{"id":"2696ad22-c392-4a92-97c1-38842ecc0cf0","arxiv_id":"1908.09062","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Using full simulation, the CEPC baseline detector reconstructs isolated photons with over 99% efficiency and purity above 1 GeV, reaching a 2.2% H to gamma gamma mass resolution after a geometry-based correction.","lead":"This paper simulates how well the proposed CEPC collider detector can find and measure photons. It reports high efficiency and purity for isolated photons and a 2.2% Higgs to two-photon mass resolution, with important caveats about how those numbers are derived.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 2.2% H→γγ mass resolution rests on an energy-dependent geometry correction calibrated only at 50 GeV, which the paper itself flags as needing an energy-dependent/iterative version; without a transfer check the headline number is not established.","rationale":"The central physics deliverable is the H→γγ mass resolution; it is the number that drives detector design and Higgs-coupling projections. That number is produced by a correction whose calibration-energy dependence is admitted explicitly in the text. The reader's weakest assumption is exactly this transferability issue, and I agree. The single-photon identification claim is well supported: Table 1 gives >99% efficiency and <1% misidentification for unconverted photons above 1 GeV with small statistical errors, and the conversion rate is consistent with the material budget. So I do not object to that part. The concern is narrow but load-bearing: one calibration sample at 50 GeV is used as a universal position-dependent scale, while Figure 10 shows the correction effect is energy dependent. Applying it to H→γγ without a cross-check leaves the 2.2% number conditional. A concrete energy-binned recalibration test would settle the issue. Secondary issues: the abstract's 'about 97% of π0 in Z→ττ can be reconstructed successfully' overreaches, since the 30 GeV critical energy corresponds to 50% reconstruction efficiency, not 100%; and Figure 9's 'roughly 1.5 million events' caption conflicts with the 10k-per-energy-point statement in §2.3. These reinforce a conditional verdict but do not change the central concern. Since the reader already issued CONDITIONAL for the same reason, my recommendation is UNCHANGED.","tokens_in":8535,"tokens_out":7320,"duration_ms":73428,"concrete_test":"Recompute the H→γγ mass resolution after replacing the single 50 GeV scale-factor map with an energy-interpolated correction: derive E_true/E_deposited maps from the existing single-photon samples at 10, 20, 30, 50, 75 and 100 GeV; for each H→γγ photon, apply the map for the nearest energy bin or a log-linear interpolation, then refit the invariant mass peak. If the resolution remains 2.2% within the statistical uncertainty of roughly 6k events (about ±0.1 percentage point) and the peak position stays at m_H, the 50 GeV-only calibration is validated. If the width moves toward the uncorrected 2.6% or the mean shifts by more than about 1%, the 2.2% headline is an artifact of the calibration choice and should be re-quoted as 2.6% pending an energy-dependent correction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In §3.3, the geometry-based energy correction uses a scale factor E'_true/E'_deposited obtained from simulated 50 GeV single photons. The same position-dependent map is then applied to H→γγ photons in §4.1 to obtain the quoted 2.2% mass resolution. The paper itself states that 'because the input sample is chosen at 50 GeV, correction at high energy is more significant' (Fig. 10) and that 'an iterative correction algorithm shall be developed in the future' (§5). For H→γγ at √s=240 GeV, the two photons have a broad energy spectrum rather than a single energy near 50 GeV. If the crack-loss fraction varies with energy, the 50 GeV map can mis-scale at least one photon, biasing the invariant mass and distorting the resolution. The quoted 2.2% is therefore not a robust statement about the baseline detector; the uncorrected 2.6% is the conservative number actually demonstrated. The 1.7% simplified-geometry value is unaffected because it does not use this correction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper characterizes photon reconstruction performance at the CEPC baseline detector using full Geant4 simulation and the Arbor particle-flow reconstruction. The authors report conversion rates, reconstruction and identification efficiencies for single photons, an energy-resolution comparison between the baseline detector and a simplified defect-free geometry, a position-dependent energy correction tested on 50 GeV photons, the H→γγ invariant-mass resolution, and π0 reconstruction efficiency as a function of π0 energy. They conclude that isolated unconverted photons above 1 GeV are identified with efficiency above 99% and misidentification below 1%, that the H→γγ mass resolution is 2.2% at the baseline detector (1.7% at the simplified geometry), and that about 97% of π0 from Z→ττ events can be reconstructed.","tokens_in":8725,"tokens_out":3758,"duration_ms":37172,"significance":"If the headline results hold, the paper provides useful quantitative benchmarks for the CEPC detector design: it gives a full-simulation estimate of the photon identification performance, quantifies the effect of ECAL geometry cracks and dead zones on the energy resolution, and demonstrates a position-based correction as a proof of principle. The separation of the simplified-geometry resolution from the baseline-detector resolution, and the comparison with CALICE test-beam results, are helpful for detector optimization. The 1.7% simplified-geometry H→γγ mass resolution is a clean, directly simulated result, and the single-photon efficiency and misidentification tables are internally consistent and reproducible in structure.","major_comments":[{"comment":"The quoted 2.2% H→γγ mass resolution rests on a geometry-correction scale factor calibrated exclusively on 50 GeV single photons and then applied to the broad photon energy spectrum of Higgs decays. The paper itself states in §3.3 and Fig. 10 that the correction is energy-dependent and that an iterative correction algorithm is deferred to future work. Because the transfer of the 50 GeV calibration to the H→γγ sample is not validated, the 2.2% number is not an established property of the baseline detector; the uncorrected 2.6% and the simplified-geometry 1.7% values are the directly demonstrated results. The abstract and conclusion should either report the 2.2% explicitly as a test-of-principle result, or the authors should add a transfer check, for example by comparing the corrected mass resolution when the calibration map is derived in several energy bins.","section":"§3.3 and §4.1, Eq. (1), Fig. 10"},{"comment":"The claim that about 97% of π0 generated in Z→ττ events can be reconstructed successfully is an inference from the energy spectrum and a single 30 GeV 'critical energy' threshold, not a direct measurement of π0 reconstruction efficiency on Z→ττ events. Figure 12 shows a smooth, energy-dependent reconstruction efficiency, so the assumption that all π0 below the critical energy are reconstructed and none above are reconstructed is not justified, and no uncertainty is propagated to the 97% figure. The abstract and conclusions should be rephrased to state that only about 3% of Z→ττ π0 have energy above the 30 GeV critical energy, which is a threshold-based estimate rather than a direct reconstruction efficiency.","section":"§4.2 and Abstract"},{"comment":"The identification-efficiency statement in the conclusion, namely 'efficiency higher than 99% and misidentification rate smaller than 1% for unconverted, isolated photons with energies larger than 1 GeV', is only demonstrated when the ToF-based separation is assumed and the cluster-level photon selection is applied. Table 1 itself shows that for E > 10 GeV the neutron misidentification rate is 0.83%, which is below 1%, but the systematic uncertainties from the assumed 50 ps time resolution and from the simplified L=2.4 m flight-distance assumption are not quantified. The authors should state these assumptions in the conclusion or add a systematic study of the ToF parameters.","section":"§3.2, Table 1"}],"minor_comments":[{"comment":"There are several typographical and grammatical issues: 'Of cause' should be 'Of course', and 'is consist with' should be 'is consistent with'.","section":"§3.2"},{"comment":"In the correction equation, the scale factor E'_true/E'_deposited is written as a scalar, but it is actually a position-dependent map defined from 50 GeV photons; please state explicitly that the scale factor is a function of the reconstructed hit position and define how the map is binned and interpolated.","section":"§3.3, Eq. (1)"},{"comment":"The energy-resolution parameterization is garbled: '17.7%√E ⊕ 2.1%/ 15.2%√E ⊕1.2%' should be written with explicit denominators, for example '17.7%/sqrt(E) ⊕ 2.1%' and '15.2%/sqrt(E) ⊕ 1.2%', and the applicable energy range and fit uncertainties should be given.","section":"§5, Conclusion"},{"comment":"The abstract reports '1.7 - 2.2%' for the Higgs mass resolution and '97%' for π0 reconstruction without mentioning that the mass resolution is evaluated in the central region and that the π0 figure is a threshold-based estimate; adding these qualifications would improve accuracy.","section":"§4.2 and Abstract"},{"comment":"The caption of Fig. 6 says 'Between 200 MeV and 500 GeV', which should be '500 MeV'; also, please define the coordinate ranges used for 'central region' and 'forward region' consistently in §3.1 and §4.1.","section":"§2.1 and Fig. 6"},{"comment":"Reference [22] appears as '[21][22]][23]' with a stray bracket; please correct the citation formatting.","section":"§3.2, references"}],"recommendation":"major_revision","confidential_remarks":"The paper contains useful simulation results, but the abstract and conclusion overstate two headline numbers: the 2.2% H→γγ mass resolution depends on a 50 GeV-calibrated correction whose energy transfer is not demonstrated, and the 97% π0 reconstruction claim is inferred from a threshold rather than measured. These are fixable within the scope of the paper by adding transfer checks or by clearly labeling the numbers as test-of-principle estimates. The simplified-geometry 1.7% result and the single-photon identification efficiencies are more robust and should be emphasized."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a solid, first dedicated photon reconstruction paper for the CEPC baseline detector. It gives conversion rates, single-photon efficiencies, identification performance, energy resolution, and a pi0 separation benchmark from full Geant4 simulation. That is genuinely new and worth having. Don't desk reject it. But don't quote the abstract's numbers without reading the caveats; two of them are over-reached.\n\nWhat is good: the single-photon results are internally consistent and plausible. The efficiency map as a function of theta and energy shows the cracks and dead zones, the identification with cluster shape plus ToF gives >99% photon efficiency with <1% mis-ID above 1 GeV, and the comparison between the baseline and a defect-free simplified geometry gives a clean way to see what geometry defects cost. The energy-dependent degradation and the correction test on 50 GeV photons are a reasonable proof-of-principle.\n\nWhere it gets soft: the stress-test concern is right. The 2.2% H→γγ mass resolution is obtained by applying a position-dependent correction factor calibrated exclusively at 50 GeV to photons over a wide energy spectrum. The paper itself says the correction is energy-dependent and that an iterative version is needed. So 2.2% is not a robust statement about the baseline detector; 2.6% uncorrected is the number actually demonstrated. The 1.7% simplified-geometry value is fine because no correction is involved.\n\nThe other over-reach is the pi0 claim. The abstract says 97% of pi0s in Z→ττ are reconstructed successfully. In the text, that's inferred from the fact that 97% of them are below an adopted 30 GeV critical energy, not from a direct reconstruction efficiency applied to those events. The paper even says a dedicated pi0 finder could improve things. That's an energy-threshold statement, not a measured efficiency.\n\nMinor issues: energy resolutions and mass resolutions are quoted without statistical uncertainties. There are a few typos (\"consist with\" rather than \"consistent with\"). The CEPC CDR and earlier ECAL optimization are cited heavily; the new numbers are independently simulated, so the citation pattern is fine.\n\nWho should read this: anyone working on CEPC detector optimization, PFA calorimetry, or future lepton colliders. It deserves a serious referee. My recommendation is to send it to review with the condition that the authors either demonstrate the correction transfers across energy or moderate the headline claims, and add numbers with uncertainties. The core single-photon characterization is the real contribution and it holds up.","headline":"Useful first full-simulation photon performance study for CEPC, but the headline 2.2% H→γγ resolution and the 97% pi0 claim are both softer than they look.","tokens_in":9256,"tokens_out":2212,"would_cite":true,"duration_ms":21982,"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":"For the CEPC baseline detector, full simulation shows isolated photons above 1 GeV are identified with over 99% efficiency and less than 1% misidentification, and the Higgs boson mass in H→γγ is reconstructed to 2.2% after a…","keywords":["photon reconstruction","electromagnetic calorimeter","particle flow detector","Higgs to diphoton","mass resolution","pi0 reconstruction","CEPC"],"falsifier":"Train the correction algorithm on 10 GeV and 100 GeV photon samples instead of 50 GeV and recompute the H→γγ invariant-mass resolution; if the 2.2% changes by more than the fit uncertainty, the single-energy calibration does not transfer and the quoted resolution depends on the training choice.","tokens_in":8321,"feed_emoji":"🎯","tokens_out":7094,"duration_ms":68436,"temperature":0.7,"pith_summary":"This paper establishes benchmark photon-reconstruction performance for the baseline detector design of the proposed Circular Electron Positron Collider, a particle-flow oriented detector with a sampling electromagnetic calorimeter. Using full simulation of single photons and diphoton events, it shows that isolated, unconverted photons above 1 GeV are identified with efficiency above 99% and misidentification below 1%, and that a position-based energy correction recovers most of the resolution lost to gaps and dead zones. At the H→γγ sample the corrected invariant-mass resolution is 2.2%, compared with 1.7% for an idealized defect-free geometry. The same reconstruction separates the two photons from π0 decays up to roughly 30 GeV, covering most π0s produced in Z→ττ events. If these numbers hold, the baseline detector meets the photon performance required for the collider's Higgs and electroweak program.","feed_headline":"CEPC photons hit 99% purity; Higgs diphoton mass 2.2%","feed_subtitle":"A full-simulation study shows the detector also reconstructs most π0s from Z→ττ.","key_machinery":"The analysis rests on three linked mechanisms: the particle-flow reconstruction that builds calorimeter clusters and separates nearby showers; a photon identifier that combines shower-shape variables with a 50 ps time-of-flight cut to reject neutral hadrons; and a geometry-based energy correction that scales each electromagnetic cluster by a position-dependent factor measured from simulated 50 GeV photons, compensating for energy lost in module and stave cracks and in the barrel–endcap dead zone. The H→γγ invariant mass then serves as the test bench for both energy resolution and two-shower separation.","core_discovery":"On its own terms, the paper claims that the CEPC baseline detector—a particle-flow oriented design with a silicon-tungsten sampling calorimeter split into staves, modules, and columns—reconstructs photons well enough for a Higgs factory. Unconverted, isolated photons above 1 GeV are identified with >99% efficiency and <1% misidentification; cluster reconstruction is near 100% for photons above 200 MeV and about 85% at 100 MeV; and roughly 6–10% of central-region photons and 25% of forward photons convert to e+e− before reaching the calorimeter. The energy resolution is about 17.7%/√E ⊕ 2.1% at the baseline detector versus 15.2%/√E ⊕ 1.2% at a simplified defect-free geometry, and a correction calibrated on 50 GeV photons removes the low-energy tail in H→γγ events so the Higgs mass resolution reaches 2.2% (1.7% ideal). The same clustering separates π0→γγ up to critical energies of 22 GeV in the barrel and 34 GeV in the endcap, so about 97% of π0s from Z→ττ are reconstructed.","pith_inferences":["An energy-dependent calibration is the obvious next step: training the correction on several photon energies and iterating would probably push the baseline H→γγ resolution from 2.2% toward the 1.7% defect-free value.","The same position-based crack correction could be transferred to any modular sampling calorimeter, not just the specific CEPC stave geometry, so the method has general value for future detectors.","The 99%/1% identification numbers apply to isolated unconverted photons; counting merged-π0 backgrounds inside jets would test where the practical purity ceiling sits for CEPC's Higgs measurements."],"forward_implications":["Isolated unconverted photons above 1 GeV can be tagged at >99% efficiency with <1% neutron contamination, making the CEPC photon sample clean enough for precision Higgs and electroweak measurements.","With the geometry correction, the Higgs mass from H→γγ is reconstructed at 2.2% relative resolution, establishing the baseline configuration as a viable starting point for the physics program.","The correction reduces the energy-resolution degradation from geometry defects from about 35% to about 10% for 40–100 GeV photons, so the main cost of a segmented calorimeter is largely recoverable.","About 97% of π0s produced in Z→ττ are reconstructed, so tau identification via decay photons is not blocked by photon merging.","Roughly 15% of π0s in inclusive Higgs events lie above the 30 GeV frontier, motivating a dedicated π0 finder beyond the generic particle-flow reconstruction."],"supporting_citations":[{"why":"Defines the baseline detector geometry, material budget, and ECAL layout used in all simulations.","marker":"[9]"},{"why":"Supplies the particle-flow reconstruction algorithm that builds neutral clusters and separates showers.","marker":"[17]"},{"why":"Provide the shower-shape discriminating variables adopted for photon identification.","marker":"[19][20]"},{"why":"Sets the 50 ps time-of-flight benchmark used to reject low-energy neutral hadrons.","marker":"[21]"},{"why":"Establishes the calorimeter energy-resolution expectations against which the simplified geometry is checked.","marker":"[24]"},{"why":"Provides the ECAL two-shower separation study that sets the π0 critical-energy comparison.","marker":"[25]"}],"fun_headline_variants":["CEPC photon ID >99% pure, Higgs mass to 2.2%","CEPC photons: 99% purity, Higgs diphoton 2.2% mass","CEPC pi0 reconstruction up to 22–34 GeV in barrel/endcap","CEPC baseline: 1.7–2.2% Higgs mass after photon correction","CEPC photon conversion up to 25% forward, still 2.2% Higgs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The position-based energy correction is calibrated on 50 GeV photons and is assumed to hold for the much wider energy spectrum of photons in H→γγ events.","fun_headline_variants_meta":{"raw":{"variants":["CEPC photon ID >99% pure, Higgs mass to 2.2%","CEPC photons: 99% purity, Higgs diphoton 2.2% mass","CEPC pi0 reconstruction up to 22–34 GeV in barrel/endcap","CEPC baseline: 1.7–2.2% Higgs mass after photon correction","CEPC photon conversion up to 25% forward, still 2.2% Higgs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000833,"raw_usage":{"total_tokens":3637,"prompt_tokens":950,"completion_tokens":2687,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":566,"completion_tokens_details":{"reasoning_tokens":2572}},"tokens_in":566,"tokens_out":2687,"duration_ms":17428,"temperature":1.0,"reasoning_tokens":2572,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:23:02.991774+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Train the correction algorithm on 10 GeV and 100 GeV photon samples instead of 50 GeV and recompute the H→γγ invariant-mass resolution; if the 2.2% changes by more than the fit uncertainty, the single-energy calibration does not transfer and the quoted resolution depends on the training choice.","supporting_citations":[{"cited_title":"The CMS HGCAL detector for HL-LHC upgrade","cited_arxiv_id":"1708.08234","evidence_quote":"Sets the 50 ps time-of-flight benchmark used to reject low-energy neutral hadrons."},{"cited_title":"PFA Oriented ECAL Optimization for the CEPC","cited_arxiv_id":"1712.09625","evidence_quote":"Provides the ECAL two-shower separation study that sets the π0 critical-energy comparison."}],"review_version":1}