{"id":"2dd1afc6-2704-4d0c-bf73-7baeca82f2d2","arxiv_id":"2507.21618","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"AstroPix4 has been characterized: it detects 14 keV gamma rays, achieves about 14% FWHM resolution at 122 keV, saturates around 250 keV, and has a roughly 90 micrometer depletion depth at -240 V.","lead":"AstroPix4, a new CMOS pixel sensor for a future gamma-ray space telescope, was tested with radioisotopes and found to detect gamma rays from 14 keV to about 250 keV, with 14% energy resolution at 122 keV. The measurements show the chip is quieter than its predecessor but still falls short of the mission's required range and resolution.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 14–250 keV dynamic range upper bound depends on an unvalidated injection-voltage linearity and a saturation-edge interpretation; direct calibration above 122 keV is needed.","rationale":"The central claim of the paper is a characterization of AstroPix4: dynamic range 14–250 keV, 16.6 keV FWHM at 122 keV, and 90 µm depletion at -240 V. Of these, the dynamic-range upper bound is the least supported. The resolution and depletion measurements follow from direct photopeak and rate measurements. The high-energy limit, however, is inferred from a CSA-saturation cutoff in the 137Cs spectrum mapped to ~250 keV using a calibration curve extended with injection data. That extension requires the assumption Q = CV ∝ E. The reader identified exactly this assumption, and I agree it is load-bearing. Low-energy agreement up to ~100 keV gives partial support, but no measured photopeak or precisely characterized spectral feature between 122 and 250 keV is used to validate the extension. The 133Ba Compton edges at 164 and 207 keV are mentioned only qualitatively; making them quantitative would be a useful but less decisive check. A dedicated beam calibration with a tunable monoenergetic source over 150–250 keV is the definitive test. If it shows the injection-based curve is biased, the dynamic range upper limit and consequent comparison to AMEGO-X requirements need revision. The unsupported power figure is a documentation gap rather than a threat to the central characterization; it does not change the verdict. Therefore, I would keep the reader's CONDITIONAL verdict.","tokens_in":5957,"tokens_out":8798,"duration_ms":116997,"concrete_test":"Perform a beam test with a tunable monoenergetic photon source at 150, 200, and 250 keV on a pixel calibrated as in Section 3, using identical readout settings; compare the measured ToT centroids with the modified calibration curve built from injection data. If any centroid deviates from the curve by more than the combined statistical and systematic uncertainty, the injection-linearity assumption is falsified and the reported ~250 keV upper bound is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 has no measured photopeak above 122 keV. The high-energy endpoint of the dynamic range is built from two steps: (i) injection data are 'corrected under the assumption of linearity between injection voltage and energy (Q = CV ∝ E)' (Fig. 3) to place calibration points above 122 keV, and (ii) the sharp ToT cutoff in the 137Cs spectrum (Fig. 2, ~140 µs) is identified as CSA saturation and mapped to ~250 keV using that extended calibration curve (Fig. 4). Both steps are load-bearing. The injection capacitor is nominally linear, and agreement below ~100 keV is encouraging, but the paper shows no direct measurement of the injector charge at high amplitude and no independent calibration point between 122 and 250 keV. The 133Ba Compton edges at 164 and 207 keV are used only qualitatively. If injector nonlinearity (e.g., switch resistance, parasitic capacitance, or voltage-dependent gain) or a ToT digitization/clamping artifact, rather than CSA saturation, sets the 137Cs cutoff, the reported ~250 keV boundary shifts and the comparison with AMEGO-X requirements changes. The power-consumption figure is also not documented, but it is secondary to the central characterization.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a laboratory characterization of AstroPix4, a 16×13-pixel HV-CMOS active pixel sensor developed for the proposed AMEGO-X mission. Using Time-over-Threshold (ToT) spectra from five radioisotopes, the authors identify six photopeaks, fit an empirical calibration curve, and use charge-injection data to extend the calibration beyond 122 keV under an explicit linearity assumption. They report a gamma-ray dynamic range from 14 keV to about 250 keV, a FWHM energy resolution of 16.6 keV (about 14%) at 122 keV, a depletion depth of about 90 micrometers at -240 V, and a power consumption of about 2 mW/cm^2. The depletion-depth data are compared with a simple PN-junction model.","tokens_in":6239,"tokens_out":5834,"duration_ms":73434,"significance":"If the reported performance figures hold, this is a useful milestone for the AstroPix development program: it demonstrates detection of the 14 keV photopeak in a 500-micrometer-pitch HV-CMOS pixel, quantifies the improvement in 60 keV resolution relative to AstroPix3, and provides a depletion-depth-versus-bias dataset checked against a physical model. The six-photopeak calibration is anchored to external isotope lines, which is a clear strength, and the paper openly states the assumption used for the injection-based extension. The main weakness is that the upper end of the dynamic range, a central claim of the paper, depends on an unvalidated linearity assumption for the charge injector and on a saturation-edge interpretation that is not independently confirmed. These issues need to be addressed before the dynamic range claim is fully supported.","major_comments":[{"comment":"The claim that the dynamic range extends to about 250 keV is load-bearing and currently rests on two unvalidated steps. First, the injection data above 122 keV are corrected under the assumption that Q = CV is proportional to deposited energy (stated in Section 3 and Fig. 3), but no high-amplitude validation of the injector's charge-versus-voltage response is shown. Second, the sharp cutoff near 140 microseconds in the 137Cs spectrum is identified as CSA saturation and mapped to about 250 keV using the extended calibration curve, but the paper does not rule out other origins for the cutoff, such as comparator saturation or ToT digitization/clamping effects. The 133Ba Compton edges at 164 and 207 keV are only used qualitatively, so there is no validated calibration point between 122 keV and the reported saturation boundary. Please provide a direct calibration point above 122 keV (for example, a monoenergetic source or beam measurement) or an independent characterization of the injector linearity at the relevant amplitudes, and propagate the resulting systematic uncertainty into the reported 250 keV upper limit.","section":"Section 3, Figs. 3 and 4"},{"comment":"The interpretation of the 137Cs cutoff as CSA saturation would be more convincing if supported by a direct measurement of the CSA response or by a controlled scan of injection amplitude and comparator threshold in the saturation region. As written, the paper asserts the saturation interpretation after the fact, and the alternative that the cutoff arises from the readout chain rather than the CSA is not quantitatively excluded. Since the upper dynamic range determines the comparison with the AMEGO-X requirement of up to 700 keV, this ambiguity is consequential for the paper's central characterization.","section":"Section 3, Fig. 4 and Section 5"}],"minor_comments":[{"comment":"The power consumption figure of about 2 mW/cm^2 is stated without measurement details, operating conditions, or uncertainty. Please specify how this value was obtained and define the active area used in the normalization.","section":"Abstract and Section 2"},{"comment":"The injection-voltage axis label 'mV / 0.08' is not self-explanatory. Please define the conversion factor between injection voltage and injected charge, and state whether it is a design value or a measured value.","section":"Fig. 3 caption and Section 3"},{"comment":"The dynamic range lower bound is presented as 14 keV based on the visibility of the 57Co photopeak, but the paper does not report the distribution of noise thresholds or the fraction of pixels that can detect 14 keV. Clarifying the pixel-to-pixel spread would make the lower-end claim more robust.","section":"Section 3"},{"comment":"The caption 'Gauss ± Gauss' is unclear; it likely means the Gaussian fit mean plus/minus the fitted sigma. Please rewrite for clarity.","section":"Fig. 7 caption"},{"comment":"Reference [8] is incomplete: 'J. of Instrum., 19 (2024)' lacks an article number. Please complete the bibliographic information.","section":"References"},{"comment":"There is a LaTeX artifact in the abstract: 'mW/{cm}^2' should be 'mW/cm^2'.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of a detector instrumentation journal and the main characterization data are valuable. The concern raised by the skeptical reader about the injection-linearity assumption is legitimate and should be addressed before acceptance; the recommended revision should include either a direct high-energy calibration measurement or a clearly quantified validation of the injector response, plus a revised uncertainty on the 250 keV dynamic range boundary."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a straightforward, honest detector characterization paper: they put radioactive sources on AstroPix4, calibrated six photopeaks, measured depletion depth, and report numbers that look internally consistent. Second, the high-energy end of the claimed 14–250 keV dynamic range is built on an assumption—linearity between charge-injection voltage and deposited energy—plus reading a cutoff in the 137Cs spectrum as CSA saturation. That upper bound is plausible but not proven, and any referee should push for direct calibration above 122 keV.\n\nWhat's genuinely new is the first gamma-ray response characterization of AstroPix4. The 14 keV photopeak detection is a real improvement from the reduced input capacitance, the 122 keV resolution (~14% FWHM) is honestly reported as not meeting the AMEGO-X requirement, and the depletion-depth versus bias measurements follow a PN junction model with a plausible resistivity estimate of ~300 Ω·cm. The injection-based calibration extension beyond 122 keV is a modest engineering solution, not a fundamental advance, but it is clearly described.\n\nThe soft spots are real but proportionate. The injection-linearity assumption is load-bearing. In Figure 3, the raw injection data deviate from the photopeak-based curve above ~100 keV, and they correct those points assuming Q = CV ∝ E. There is no independent measurement of injector charge at high amplitude and no photopeak between 122 and 250 keV to anchor the correction. If the injector saturates or becomes nonlinear, or if the 137Cs cutoff is a ToT digitization artifact rather than true CSA saturation, the ~250 keV upper limit shifts. The authors flag the lack of higher-energy photopeaks and mention beam tests, which is honest, but the assumption remains unvalidated. The power-consumption figure (~2 mW/cm²) is quoted without any measurement detail; it is secondary but should be documented. The explanation for the degraded 122 keV resolution (photoelectron path length exceeding depletion depth) is plausible but not demonstrated.\n\nThis paper is for the AstroPix/AMEGO-X community and people working on HV-CMOS active pixel sensors for gamma-ray tracking. It deserves a serious referee rather than a desk reject. A revision that directly addresses injection-linearity validation and documents the power measurement would make the central claims solid. I'd send it to review with that request.","headline":"A solid, honest AstroPix4 characterization with a plausible 14–250 keV dynamic range, but the upper end rests on an unvalidated injection-linearity assumption that a referee should press on.","tokens_in":6756,"tokens_out":2999,"would_cite":true,"duration_ms":32465,"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":"AstroPix4 extends its gamma-ray sensitivity down to 14 keV and up to about 250 keV, with an energy resolution of 16.6 keV at 122 keV and a depletion depth near 90 micrometers.","keywords":["AstroPix4","HV-CMOS active pixel sensor","gamma-ray detector","Time-over-Threshold","energy calibration","depletion depth","AMEGO-X","MeV gamma-ray astronomy"],"falsifier":"Measure a known calibration line between 122 and 250 keV such as the 244.7 keV line of $^{152}$Eu, reconstruct its energy with the injection-corrected calibration curve, and compare; a mismatch beyond the quoted 68% confidence interval would show that the linear-injection assumption breaks down and the dynamic-range ceiling is not reliable.","tokens_in":5791,"feed_emoji":"📡","tokens_out":6849,"duration_ms":72556,"temperature":0.7,"pith_summary":"AstroPix4 is the latest iteration of a high-voltage CMOS active pixel sensor being developed for the proposed AMEGO-X gamma-ray space telescope, and this paper sets out to measure how well it actually detects gamma rays. The central result is that the chip's reduced input capacitance lowers the noise floor enough to see the 14 keV photopeak of $^{57}$Co, while the charge-sensitive amplifier saturates near 250 keV, so the estimated dynamic range is roughly 14 to 250 keV. At 122 keV the energy resolution is 16.6 keV (about 14% FWHM), which does not yet meet AMEGO-X's $<10\\%$ requirement, and the measured depletion depth is about 90 $\\mu$m at $-240$\\,V on a substrate whose resistivity cannot support full depletion. These numbers matter because they quantify how close AstroPix4 comes to the AMEGO-X requirements and directly motivate the planned AstroPix5 changes: a higher-range amplifier, fewer metal layers, and higher-resistivity wafers.","feed_headline":"AstroPix4 gamma-ray sensor spans 14 to 250 keV","feed_subtitle":"The lower-noise chip sees 14 keV gamma rays, but saturation and a thin depletion layer still miss AMEGO-X goals.","key_machinery":"The load-bearing object is the AstroPix4 pixel: each 500 $\\mu$m-pitch pixel collects ionization electrons in a deep N-well, reads them out through an intra-pixel charge-sensitive amplifier and comparator, and digitizes the pulse duration as a Time-over-Threshold value. Two analysis techniques carry the results: an empirical calibration function $\\mathrm{ToT} = aE + b(1-e^{-E/c}) + d$ fitted to six photopeaks, then extended above 122 keV by charge-injection data corrected under the assumption $Q = CV \\propto E$; and a photopeak-counting method using $^{241}$Am 60 keV gamma rays to derive depletion depth, with corrections for readout dead time and for photoelectron path length. The central mechanism being tested is whether the sensor's depletion layer and signal chain behave as a simple PN junction with linear charge collection.","core_discovery":"The paper's claim is that AstroPix4 has a gamma-ray response spanning from 14 keV to roughly 250 keV, set at the low end by noise (now reduced thanks to lower input capacitance) and at the high end by saturation of the in-pixel charge-sensitive amplifier. Using a modified calibration curve that extends photopeak fits with charge-injection data, the authors reconstruct spectra from five radioisotopes and confirm the upper cutoff in $^{137}$Cs. They report an energy resolution of $16.6^{+4.0}_{-2.6}$ keV FWHM at 122 keV (about 14%), a 30% improvement at 60 keV over AstroPix3, and a depletion depth of about 90 $\\mu$m at $-240$\\,V following a PN-junction model with inferred substrate resistivity of about $299\\pm5\\ \\Omega\\cdot\\text{cm}$.","pith_inferences":["If the injection-voltage linearity survives direct testing, the same calibration technique could be reused for AstroPix5's high-dynamic-range test columns, letting the team characterize saturation behavior with electronic pulses instead of radioactive sources in the 300-700 keV gap.","The hypothesis that 122 keV resolution is limited by photoelectron escape from a 90 $\\mu$m layer is directly testable by simulating photoelectron transport for that active depth and comparing the predicted 122 keV line width with the observed 16.6 keV.","A natural follow-up would be to check whether the 88 keV and 122 keV line widths scale with the inverse square root of collected charge; if they do not, the excess width at 122 keV has a source other than counting statistics, such as incomplete charge collection.","An independent anchor for the dynamic-range ceiling could come from a source like $^{152}$Eu, whose 244.7 keV gamma line lies just below the suspected saturation point, providing a real photopeak to test the injection-corrected calibration curve."],"forward_implications":["The chip's 14 keV sensitivity beats the 25 keV lower bound AMEGO-X asks for, but the roughly 250 keV saturation cutoff is far short of the 700 keV requirement, so the next sensor needs a higher-saturation amplifier.","A 14% FWHM resolution at 122 keV misses the required 10%, and the paper attributes this to 122 keV photoelectrons not being fully absorbed in the roughly 90 $\\mu$m depletion layer; deeper depletion should improve it.","Power consumption falls to about 2 mW/cm$^2$, half of AstroPix3, but still above the 1.5 mW/cm$^2$ AMEGO-X target.","The depletion depth grows with bias as a PN-junction model predicts, implying that fabricating AstroPix5 on higher-resistivity wafers should yield substantially larger depletion rather than requiring a new operating principle.","The injection-data calibration extension offers a way to map the sensor's upper response without waiting for rare high-energy photopeaks, and it is what locates the CSA saturation at about 250 keV."],"supporting_citations":[{"why":"Defines AMEGO-X mission context and the science case that sets AstroPix performance goals.","marker":"[1]"},{"why":"Sets the quantitative requirements table (energy range, resolution, power) against which AstroPix4 is judged.","marker":"[2]"},{"why":"Provides the depletion-depth analysis procedure and the AstroPix3 60 keV resolution baseline used for comparison.","marker":"[6]"},{"why":"Documents the AstroPix4 design changes, notably input-capacitance reduction, behind the improved noise floor.","marker":"[8]"},{"why":"Supplies the founding HV-CMOS monolithic pixel concept on which the AstroPix sensor architecture is based.","marker":"[3]"}],"fun_headline_variants":["AstroPix4 detects gamma rays from 14 to 250 keV","AstroPix4 sensor expands gamma range down to 14 keV","AstroPix4 reaches 14 keV but still short of AMEGO-X spec","AstroPix4: 14–250 keV gamma detection, with trade-offs","AstroPix4 gamma sensor spans 14–250 keV, not full spec"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The assumption that injection voltage is linearly proportional to deposited energy ($Q = CV \\propto E$) is load-bearing: it is used to correct injection data above 122 keV and to set the upper end of the stated dynamic range at roughly 250 keV, so if that linearity fails the high-energy calibration falls with it.","fun_headline_variants_meta":{"raw":{"variants":["AstroPix4 detects gamma rays from 14 to 250 keV","AstroPix4 sensor expands gamma range down to 14 keV","AstroPix4 reaches 14 keV but still short of AMEGO-X spec","AstroPix4: 14–250 keV gamma detection, with trade-offs","AstroPix4 gamma sensor spans 14–250 keV, not full spec"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000318,"raw_usage":{"total_tokens":1840,"prompt_tokens":1029,"completion_tokens":811,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":645,"completion_tokens_details":{"reasoning_tokens":704}},"tokens_in":645,"tokens_out":811,"duration_ms":8528,"temperature":1.0,"reasoning_tokens":704,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:33:04.203706+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure a known calibration line between 122 and 250 keV such as the 244.7 keV line of $^{152}$Eu, reconstruct its energy with the injection-corrected calibration curve, and compare; a mismatch beyond the quoted 68% confidence interval would show that the linear-injection assumption breaks down and the dynamic-range ceiling is not reliable.","supporting_citations":[{"cited_title":"Caputo et al., ``All-sky Medium Energy Gamma-ray Observatory eXplorer mission concept,'' J","cited_arxiv_id":null,"evidence_quote":"Sets the quantitative requirements table (energy range, resolution, power) against which AstroPix4 is judged."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the depletion-depth analysis procedure and the AstroPix3 60 keV resolution baseline used for comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the AstroPix4 design changes, notably input-capacitance reduction, behind the improved noise floor."},{"cited_title":"Brewer et al., ``Developing the future of gamma-ray astrophysics with monolithic silicon pixels,'' Nucl","cited_arxiv_id":null,"evidence_quote":"Supplies the founding HV-CMOS monolithic pixel concept on which the AstroPix sensor architecture is based."}],"review_version":1}