{"id":"e8e76113-6292-45a2-b433-a778a7194c92","arxiv_id":"2411.09345","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The DarkSHINE baseline design projects 90% C.L. sensitivity to dark photon kinetic mixing epsilon^2 as low as ~1e-12 for masses around 1-100 MeV with 3e14 electrons on target.","lead":"DarkSHINE is a proposed fixed-target experiment to search for invisibly decaying dark photons at the SHINE accelerator. The report presents a detector design and projected sensitivity that could improve existing dark photon limits by up to two orders of magnitude.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Exposure inconsistency: the stated 1 MHz single-electron beam yields ~3×10^13 EOT/year, making the 3×10^14 baseline and the one/three/five-year labels off by roughly an order of magnitude.","rationale":"The reader's conditional verdict focuses on the un-reproduced background extrapolation from ref. [22] and on the 1 MHz/3×10^14 EOT assumption. I agree the background estimate is opaque, but it is not the most load-bearing element: Eq. (3.1) gives nearly the same 90% CL signal limit for b = 0.015 and b = 1 (about 2.31 vs 2.89 signal events), and even b = 10 only worsens the limit by a factor of about 2.3. The exposure L, by contrast, enters Eq. (3.2) directly, and the stated beam parameters are internally inconsistent. Sec. 2.1 quotes ~10^13 EOT/year for a 1 MHz single-electron beam, while the conclusion maps 3×10^14, 9×10^14, and 1.5×10^15 to one, three, and five years. Simple arithmetic shows these labels are off by a factor of about 10. If L is really 3×10^13 EOT/year, the ε² limit worsens by sqrt(10), weakening the 'nearly two orders of magnitude' claim. If L = 3×10^14/year is intended, the beam must contain ~10 electrons per bunch, invalidating the single-particle simulation assumption. This is a concrete internal tension that can be settled by checking the facility beam-delivery parameters. The design remains plausible if the exposure is correctly relabeled, so the reader's CONDITIONAL verdict is unchanged.","tokens_in":43884,"tokens_out":11260,"duration_ms":109487,"concrete_test":"Check the SHINE DarkSHINE beam-delivery plan: compute the integrated EOT per year from the stated 1 MHz single-electron rate (1e6 e−/s × 3.15e7 s ≈ 3.15×10^13 EOT/year) and compare with the 3×10^14 EOT baseline and the one/three/five-year labels in the conclusion; if the facility cannot deliver ~10 electrons per bunch without violating the single-electron assumption, the exposure normalization and the projected sensitivity curves must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The projected sensitivity in Eq. (3.2) is normalized to L = 3×10^14 EOT, and the conclusion labels 3×10^14, 9×10^14, and 1.5×10^15 as one, three, and five years of running. However, Sec. 2.1 states that DarkSHINE uses a 1 MHz single-electron beam and estimates ~10^13 EOT/year. One year at 1 MHz with one electron per bunch gives 3.15×10^13 EOT, so 3×10^14 corresponds to ~9.5 years (or ~30 years using the report's own 10^13/year estimate), not one year. If the first run delivers only ~3×10^13 EOT, the ε² limit in Fig. 3.9 worsens by roughly sqrt(10) ≈ 3.2, and the headline 'nearly two orders of magnitude' improvement over existing experiments is reduced to about one order. Conversely, if 3×10^14 EOT per year is really achievable at 1 MHz, each bunch must carry about 10 electrons on average, which contradicts the single-electron assumption used throughout the tracker, ECAL, and background simulations. The report does not resolve this tension: either the exposure must be relabeled to a multi-year program, or the simulations must be redone with realistic per-bunch electron multiplicity and pileup.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript is the baseline design report of the DarkSHINE experiment, a proposed fixed-target search for invisibly decaying dark photons at the SHINE 8 GeV electron facility. It specifies a detector consisting of a tagging and a recoil silicon-strip tracker, a 350-micron tungsten target, a homogeneous 20x20x11 LYSO electromagnetic calorimeter, and a large iron/scintillator hadronic calorimeter, together with the associated magnets, mechanics, and readout electronics. The physics studies use CalcHEP-generated dark-bremsstrahlung events and Geant4 simulations of inclusive and rare Standard Model backgrounds, supplemented by prototype measurements of AC-LGAD strip sensors, LYSO/SiPM modules, HCAL scintillator strips, and 1 MHz readout electronics. The central quantitative results are a background estimate of 0.015 events per 3e14 electrons on target and projected 90% C.L. exclusion limits on the kinetic mixing parameter epsilon^2 versus dark-photon mass, claimed to improve on current experiments by nearly two orders of magnitude, together with a thermal-relic dark-matter projection in the MeV range.","tokens_in":44148,"tokens_out":16422,"duration_ms":158303,"significance":"If the projected sensitivity were realized, DarkSHINE would provide an independent and competitive probe of sub-GeV invisible dark photons, complementary to NA64 and LDMX, and could reach thermal-relic dark-matter parameter space in the MeV region. The report's strengths are its detailed full-detector Geant4 simulation, biased rare-process samples with quoted effective exposures, and a substantial set of small-scale prototype measurements: AC-LGAD spatial resolutions of 6.5-12.3 microns, DESY beam-test data on LYSO/SiPM modules, electronics tests at 1 MHz, and HCAL neutron-veto studies. These elements make the detector concept credible as a baseline. However, the quantitative sensitivity claim inherits load-bearing inconsistencies: the one-year exposure label disagrees with the stated beam rate by an order of magnitude, the beam energy differs between sections, and the central background yield is not reproduced in this report and appears to omit the neutrino-production backgrounds estimated in Table 2.3. The physics case is plausible, but the headline projection should be treated as preliminary until these points are corrected or documented.","major_comments":[{"comment":"The beam energy is stated inconsistently. Section 2.4 says DarkSHINE uses a 4 GeV electron beam and that the tagging tracker must track 4 GeV incoming electrons, while Sections 2.5, 3.2, and 3.3 and Fig. 3.10 use an 8 GeV beam, and Table 2.3 is labeled with Ebeam = 10 GeV. The dark-photon production cross section, recoil kinematics, tracker momentum acceptance, and calorimeter containment all depend on the beam energy, so these values cannot all be correct simultaneously. The report should state a single baseline beam energy and verify that the tracker performance numbers in Figs. 2.6-2.8 and the signal efficiencies in Fig. 3.7 are computed with it.","section":"Sec. 2.4, Sec. 2.5, Sec. 3.2, Sec. 3.3, Table 2.3"},{"comment":"The exposure per year is off by an order of magnitude. Section 2.1 describes a 1 MHz single-electron beam and estimates about 1e13 EOT per year; at 1 MHz continuous operation, one year is 3.15e13 electrons. Yet Section 2.5 says 'after the first run DarkSHINE can collect 3e14 electron-on-target', and the Conclusion labels 3e14, 9e14, and 1.5e15 as one, three, and five years of running, while also listing 1e16 EOT without a time label. Equation (3.2) and Fig. 3.9 normalize the sensitivity to L = 3e14 EOT. If the real one-year exposure is about 3e13, the projected epsilon^2 limits worsen by roughly sqrt(10) about 3.2; if 3e14 per year is intended, the beam must carry about 10 electrons per bunch on average, contradicting the single-electron assumption used in the tracker, ECAL, and background simulations. The beam rate, per-bunch multiplicity, and exposure labels must be reconciled.","section":"Sec. 2.1, Sec. 2.5, Sec. 3.3, Conclusion"},{"comment":"The neutrino-production backgrounds estimated in Section 2.5 are not included in the background count. Table 2.3 reports 3.6e-4 events (Moller + CCQE) and 0.3-0.6 events (CCQE with exclusive pi0) per 3e14 EOT for the tungsten target, and the text says such real-missing-energy backgrounds are 'an important component'. Section 3.2 then states that 'neutrino production reactions can be ignored for now', and Table 3.3 contains no samples for these processes. The extrapolated total of 0.015 events therefore appears to exclude the CCQE-with-pi0 contribution, which is 20-40 times larger than 0.015 before any selection. If these events are rejected by the one-track, ECAL, or HCAL cuts, the survival fractions must be demonstrated; if they are not rejected, the background estimate must be revised. The use of Ebeam = 10 GeV in Table 2.3 should also be reconciled with the 8 GeV baseline.","section":"Sec. 2.5, Table 2.3, Sec. 3.2, Table 3.3"},{"comment":"The central background estimate of 0.015 events per 3e14 EOT is not supported within this manuscript. Section 3.2 states that after the listed cuts 'no background survives' and then obtains 0.015 by 'an extrapolation method [22]', but the per-process surviving counts, effective exposures, extrapolation procedure, and uncertainties are not given. Reference [22] is also the source for the ECAL signal-region definition in Section 2.6.1. Since this number enters directly into the expected-limit formula, Eq. (3.1)-(3.2), and since only the GMM-target sample has effective EOT (4.3e14) larger than the 3e14 baseline, the central result is not independently checkable from this report. The authors should present the extrapolation in detail, state whether 0.015 is a best estimate or an upper limit, and show how the sensitivity changes for plausible values of the background, for example b = 0, 0.015, 0.1, and 1.","section":"Sec. 3.2, Eq. (3.1)-(3.2), Sec. 2.6.1"}],"minor_comments":[{"comment":"The quantity 'effective EOT' is not defined; the report should state how biasing weights convert generated events to effective electrons-on-target and how statistical uncertainties on the extrapolated yields are assessed.","section":"Sec. 3.2, Table 3.3"},{"comment":"The signal-efficiency axis extends above 1.0, which conflicts with the text's statement that the overall acceptance is 'over 60%'; please clarify whether the plotted quantity is a cumulative efficiency or includes an additional normalization.","section":"Fig. 3.7"},{"comment":"The ECAL smearing parameters A, B, C are extracted from a single-crystal optical simulation and applied to the full 20x20x11 array; the report should state explicitly that the full-array response has not yet been validated by beam data and should add this validation to the R&D roadmap.","section":"Sec. 2.6.2-2.6.3, Table 2.5"},{"comment":"A proofreading pass is needed: examples include 'statitics' in the Conclusion, 'F uture Plan' in the table of contents, 'electromagetic' in Section 2.2, 'ralative rate' in Section 3.2, and 'magentic' in the Fig. 3.1 caption.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"To the editor: this is a credible baseline design report, but the central sensitivity projection depends on several numbers that are internally inconsistent or delegated to the collaboration's earlier paper [22]. I am not asking the authors to build a full prototype before publication, but I would require them to harmonize the beam energy and exposure numbers and to include the background-extrapolation details in the paper or an appendix. The spread of 4/8/10 GeV across sections and the mismatch between the 1 MHz single-electron rate and the quoted one-year exposure suggest that the text has not been fully harmonized; a careful consistency pass is advisable before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a serious baseline design report for a fixed-target dark photon search at SHINE, and it earns its place as a reference for the proposed experiment. The genuinely new pieces are the facility-specific application and the concrete detector choices—AC-LGAD strip tracker, LYSO crystal ECAL with digitization and saturation modeling, large sampling HCAL as neutron/muon veto—plus real prototype work: DESY beam tests on a crystal module, SiPM dynamic range measurements, and radiation dose maps. The projected sensitivity curves are useful for planning, and if 3e14 EOT is actually collectable, the reach is competitive with LDMX Phase 1 and NA64 and touches thermal-relic MeV dark matter. Credit where due: the collaboration reports prototype measurements rather than only simulation, and the simulation chain is described in enough detail to be reproducible in principle.\n\nNow the soft spots. The exposure bookkeeping is off by about an order of magnitude. Section 2.1 says the 1 MHz single-electron beam gives ~1e13 EOT/year. One year of 1 MHz single electrons is ~3.1e13. Yet the baseline sensitivity uses L=3e14 and the conclusion labels 3e14, 9e14, and 1.5e15 as one, three, and five years. That cannot all be true. If the real first-run exposure is ~3e13, the epsilon^2 limit gets worse by sqrt(10) ~ 3, and the \"nearly two orders of magnitude\" improvement shrinks to roughly one order. This is fixable—relabel the curves as multi-year totals or explain how the rate is actually higher—but it changes how the headline result should be read.\n\nThe second soft spot is the background estimate. The 0.015 events per 3e14 is inherited from the self-cited ref [22] via an extrapolation, and the report does not reproduce the method or give enough external detail to check it. Combined with the ECAL smearing being fitted to the collaboration's own single-crystal simulation, the sensitivity projection is a design target, not a quasi-measurement. That is normal for a design report, but the background extrapolation is load-bearing enough that it should be documented in the report itself or made publicly reproducible before the projection is relied on.\n\nMinor issues: typos, a few sloppy figure labels, and the write-up is long. None of that changes the substance. Who this is for: the DarkSHINE collaboration, SHINE planners, and anyone comparing proposed fixed-target dark photon experiments. It deserves a serious referee, not a desk reject. I would send it to review and ask that the exposure inconsistency be resolved and that the background extrapolation be either fully described or cited with the key numbers reproduced. That is a major revision, but the design work and prototype results justify the effort.","headline":"A credible, detailed baseline design for a fixed-target dark photon search at SHINE, but the exposure numbers do not add up—3e14 EOT is not one year at 1 MHz single-electron—and the background estimate rests on an under-documented extrapolation.","tokens_in":44871,"tokens_out":3814,"would_cite":true,"duration_ms":40006,"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":"This paper makes the case that a new fixed-target electron-beam experiment can search for invisibly decaying dark photons with almost no background, projecting sensitivity roughly two orders of magnitude beyond current limits.","keywords":["dark photon","invisible decay","fixed-target experiment","missing energy signature","kinetic mixing parameter","dark bremsstrahlung","LYSO calorimeter","hadronic calorimeter veto"],"falsifier":"Take data with the full detector and beam on a 0.35 mm tungsten target for the equivalent of 3×$10^{14}$ electrons on target and count events passing the five selection cuts: one track, missing momentum above 4 GeV, ECAL total energy below 2.5 GeV, HCAL total energy below 0.1 GeV, and HCAL maximum cell energy below 2 MeV. If more than about one event is observed, the 0.015 background estimate is falsified and the projected ε² exclusion curve does not hold at face value.","tokens_in":43657,"feed_emoji":"⚛️","tokens_out":6357,"duration_ms":74116,"temperature":0.7,"pith_summary":"The paper lays out a baseline design for DarkSHINE, a proposed fixed-target experiment that would use the SHINE facility's 1 MHz, 8 GeV single-electron beam to search for dark photons that decay invisibly into dark matter. The experimental signature is a single recoiling electron carrying large missing momentum and little energy in the calorimeters, produced by dark bremsstrahlung off a thin tungsten target. Simulating the full detector, the authors estimate 0.015 background events for 3×$10^{14}$ electrons on target after all selection cuts, and derive a projected 90% confidence exclusion limit on the kinetic-mixing parameter ε² as a function of dark photon mass that improves on current experiments by nearly two orders of magnitude. If the projection holds, the setup could also probe thermally produced dark matter in the MeV mass range.","feed_headline":"DarkSHINE design targets dark photons 100x beyond current limits","feed_subtitle":"Missing-energy search at a 1 MHz electron beam projects 0.015 background events and MeV-scale dark matter reach.","key_machinery":"The mechanism that carries the argument is the missing-momentum signature combined with a layered veto. Dark bremsstrahlung $e^-Z \\to e^-Z A'$ with the dark photon $A'$ decaying invisibly leaves one low-energy, high-transverse-momentum electron; backgrounds leave extra particles and more calorimeter energy. The paper's selection requires exactly one track, tagging-minus-recoil momentum greater than 4 GeV, ECAL total energy below 2.5 GeV, and HCAL total and maximum-cell energy below 0.1 GeV and 2 MeV. The LYSO crystal ECAL gives better than 2% statistical energy resolution and the 4×4 m² sampling HCAL provides a neutron veto with inefficiency below $10^{-5}$ for energetic neutrons, together reducing the estimated background to 0.015 events in the nominal dataset.","core_discovery":"The central claim is that an experiment built around a high-repetition-rate electron beam can detect the invisible decay of dark photons through a missing-energy and missing-momentum signature with essentially background-free sensitivity. DarkSHINE would use a 0.35 mm tungsten target and measure the incoming and outgoing electron with silicon strip trackers in a 1.5 T dipole field, then require a single recoil track, missing momentum above 4 GeV, ECAL energy below 2.5 GeV, and HCAL energy below 0.1 GeV. With these selections no simulated background event survives; an extrapolation procedure yields 0.015 expected background events per 3×$10^{14}$ electrons on target. The paper derives a 90% C.L. exclusion curve on ε² versus dark photon mass for 3×$10^{14}$ to $10^{16}$ electrons on target, claiming sensitivity nearly two orders of magnitude better than current experiments and reach into the MeV-scale thermal relic dark matter parameter space.","pith_inferences":["If the background extrapolation is the weakest link, a dedicated control run with an empty target and a beam-off-target sample would directly calibrate the rare-process rates before the physics run; the report does not describe such a control.","The same detector, with vertex reconstruction added, could search for visible dark photon decays and other long-lived signatures, extending the physics case beyond invisible decays.","Because the projected reach hinges on the 1 MHz single-electron beam and 3×10^14 EOT per year, a near-term prototype test of beam delivery and triggerless readout on the SHINE linac would be the fastest way to de-risk the sensitivity claim."],"forward_implications":["If the projected sensitivity is correct, the experiment can set the strongest laboratory limits on invisibly decaying dark photons in the roughly 1 MeV to 1 GeV mass range.","The reach in the dark-matter interaction-strength plane would cover benchmark thermal-relic dark matter models for dark matter masses around the MeV scale.","At 3×10^14 electrons on target the sensitivity is comparable to the LDMX Phase-1 design; at 10^16 EOT it complements LDMX, doing better at low mass because of the precise LYSO calorimeter.","The staggered ECAL layout studied in the report would reduce inclusive backgrounds entering the signal region by more than an order of magnitude relative to the uniform layout.","A 1.5 m × 1.5 m HCAL transverse size meets the low-energy-neutron veto requirement while reducing weight, an option for the real hall constraints."],"supporting_citations":[{"why":"Supplies the extrapolation method that turns the simulated rare-process samples into the 0.015 background yield per 3×10^14 EOT.","marker":"[22]"},{"why":"The CalcHEP generator used to produce dark photon signal events and inclusive cross sections for the sensitivity calculation.","marker":"[45]"},{"why":"The Geant4 simulation toolkit used for the full detector response and for modeling signal and background processes.","marker":"[43]"},{"why":"The current leading fixed-target exclusion limit that DarkSHINE aims to improve on by nearly two orders of magnitude.","marker":"[46]"},{"why":"One of the existing experimental constraints used as a reference point in the exclusion-limit plots.","marker":"[48]"},{"why":"The comparable fixed-target proposal whose projected sensitivity DarkSHINE benchmarks against in the comparison figures.","marker":"[3]"},{"why":"Provides the estimates of irreducible neutrino backgrounds used in the target material selection and background budget.","marker":"[17]"},{"why":"Documents the ECAL readout electronics prototype that meets the 1 MHz event-rate requirement, supporting the rate-capability claim.","marker":"[20]"}],"fun_headline_variants":["DarkSHINE: near-zero background dark photon search","Two-order sensitivity leap for dark photons at DarkSHINE","0.015 background events in DarkSHINE dark matter probe","DarkSHINE opens window to MeV-scale dark matter","Missing-energy signature targets dark photons 100x better"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole sensitivity projection rests on the estimated background of 0.015 events per 3×$10^{14}$ electrons on target, obtained by extrapolating from simulated rare processes; if the real detector produces even a few times more background in the signal box, the claimed limits weaken substantially.","fun_headline_variants_meta":{"raw":{"variants":["DarkSHINE: near-zero background dark photon search","Two-order sensitivity leap for dark photons at DarkSHINE","0.015 background events in DarkSHINE dark matter probe","DarkSHINE opens window to MeV-scale dark matter","Missing-energy signature targets dark photons 100x better"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1536,"prompt_tokens":836,"completion_tokens":700,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":452,"completion_tokens_details":{"reasoning_tokens":631}},"tokens_in":452,"tokens_out":700,"duration_ms":6972,"temperature":1.0,"reasoning_tokens":631,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:44:13.700094+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take data with the full detector and beam on a 0.35 mm tungsten target for the equivalent of 3×$10^{14}$ electrons on target and count events passing the five selection cuts: one track, missing momentum above 4 GeV, ECAL total energy below 2.5 GeV, HCAL total energy below 0.1 GeV, and HCAL maximum cell energy below 2 MeV. If more than about one event is observed, the 0.015 background estimate is falsified and the projected ε² exclusion curve does not hold at face value.","supporting_citations":[{"cited_title":"Chen et al., Sci","cited_arxiv_id":null,"evidence_quote":"Supplies the extrapolation method that turns the simulated rare-process samples into the 0.015 background yield per 3×10^14 EOT."},{"cited_title":"Belyaev, N","cited_arxiv_id":null,"evidence_quote":"The CalcHEP generator used to produce dark photon signal events and inclusive cross sections for the sensitivity calculation."},{"cited_title":"Design of High-speed readout electronics for the DarkSHINE electromagnetic calorimeter","cited_arxiv_id":"2407.20723","evidence_quote":"Documents the ECAL readout electronics prototype that meets the 1 MHz event-rate requirement, supporting the rate-capability claim."}],"review_version":1}