{"id":"05cec34f-0f12-47c6-a188-62371bab732d","arxiv_id":"2502.02244","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Gain suppression in FBK LGADs depends on detector thickness, proton energy, bias voltage, and angle, and is stronger for FBK than for previously measured HPK sensors.","lead":"This paper reports how much signal boost is reduced in thin silicon detectors (LGADs) when protons deposit large amounts of energy. The measurements cover three detector thicknesses, several beam angles and voltages, and inform detector choice for the PIONEER particle physics experiment.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative gain-suppression values for the 100 µm device rest on a PIN normalization that the paper itself shows to be ~20% wrong; the qualitative angle trend is likely robust.","rationale":"The central claim is a measurement claim with two parts: (i) all devices show substantial gain suppression relative to MIP gain, and (ii) suppression varies with angle, peaking around 40–60°. The load-bearing assumption is that the PIN reference gives the primary-ionization charge for each LGAD. The paper itself flags a thickness discrepancy for W8, making part (i) quantitatively uncertain for that device. The angle trend (ii) is less affected because it is seen consistently in all three thicknesses and in both stopping and punch-through configurations; an 80 vs 100 µm difference cannot easily produce a non-monotonic peak in the same 40–60° window in all devices. However, the paper's stated purpose (energy linearity for PIONEER) requires quantitative gain values, so the conditional verdict is appropriate. No independent evidence (simulation, error bars, cross-calibration) currently bounds the normalization bias, which is why this is the single most load-bearing concern.","tokens_in":4894,"tokens_out":7027,"duration_ms":72293,"concrete_test":"Re-analyze the W8 100 µm data using an 80 µm PIN active thickness: compute Q_PIN(θ,E) from SRIM/Geant4 stopping-power tables for each energy and angle, treating protons as stopping when the range exceeds the slanted chord 80 µm/cosθ, and re-derive the gains in Fig. 2b. If any corrected gain shifts by more than 20% or the angle of maximum gain moves by more than 5–10°, the quantitative W8 results (and the cross-thickness comparison in Section 6) should be labeled as bounds only, while the 50 µm and 150 µm conclusions remain.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5 defines gain for each proton run as G = Q_LGAD(θ,E,V) / Q_PIN(θ,E), implicitly assuming the corresponding PIN has the same active thickness and full charge collection as the LGAD apart from the gain layer. Section 4 and Fig. 1 (left) contradict this for the 100 µm device: the W8 PIN collects charge consistent with an 80 µm active thickness, a 20% normalization error. For stopping protons the error is not a simple 0.8 scale factor: a proton that stops in the 100 µm LGAD deposits roughly a 90 µm track in silicon, but the 80 µm PIN lets the last ~10 µm escape, so the denominator is missing not just 20% of the charge but the Bragg-peak tail, and the missing fraction depends on angle as the geometric path length crosses the range. Thus the reported W8 suppression numbers, and the device-to-device comparison in Section 6, carry an unquantified systematic error. The qualitative angle dependence is nevertheless supported by the 50 µm and 150 µm devices, whose PIN thicknesses match, so the concern does not overturn the central qualitative claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an experimental study of gain suppression in three FBK LGADs with active thicknesses of 50, 100, and 150 micrometers, using MeV-range protons from the CENPA tandem accelerator and a 90Sr MIP reference. Gain is defined as the ratio of collected charge in an LGAD to that in a matched PIN diode without a gain layer, measured as a function of bias voltage, incidence angle, and proton energy. The main finding is that all LGADs show substantial gain suppression relative to laboratory MIP gain, with a non-monotonic angular dependence: suppression increases up to 40-60 degrees and then decreases because of charge deposition near the gain layer. The authors also compare with earlier HPK measurements and attribute the stronger FBK suppression to a very shallow gain layer.","tokens_in":5081,"tokens_out":4749,"duration_ms":45329,"significance":"The paper provides a useful dataset for LGAD applications such as the PIONEER active target, where energy linearity for stopping particles is critical. The systematic variation of angle across three thicknesses is valuable, and the direct comparison with an external MIP benchmark is a strength. The paper does not rely on free parameters or fitting; the gain is directly measured. However, the quantitative gain values for the 100 micrometer device are unreliable without a corrected normalization, and the absence of error bars limits the precision of all reported numbers. The central qualitative claims are likely robust, but the paper needs revision to support its quantitative statements.","major_comments":[{"comment":"The 100 micrometer PIN device (W8 PIN) is shown in Sec. 4 to collect charge consistent with an 80 micrometer active thickness, yet the gain in Sec. 5 for the 100 micrometer LGAD is computed as the ratio to this same PIN. For stopping protons, the PIN's missing ~20 micrometers of active silicon removes not only about 20% of the deposited charge but also the Bragg-peak tail near the end of the proton range; this missing fraction is angle- and energy-dependent because the geometric track length changes with angle. The resulting W8 gain values and any device-to-device comparisons in Sec. 6 therefore carry an unquantified systematic error. Please correct the normalization, propagate the systematic uncertainty, or explicitly restrict the quantitative claims for the 100 micrometer device.","section":"Sec. 4 and Fig. 1 (left); Sec. 5"},{"comment":"No error bars or statistical uncertainties are shown for any of the gain points in Fig. 2, and the caption states that some runs were excluded because of amplifier or digitizer saturation. Since saturation occurs at the highest gains, the exclusion is correlated with the high-bias and large-angle points that are central to the claimed bias and angle trends. Please provide per-point uncertainties and a quantitative assessment of how the excluded runs affect the stated trends, or at least identify which points are affected and justify that the conclusions remain unchanged.","section":"Fig. 2 and Sec. 5"}],"minor_comments":[{"comment":"The first sentence says 'protons of 1.8 MeV, 3 MeV, etc. momenta were used'; since the listed quantities are kinetic energies, the wording should be 'energies' or 'kinetic energies' for consistency.","section":"Sec. 2"},{"comment":"The y-axis of the collected-charge plot has no units stated in the text or caption; please add units (e.g., fC) to the figure or the caption.","section":"Fig. 1 (left)"},{"comment":"The sentence 'The charge values reflect the PIN thickness at 50 um and 150 um (around 0.1 fC per 10 um)' would be clearer if the expected charge per micrometer were defined explicitly, since the reader must infer the conversion from energy loss.","section":"Sec. 4"},{"comment":"The comparison with HPK sensors from Ref. [7] is made without stating whether those data were reanalyzed with the same analysis package and similar run selections; please add a sentence clarifying the consistency of the comparison.","section":"Sec. 6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a straightforward experimental contribution that fits the journal's scope. The main issue is the unquantified normalization of the 100 micrometer device, which the authors themselves flag but do not propagate into the analysis. With a corrected normalization or a clear restriction of the quantitative claims, the paper could be suitable for publication. The lack of error bars also needs attention, though the qualitative trends appear credible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a clean, useful follow-up measurement of LGAD gain suppression for FBK sensors. The genuinely new content is the systematic angle, energy, and thickness dependence for 50, 100, and 150 µm devices, plus the direct comparison to their earlier HPK data. The measurement approach is straightforward — RBS proton beam, PIN normalization, 90Sr MIP reference — and the paper is honest about its main limitation: the 100 µm PIN behaves as if it were 80 µm thick.\n\nWhat it does well: the experimental campaign is careful, the data cover a useful parameter space for the PIONEER active target, and the reported qualitative trend — suppression grows with bias and with angle up to 40–60°, then drops at large angles due to charge deposited near the gain layer — is clearly supported by the figures. The discussion of the shallow FBK gain layer as a likely cause of stronger suppression than HPK is reasonable and appropriately labeled as a hypothesis.\n\nSoft spots, in proportion: the lack of error bars on any of the gain values is a real gap. Some runs are excluded for amplifier saturation, and the reader doesn't get a sense of run-to-run stability. The 100 µm PIN thickness discrepancy is acknowledged in Section 4 but not propagated into the quoted numbers. The stress-test analysis is right that this is not just a 20% scale error: for stopping protons, the missing last ~10 µm of silicon in the PIN removes part of the Bragg peak, so the denominator is too small in a way that depends on angle and energy. That means the 100 µm gain values, and any device-to-device comparison involving them, carry an unquantified systematic shift. The qualitative angle dependence for that device is still visible, and the 50 and 150 µm devices don't have this problem, so the central claim survives.\n\nVerdict: the paper deserves peer review and likely publication after the authors add uncertainty estimates, clarify the 100 µm thickness calibration, and either renormalize or caveat that dataset. The reader's conditional verdict is about right; I'd be slightly more positive about the qualitative contribution.","headline":"A modest, honest LGAD gain-suppression measurement for FBK sensors with a real normalization caveat on the 100 µm device, but the qualitative angle trend holds up.","tokens_in":5631,"tokens_out":3187,"would_cite":true,"duration_ms":28943,"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":"Gain suppression in LGADs peaks at 40–60 degrees for MeV proton deposits, growing with bias voltage.","keywords":["low-gain avalanche detector","LGAD","gain suppression","charge density","proton beam","energy linearity","active target","particle identification"],"falsifier":"Measure the active thickness of each reference PIN with an independent technique—capacitance–voltage profiling or alpha-particle energy loss—and recompute the gain from the raw LGAD and PIN charge data. If the 100 micrometer PIN is truly 80 micrometers, the absolute gain values for that sensor shift; the central claim survives only if the corrected gains still show the same strong bias-dependent suppression and the same angular peak at intermediate angles across all three thicknesses.","tokens_in":4726,"feed_emoji":"📉","tokens_out":12708,"duration_ms":115039,"temperature":0.7,"pith_summary":"This paper reports measurements of how the internal gain of low-gain avalanche detectors (LGADs) responds to dense deposits of charge from a tuned proton beam, using sensors with 50, 100, and 150 micrometers of active thickness. The central result is that for MeV-range protons, the measured gain is far below the gain the same detectors show for minimum-ionizing particles, and the shortfall grows as the bias voltage is increased. The suppression also changes with the angle of incidence: it gets stronger up to roughly 40–60 degrees and then weakens, because at large angles the proton's Bragg peak deposits more charge close to the gain layer. The authors conclude that gain suppression is a serious nonlinearity for any experiment using LGADs as an active target for energy measurement, because particle identification relies on a linear energy response.","feed_headline":"40 to 60 degrees is where LGAD gain suppression peaks","feed_subtitle":"Measured with MeV protons across 50, 100, and 150 µm sensors; suppression grows with bias voltage.","key_machinery":"The central object is the highly doped gain layer of an LGAD, the thin region where avalanche multiplication begins. The mechanism being studied is gain suppression: a dense cloud of ionization from a stopping proton drifts into the gain layer and, once multiplied, partially shields the electric field that sustains the avalanche, so the effective gain falls below its small-deposit value. The measurements trace this suppression by dividing the charge collected in an LGAD by the charge collected in a geometrically identical PIN diode without a gain layer, varying the bias voltage, proton energy, and angle of incidence; the shallow gain layer of the tested devices is the property they single out to explain the strength of the observed effect.","core_discovery":"The paper's central claim is that all tested LGADs compress their gain substantially when a MeV proton deposits a high density of charge, and that this compression is governed by bias voltage, incidence angle, and proton energy. The gain at a given bias is not constant: it drops sharply from its minimum-ionizing value, and the drop is biggest at intermediate angles of incidence rather than at perpendicular incidence. The authors attribute the angular behavior to two competing effects: angled tracks spread the charge across the gain layer, which reduces the local charge density and lessens suppression, while at large angles more of the Bragg-peak charge is deposited near the gain layer, which increases suppression because charge generated close to the gain layer is less effective at producing gain. They further note that the tested devices, which have a very shallow gain layer, show stronger suppression than comparable sensors with a different doping profile, suggesting that the compact high-field region amplifies the shielding of the multiplication field by the moving charge cloud.","pith_inferences":["A quantitative model could be built from the data: plot the measured suppression against the charge density projected on the gain layer, accounting for the proton's stopping profile and lateral diffusion; if a single critical density triggers the suppression, the angular peak at 40–60 degrees should emerge naturally from the geometry.","The same mechanism should affect the timing resolution of LGADs for highly ionizing particles, because the leading-edge signal depends on the amount and arrival time of charge at the gain layer; the paper does not measure timing, so this is a testable extension.","If a deeper gain layer reduces suppression, then devices currently being developed for large timing detectors could be tuned toward better linearity at the cost of some time resolution; this trade-off is a design question the paper's comparison motivates but does not resolve.","The 100 micrometer normalization problem implies that any absolute gain comparison between this dataset and other measurements should be treated with caution until the PIN thickness is independently verified."],"forward_implications":["For any dense deposit, the LGAD gain is a strong nonlinear function of deposited charge, so an active target must apply per-pulse, angle-dependent gain corrections to recover linear energy response.","Raising the bias voltage to increase gain for minimum-ionizing tracks amplifies the suppression for MeV protons, forcing a trade-off between timing resolution and energy linearity.","The suppression pattern means a stopping particle and a through-going particle of the same kinetic energy will register different charge, so particle identification based on energy deposits requires modeling the Bragg-peak position.","Because the angular peak of suppression lies near 40–60 degrees, tracks at these angles in a detector will be the most distorted if the calibration is done only at normal incidence.","The comparison between gain-layer profiles indicates that the doping depth of the gain layer is a design handle for reducing suppression, since a shallower gain layer is associated with stronger nonlinearity."],"supporting_citations":[{"why":"Introduces LGADs and the operating principle of the internal gain layer this paper studies.","marker":"[1]"},{"why":"Describes the 90Sr minimum-ionizing-particle calibration setup used to define the reference gain.","marker":"[2]"},{"why":"The previous gain-suppression measurement campaign at the same accelerator, whose data analysis package is reused here.","marker":"[7]"},{"why":"An independent study of ionization charge-density-induced gain suppression that provides prior evidence for the effect.","marker":"[8]"},{"why":"Documents the gain reduction mechanism of field shielding by multiplied carriers in LGADs, which the paper invokes to explain its results.","marker":"[9]"},{"why":"Shows that the depth at which X-rays are absorbed changes the LGAD response, supporting the near-gain-layer charge-deposition argument.","marker":"[13]"},{"why":"Extends that depth-dependence finding to focused X-rays and other LGAD geometries, used to interpret the large-angle behavior.","marker":"[14]"}],"fun_headline_variants":["LGAD gain suppression peaks at 40-60 degrees, not perpendicular","Angled proton beams worsen LGAD gain suppression","Bias and angle drive LGAD gain suppression, not thickness","MeV protons expose LGAD gain nonlinearity at mid angles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The gain is defined as the charge collected by an LGAD divided by the charge collected by a same-thickness PIN detector, which assumes the PIN has the same active thickness and charge-collection efficiency as the LGAD apart from the gain layer; the paper reports that its 100 micrometer PIN behaves as if it were 80 micrometers thick, so the normalization for that sensor is uncertain and the reported gains carry an unquantified systematic shift.","fun_headline_variants_meta":{"raw":{"variants":["LGAD gain suppression peaks at 40-60 degrees, not perpendicular","Angled proton beams worsen LGAD gain suppression","Bias and angle drive LGAD gain suppression, not thickness","MeV protons expose LGAD gain nonlinearity at mid angles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000409,"raw_usage":{"total_tokens":2103,"prompt_tokens":910,"completion_tokens":1193,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":526,"completion_tokens_details":{"reasoning_tokens":1123}},"tokens_in":526,"tokens_out":1193,"duration_ms":9068,"temperature":1.0,"reasoning_tokens":1123,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T12:48:30.356841+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the active thickness of each reference PIN with an independent technique—capacitance–voltage profiling or alpha-particle energy loss—and recompute the gain from the raw LGAD and PIN charge data. If the 100 micrometer PIN is truly 80 micrometers, the absolute gain values for that sensor shift; the central claim survives only if the corrected gains still show the same strong bias-dependent suppression and the same angular peak at intermediate angles across all three thicknesses.","supporting_citations":[{"cited_title":"Pellegrini et al.,Technology developments and first measurements of Low Gain Avalanche Detectors (LGAD) for high energy physics applications,Nucl","cited_arxiv_id":null,"evidence_quote":"Introduces LGADs and the operating principle of the internal gain layer this paper studies."},{"cited_title":"Comparison of 35 and 50 {\\mu}m thin HPK UFSD after neutron irradiation up to 6*10^15 neq/cm^2","cited_arxiv_id":"1803.02690","evidence_quote":"Describes the 90Sr minimum-ionizing-particle calibration setup used to define the reference gain."},{"cited_title":"Gain suppression study on LGADs at the CENPA tandem accelerator","cited_arxiv_id":"2405.02550","evidence_quote":"The previous gain-suppression measurement campaign at the same accelerator, whose data analysis package is reused here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"An independent study of ionization charge-density-induced gain suppression that provides prior evidence for the effect."},{"cited_title":"Currás, M","cited_arxiv_id":null,"evidence_quote":"Documents the gain reduction mechanism of field shielding by multiplied carriers in LGADs, which the paper invokes to explain its results."},{"cited_title":"Synchrotron light source X-ray detection with Low-Gain Avalanche Diodes","cited_arxiv_id":"2306.15798","evidence_quote":"Shows that the depth at which X-rays are absorbed changes the LGAD response, supporting the near-gain-layer charge-deposition argument."},{"cited_title":"Synchrotron light source focused X-ray detection with LGADs, AC-LGADs and TI-LGADs","cited_arxiv_id":"2504.18638","evidence_quote":"Extends that depth-dependence finding to focused X-rays and other LGAD geometries, used to interpret the large-angle behavior."}],"review_version":1}