{"id":"b87ac199-03fc-4430-ba7f-7594b72c6a00","arxiv_id":"2501.07431","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A test-beam study shows silicon and GaAs pad sensors with edge readout meet signal-to-noise requirements for a compact calorimeter, with GaAs showing a notable signal drop between pads.","lead":"This paper reports test-beam measurements of novel silicon and gallium arsenide pad sensors for a compact electromagnetic calorimeter. It finds both sensor types detect 5 GeV electrons with good signal-to-noise, but the GaAs sensors lose up to 40% of signal between pads.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'very promising' conclusion rests on an unquantified tolerance of the measured 40% GaAs inter-pad signal loss; the paper itself calls for the missing simulation in Sec. 6.4.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the 40% GaAs inter-pad loss has not been shown to be correctable or acceptable. I examined other possible objections, such as the Monte Carlo agreement being softened by a fitted normalization factor (Sec. 6.2), the untested full dynamic range of the FLAME ASIC, and cross-talk estimates based partly on capacitance rather than direct measurement. These are genuine but secondary. The inter-pad loss is the only measured effect that the authors themselves flag as potentially jeopardizing the intended application, and the paper explicitly defers the decisive simulation. The raw measurements appear credible: the alignment is cross-checked, the system test is complete, and the conclusions are honestly hedged. The correct remedy is a detector-level simulation using the measured loss map, not a rejection of the sensor data. The reader's CONDITIONAL verdict is therefore appropriate, and no change to the verdict is needed, but the condition should be explicit: the GaAs technology is promising only if the inter-pad loss map is shown to be tolerable for the LUXE ECAL performance goals.","tokens_in":12173,"tokens_out":5000,"duration_ms":55237,"concrete_test":"Build a Geant4 model of the LUXE ECAL from Ref. [4], replacing the GaAs active layers with the measured position-dependent response map from Sec. 6.4 (40% loss over about 0.5 mm in x, 10% in y, interpolated from Figs. 21-22), and simulate 5 GeV single electrons plus representative multi-electron events. Reconstruct the shower energy and barycenter with and without a software correction based on that map, and compare with the LUXE ECAL resolution requirements. If the corrected or uncorrected position or energy resolution fails the requirements, the 'very promising' claim should be replaced by a conditional statement pending a mitigation strategy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that both Si and GaAs sensor planes are 'very promising' for a compact sampling calorimeter. For Si, the beam data support this: no inter-pad loss, SNR of 15, and intra-pad response variations of about 1%. For GaAs, however, Sec. 6.4 measures a 40% signal deficit in the horizontal direction (10% in the vertical) extending over about 0.5 mm around the inter-pad gap, attributed to the embedded aluminum traces. This is a local charge-collection inefficiency, not a simple geometrical dead area: the loss depends on the unknown impact point within a pad. In an electromagnetic shower, individual particle positions inside a pad are not known, so the loss cannot be removed by a single per-pad calibration factor; it contributes both a bias and an event-by-event fluctuation to the reconstructed shower energy and barycenter. The authors state that this effect 'may jeopardize the shower energy and position reconstruction' and that 'a detailed simulation study should quantify the effect,' but no such study is presented. The quoted SNR and homogeneity results use centered single-electron hits and therefore do not address this issue. The 'very promising' conclusion thus depends on the untested assumption that the inter-pad loss is either correctable or tolerable at the calorimeter level.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports test-beam measurements of 500-micrometer-thick silicon and GaAs pad sensors read out by FLAME ASICs, intended as active layers for a compact sampling calorimeter such as the LUXE ECAL. Using 5 GeV electrons and a pixel telescope, the authors measure single-pad signal distributions, signal-to-noise ratios (15 for Si, 17 for GaAs), intra-pad response uniformity, inter-pad signal losses, pad-to-pad response spreads, and signals induced by readout traces. They compare the measured distributions with a Geant4 simulation and conclude that both technologies are very promising for a highly compact and granular electromagnetic calorimeter.","tokens_in":12424,"tokens_out":4410,"duration_ms":43895,"significance":"If the conclusions hold, this is a valuable system-level validation of two relatively new thin-sensor readout concepts: Kapton fanouts with conductive glue for Si and embedded aluminium traces for GaAs. The strengths are the direct telescope-based alignment, the use of a dedicated ASIC readout chain, the explicit characterization of inter-pad losses and trace-induced signals, and the fact that the measurements address quantities directly needed for the LUXE ECAL design. The main limitation is that the 'very promising' conclusion for GaAs rests on an unquantified calorimeter-level tolerance of a measured 40% inter-pad signal loss; the paper itself identifies the missing simulation study. This does not invalidate the direct experimental results, but it makes the central claim for GaAs conditional.","major_comments":[{"comment":"The central claim that GaAs sensors are 'very promising' for a compact calorimeter is not supported quantitatively. Section 6.4 reports a roughly 40% loss in the horizontal direction and 10% in the vertical direction in the summed MPV around the inter-pad gap, and the Summary states that this 'may jeopardize the shower energy and position reconstruction'; however, the paper does not provide the simulation study it calls for. Because the loss depends on the unknown impact point of each particle within a pad, it cannot be removed by a single per-pad calibration factor and will contribute both a bias and event-by-event fluctuations to reconstructed showers. Please add a calorimeter-level simulation or an analytic estimate quantifying the effect on energy and position resolution, or explicitly reduce the GaAs conclusion to a conditional statement.","section":"6.4"},{"comment":"The claim of 'very good modelling' of the data by Geant4 is weakened by the free multiplicative correction factor of 1.05 introduced to match the MPV. Since this factor is fitted to the data, it absorbs unknown systematic effects such as charge-collection efficiency, gain calibration, or inactive layers, and the agreement is not parameter-free. Please quote the uncertainty on this factor, give a physical justification for its magnitude, and propagate its uncertainty into the simulation comparisons.","section":"6.2"},{"comment":"Quantitative claims about intra-pad uniformity and inter-pad losses are presented without statistical uncertainties. Figures 19-22 show MPV values or MPV ratios as functions of position with no error bars; the statements 'variations of about 1%' for Si and 'about 40%' and 'about 10%' for GaAs need uncertainties derived from the Landau-Gauss fits and from the finite number of tracks per strip. Please add error bars, or a table of fit results with errors, to support these quantitative conclusions.","section":"6.3-6.4"}],"minor_comments":[{"comment":"The caption of Fig. 4 refers to a 'SO2 passivation layer'; this should read 'SiO2'.","section":"2 (Fig. 4)"},{"comment":"There is a duplicated word in 'in the vertical direction direction' in the description of the GaAs inter-pad loss.","section":"6.4"},{"comment":"The phrase 'compared the the distribution' in the discussion of Fig. 25 contains a typo and should be corrected.","section":"6.6"},{"comment":"The integrated wrong-assignment ratio of about 2e-4 assumes a flat beam intensity; please state how this estimate changes if the measured non-uniform beam profile is used instead.","section":"6.6"},{"comment":"The expansion 'FcaL Asic forMultiplane rEadout' should be typeset as 'FCAL ASIC for Multiplane Readout'.","section":"3"},{"comment":"The abstract in the submission header and the abstract in the full text list slightly different topics (the header omits cross-talk and wrongly assigned signals); please make the two versions consistent.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"This is a solid experimental paper well within the scope of the journal. The direct measurements are useful and clearly described; the main issue is that the GaAs conclusion is stated more strongly than the evidence supports, because the calorimeter-level impact of the 40% inter-pad loss is deferred to future simulation. A revision that adds that simulation or tempers the conclusion would be satisfactory. I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: this is a solid, honest test-beam characterization of two pad-sensor technologies for a compact sampling calorimeter. The genuinely new item is the GaAs sensor with aluminium readout traces embedded in the gaps between pads—that design is reported for the first time here, and the paper gives it a thorough workout. The Si sensor uses a previously described Kapton fanout, but the full system integration with FLAME readout and telescope-based position measurement is useful.\n\nWhat the paper does well: the measurements are direct and carefully described. They report S/N of 15 (Si) and 17 (GaAs) for 5 GeV electrons, intra-pad response uniformity at the few-percent level for Si and about 1% strips, a careful Monte Carlo comparison using Geant4 with a readout model, and a cross-talk study. The identification of the wrong-pad signal fraction from hits on the GaAs traces—about 2e-4 relative to pad signals—is a nice piece of detective work. The authors are also clear about what they did not measure.\n\nThe soft spots are real but they are mostly acknowledged in the paper. The 40% signal loss between GaAs pads (10% vertically) is the big one. It's not a simple geometric dead area; it depends on impact point inside the pad, so a single per-pad calibration won't remove it. The authors say this 'may jeopardize' energy and position reconstruction and call for a detailed simulation—which is exactly right, but it means the 'very promising' conclusion is conditional until that simulation exists. A reviewer should ask for that study or for a softened claim. The MC comparison has a free normalization factor of 1.05 fitted to the data before claiming agreement; that's minor because the experimental results don't rest on it. Some uniformity plots lack error bars; that's also minor, since the effects are large and consistent.\n\nBottom line: this is a genuinely useful paper for anyone building a compact calorimeter, especially for LUXE or a luminometer. It deserves a serious referee, and the referee's main job is to push on the GaAs signal-loss question. I'd cite it for the GaAs design and the test-beam data.","headline":"A solid, honest test-beam study of two pad-sensor technologies; the GaAs inter-pad signal loss is a real but clearly flagged limitation that makes the 'very promising' conclusion conditional.","tokens_in":13060,"tokens_out":1884,"would_cite":true,"duration_ms":17447,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Vj","29.40.Wk"],"model":"deepseek-v4-flash","headline":"The paper reports full-system tests showing that ultra-thin silicon and GaAs pad sensors with edge readout reach signal-to-noise ratios of 15 and 17 and are promising active layers for a compact sampling calorimeter.","keywords":["silicon pad sensors","GaAs pad sensors","compact electromagnetic calorimeter","FLAME readout ASIC","edge readout traces","Kapton fanout","test-beam characterization","inter-pad signal loss"],"falsifier":"A simulation of the full calorimeter that plugs in the measured inter-pad response gaps and shows that reconstructed shower energy or position resolution fails the design target would falsify the paper's 'very promising' claim for the GaAs sensors; equivalently, a dedicated measurement of the missing charge in the 0.5 mm gap as a function of beam position and bias voltage would show whether the loss is recoverable.","tokens_in":11992,"feed_emoji":"⚡","tokens_out":7796,"duration_ms":72427,"temperature":0.7,"pith_summary":"The paper is testing whether two ways of making very thin active sensor planes can be used inside a compact electromagnetic sampling calorimeter, where the gap between tungsten absorber plates must stay small to preserve a small Molière radius. It puts 500 µm silicon pad sensors, read out through copper traces on a Kapton foil glued to the pads, and 500 µm GaAs pad sensors, read out through aluminium traces embedded between pads, in a 5 GeV electron beam with a tracking telescope and the FLAME readout ASIC. It reports that single-electron signals are described well by a Landau distribution convoluted with a Gaussian, with signal-to-noise of 15 for Si and 17 for GaAs, and that pad-to-pad response varies by about 3% and 2%. The paper concludes that the two technologies are very promising for a highly compact and granular electromagnetic calorimeter, while noting that the up to 40% signal loss measured between GaAs pads needs calibration and a simulation study to quantify its effect.","feed_headline":"Edge-readout thin sensors hit signal-to-noise of 15–17","feed_subtitle":"A full 5 GeV beam test shows silicon and GaAs planes can be the active layers of a compact calorimeter.","key_machinery":"The load-bearing elements are the two edge-readout schemes and the FLAME readout chain. In the silicon sensors, copper traces on a Kapton foil are attached to each pad with conductive glue, so the front-end ASICs can sit at the sensor edge instead of above the pads. In the GaAs sensors, 1 µm aluminium traces are embedded in the gaps between pads on a SiO2 passivation layer, routing signals to bond pads at the edge without any flexible printed circuit. FLAME, a 32-channel CMOS ASIC with a 10-bit ADC and a CR–RC shaper, amplifies and digitises the pad signals, and an FPGA applies a three-sample deconvolution to reconstruct pulse amplitudes. A beam telescope with six pixel planes gives the track impact point with about 37–40 µm uncertainty, which is what allows the response inside a pad and across the inter-pad gaps to be mapped. Together these parts make it possible to measure whether a sensor plane thinner than 1 mm can be read out without losing the signal or the shower position information.","core_discovery":"The central claim is that both sensor technologies can form the active layers of a compact calorimeter despite the unconventional routing of signals to the sensor edges. For the silicon sensors, no signal is lost in the inter-pad transition, the response is uniform to about 1% within a pad, and no cross-talk is observed. For the GaAs sensors, the embedded aluminium traces eliminate the need for a flexible fanout, but the wider dead gap between pads produces a response drop of up to 40% in one direction and 10% in the other, and a small fraction of the charge deposited under the traces is assigned to the wrong pads; the integrated size of those wrongly assigned signals is about $2\\times10^{-4}$ of a pad signal. The paper's own summary states that with this qualification the technologies, read out via FLAME ASICs, are very promising for a highly compact and granular electromagnetic calorimeter.","pith_inferences":["If the GaAs inter-pad loss cannot be reduced, a hybrid choice—Si planes where maximum uniformity is needed and GaAs planes where the more compact trace routing outweighs calibration effort—would be a natural design outcome, but this is my inference rather than the paper's conclusion.","The same edge-routing idea could be extended to sensors with many more pads or to thinner substrates, where wire-bond fanouts become mechanically impractical; the tests here establish that glued Kapton fanouts and embedded traces both work at the level required for a prototype.","A direct testable extension would be to measure the GaAs inter-pad signal loss as a function of bias voltage: the paper attributes the loss to the gap structure, and if it is mainly a charge-collection effect, higher bias might recover part of the missing signal.","The 2%–3% pad-to-pad response spread, measured with test-beam data, could be used to set production quality-control criteria for the sensor batch, since each pad's gain is separately correctable in the analysis chain."],"forward_implications":["A calorimeter built with these planes can keep the gap between tungsten plates near the minimum set by the absorber, preserving a small effective Molière radius.","The Si technology can provide a fine-grained active layer with no inter-pad dead region, so shower energy and position reconstruction need no correction for lost charge between pads.","The GaAs technology trades the Kapton fanout for on-substrate traces, at the price of a response dip between pads that must be calibrated and studied in simulation before the design is finalised.","The measured signal-to-noise values of 15 and 17 mean the sensors can see minimum-ionising particles with high efficiency, which is what is needed for in-calorimeter alignment and per-channel calibration.","The observed wrong-pad signal fraction of about $2\\times10^{-4}$ for electrons hitting GaAs traces is small enough that it should not dominate the calorimeter response, although the paper does not yet provide an end-to-end shower simulation."],"supporting_citations":[{"why":"defines the compact electromagnetic calorimeter requirements that motivate the sensor choice.","marker":"[4]"},{"why":"provides the earlier edge-readout segmented strip-sensor technology that the embedded-trace idea extends.","marker":"[8]"},{"why":"documents the Kapton-fanout readout scheme and the compact prototype whose performance motivates this system test.","marker":"[9]"},{"why":"supplies the high-resistivity GaAs:Cr sensor technology under test.","marker":"[12]"},{"why":"presents the FLAME readout ASIC whose specifications and block diagram anchor the electronics chain.","marker":"[13]"},{"why":"gives the three-sample deconvolution algorithm implemented in the FPGA for pulse reconstruction.","marker":"[15]"},{"why":"describes the test-beam facility used for the 5 GeV electron data.","marker":"[16]"},{"why":"provides the Monte Carlo particle-transport toolkit used to model the sensor response and compare it with data.","marker":"[23]"}],"fun_headline_variants":["Beam test shows silicon and GaAs suit compact calorimeter","Edge-readout silicon and GaAs sensors pass 5 GeV test","Silicon uniform, GaAs gaps: compact calorimeter test results","Test beam validates silicon and GaAs for compact calorimetry","5 GeV beam test: silicon and GaAs sensors for compact calorimeter"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The positive conclusion for the GaAs sensors rests on the untested assumption that the measured loss of up to 40% of the signal between pads can be corrected in software or accepted without degrading the calorimeter's shower energy and position reconstruction beyond its requirements.","fun_headline_variants_meta":{"raw":{"variants":["Beam test shows silicon and GaAs suit compact calorimeter","Edge-readout silicon and GaAs sensors pass 5 GeV test","Silicon uniform, GaAs gaps: compact calorimeter test results","Test beam validates silicon and GaAs for compact calorimetry","5 GeV beam test: silicon and GaAs sensors for compact calorimeter"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000824,"raw_usage":{"total_tokens":3614,"prompt_tokens":966,"completion_tokens":2648,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":582,"completion_tokens_details":{"reasoning_tokens":2573}},"tokens_in":582,"tokens_out":2648,"duration_ms":20216,"temperature":1.0,"reasoning_tokens":2573,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:41:41.810914+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A simulation of the full calorimeter that plugs in the measured inter-pad response gaps and shows that reconstructed shower energy or position resolution fails the design target would falsify the paper's 'very promising' claim for the GaAs sensors; equivalently, a dedicated measurement of the missing charge in the 0.5 mm gap as a function of beam position and bias voltage would show whether the loss is recoverable.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the earlier edge-readout segmented strip-sensor technology that the embedded-trace idea extends."},{"cited_title":"Performance and Moli`ere radius measurements using a compact prototype of LumiCal in an electron test beam","cited_arxiv_id":"1812.11426","evidence_quote":"documents the Kapton-fanout readout scheme and the compact prototype whose performance motivates this system test."},{"cited_title":"IEEE NUCLEAR SCIENCE SYM- POSIUM AND MEDICAL IMAGING CONFERENCE (2021)","cited_arxiv_id":null,"evidence_quote":"supplies the high-resistivity GaAs:Cr sensor technology under test."},{"cited_title":"https://indico.cern.ch/event/1127562/contributions/4904506/ attachments/2512388/4318796/moron TWEPP 2022 09 21.pdf","cited_arxiv_id":null,"evidence_quote":"presents the FLAME readout ASIC whose specifications and block diagram anchor the electronics chain."},{"cited_title":"Acta Phys","cited_arxiv_id":null,"evidence_quote":"gives the three-sample deconvolution algorithm implemented in the FPGA for pulse reconstruction."}],"review_version":1}