$ relation, would falsify the linear-calibration claim. A complementary simulation check would be to run the radiative Bhabha sample through a full LumiCal cluster reconstruction and verify that the resulting bias falls between the leading-order and loose-cut brackets used in the paper.","supporting_citations":[{"cited_title":"Abada et al.,FCC-ee: The Lepton Collider, The European Physical Journal Special Topics228 (Jun, 2019) 261–623","cited_arxiv_id":null,"evidence_quote":"Supplies the FCC-ee CDR beam parameters, LumiCal geometry, and the 10^-4 luminosity precision targets used throughout."},{"cited_title":"Rimbault, P","cited_arxiv_id":null,"evidence_quote":"Extends the Guinea-Pig code to track Bhabha events in the beam fields, the method adopted and adapted here."},{"cited_title":"Closed Expression for the Electrical Field of a Two-dimensional Gaussian Charge","cited_arxiv_id":null,"evidence_quote":"Gives the closed expression for the field of a two-dimensional Gaussian bunch used in the numerical cross-check."},{"cited_title":"Agostinelli et al.,GEANT4: A Simulation toolkit, Nucl","cited_arxiv_id":null,"evidence_quote":"Simulates the LumiCal response and provides the polar-angle resolution used to estimate the statistical precision of Acol."}],"review_version":1}