{"id":"de5e85e4-ccef-49f9-9810-2c0ac93726ec","arxiv_id":"2412.13780","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"First IV/CV measurements of carbon-infused LGAD sensors after non-uniform 24 GeV/c proton irradiation up to 1e16 p/cm2 show a common operating voltage may remain viable despite a roughly tenfold dose gradient.","lead":"A team measured how silicon timing detectors called LGADs behave after deliberate, uneven proton damage mimicking the radiation close to a collider beam. They report that even with a tenfold dose gradient across one sensor, a single high-voltage working point may still cover all tested pixels.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"IV/CV operating ranges cannot certify a timing working point: acceptor removal may require more bias for the 1e16 pixel than the least-irradiated pixel's breakdown limit.","rationale":"The reader's weakest assumption identifies exactly the load-bearing gap: the operating voltage windows are derived from IV derivative features, while the detector's intended function is precise timing. This is not a criticism of the measured data, which appear carefully acquired and honestly reported, but of the interpretive step from 'the current starts rising sharply and breakdown is not reached' to 'a common HV working point can operate the timing detector'. The physics makes the failure mode concrete: radiation-induced acceptor removal reduces the gain-layer doping, so the bias needed for adequate avalanche gain is likely higher than the depletion knee. The paper's own Fig. 10 shows a large acceptor-removal fraction at the highest doses, and Section 7 explicitly lists efficiency and time resolution as future measurements. Therefore the central operational claim should be treated as conditional pending those measurements, which is the reader's verdict. Since the reader already flagged this precise concern and the paper itself is appropriately hedged, no verdict adjustment is needed. The proposed gain-and-timing scan on the same or equivalent devices would settle whether the inferred common working point survives the transition from IV/CV to detector performance.","tokens_in":10593,"tokens_out":5614,"duration_ms":56106,"concrete_test":"Take the same non-uniformly irradiated device (or equivalent FBK W18 UFSD4 sensors uniformly irradiated to 0.3, 1, 5, and 10e15 p/cm2) and measure collected charge (gain) and time resolution as a function of bias at -20C, using a 90Sr source or a sub-ps laser with the intended readout. For each pixel define a timing-relevant bias window: V_low is the smallest bias where gain and time resolution meet the application requirement (for example, time resolution below about 35 ps and efficiency above 95%), and V_high is the bias where the least-irradiated pixel reaches the breakdown onset quoted in Table 1 (about 195 V for the 1e16 devices). If the intersection of these windows across the most- and least-irradiated pixels is non-empty, the central claim stands; if the most-irradiated pixel's required V_low exceeds the least-irradiated pixel's V_high, it falls.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central conclusion is that a single bias can operate all pixels despite a factor ~10 dose gradient (Section 7, Table 1). The evidence is entirely IV/CV. The operating range in Section 6.2 is defined as the voltage interval between the current-derivative knee and the breakdown onset (50% of the derivative peak). This is a bulk-depletion/breakdown proxy, not a timing-performance metric. Radiation damage in LGADs removes acceptors from the gain layer (Section 6.3); to restore the electric field and hence gain/timing after a 1e16 p/cm2 dose, the pixel likely needs a bias well above its depletion knee. The k-factor data (Fig. 10) show substantial acceptor removal at the highest dose. Thus a bias chosen inside the 70-195 V overlap could leave the most-irradiated pixel with unacceptably low gain and poor time resolution, even though its IV curve looks 'operational'. The authors explicitly defer efficiency and time-resolution measurements to future work (Section 7), confirming this gap. The concern is not that the IV measurements are wrong, but that the inference from dI/dV shape to a usable timing working point is not supported by the data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports IV and CV measurements of carbon-infused FBK UFSD4 LGAD sensors before and after highly non-uniform 24 GeV/c proton irradiation, with peak fluences up to 1e16 p/cm2 and a dose gradient of about a factor of 10 across the sensor. The authors define per-pixel operating voltage ranges from the derivative of the IV curves and find an overlap between the minimum operating voltage of the most irradiated pixels and the maximum operating voltage of the least irradiated pixels, concluding that a common high-voltage working point may be found for all pixels. They also extract the acceptor-removal fraction using the k-factor method and compare it with previous uniform-irradiation results.","tokens_in":10772,"tokens_out":6517,"duration_ms":59214,"significance":"This is the first measurement of post-irradiation IV/CV behavior of LGADs under a non-uniform dose profile relevant to forward proton timing detectors at the HL-LHC. The experimental work is careful: measurements are done at -20 C with per-pixel before/after comparisons, hysteresis checks, multiple devices, and a non-irradiated control. The data are credible and the study addresses a practical question that has not been directly measured before. If the common-operating-voltage claim is confirmed by later timing measurements, it would be of direct practical value. The main weakness is interpretive: the operating range is defined from IV/CV features, not from timing performance, and the paper explicitly defers gain and time-resolution measurements to future work.","major_comments":[{"comment":"The central claim that a common high-voltage working point may be found for all pixels is based entirely on IV-derived voltage ranges: the minimum is the voltage at the initial dI/dV peak (depletion knee) and the maximum is the voltage where dI/dV reaches 50% of that peak on the breakdown side. This is a depletion/breakdown proxy, not a timing-performance criterion. The paper's own acceptor-removal data (Section 6.3, Fig. 10) show that the gain-layer voltage drops to roughly 40-50% of its initial value at 1e16 p/cm2; maintaining sufficient gain for timing after such acceptor removal typically requires a bias well above the depletion knee, which could exceed the breakdown limit of the least irradiated pixel. Because no gain, efficiency, or time-resolution data are presented (as acknowledged in the final paragraph of Section 7), the conclusion 'a common high voltage working point may be found for all pixels' is not supported for the intended timing application. Please either (a) explicitly qualify the claim as an electrical operating point only, or (b) provide modeling or literature-based evidence that the IV/CV-defined range brackets the bias needed for adequate timing performance.","section":"Section 6.2 and Section 7 (Table 1)"},{"comment":"For the most irradiated pixel (0,0, 1e16 p/cm2), the 1/C^2 curve does not reach a plateau up to the maximum measured voltage of 65 V, so the full depletion voltage for that pixel is not actually bracketed by the CV data. Since the IV-derived minimum operating voltage for this pixel is about 70 V, the assertion that this pixel is above full depletion at its minimum operating voltage is not directly confirmed. Please either extend the CV measurement to higher voltages or explicitly state that the full-depletion voltage is a lower limit from the IV analysis only.","section":"Section 6.4, Fig. 11"},{"comment":"The horizontal dose axis uses per-pixel doses estimated from the BPM-derived beamspot profile, but the systematic uncertainty of this mapping is not quantified. The quoted 7% uncertainty is from foil activation only; the conversion from the beamspot shape to integrated dose per 1.3x1.3 mm2 pixel, especially across a steep gradient, likely has additional uncertainty. Please discuss or estimate this systematic effect; the qualitative conclusion is probably robust, but the comparison to uniform-irradiation data in Fig. 10 should account for this.","section":"Section 6.3, Fig. 10"}],"minor_comments":[{"comment":"The axis labels in Fig. 10 appear garbled in the manuscript (e.g., '15 1016 10] 2 [p/cm2' and '(0) gl )/V Φ ( gl V'). Please fix the LaTeX rendering of the axis titles.","section":"Figure 10"},{"comment":"The choice of 50% of the initial dI/dV peak as the definition of the maximum operating voltage is arbitrary; please add a sentence discussing the sensitivity of the overlap in Table 1 to this threshold, or cite a prior justification.","section":"Section 6.2"},{"comment":"The sentence 'The dose measured at the center of the beam was found to be 3-11% higher than the nominal target, with an uncertainty of 7%' is ambiguous: clarify whether the 7% uncertainty applies to the foil activation measurement of the absolute dose or to the correction factor.","section":"Section 4"},{"comment":"The captions of Figs. 5 and 6 do not explicitly state which color corresponds to which pixel for the pre-irradiation open markers; the text does this, but a short caption note would improve readability.","section":"Figures 5 and 6"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for NIM A and reports a useful first dataset. The main revision needed is to temper the conclusion about a common working point so that it clearly refers to an electrical operating range, given that timing performance is not measured. This is a framing issue rather than a flaw in the measurements themselves; if the authors are willing to add the qualification, the paper could become acceptable after minor changes, but as written the claim overreaches the data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a genuine first measurement of LGAD current/voltage behavior under a factor ~10 non-uniform proton dose, and the authors' main claim—that one bias voltage may cover all pixels—is supported by their IV/CV data but only at the depletion/breakdown level. Timing performance is explicitly deferred. Read the abstract as promise, not proof.\n\nWhat is new: previous LGAD irradiation studies were uniform, which is not the forward-proton condition. The paper measures three pixels along the diagonal of four irradiated sensors plus a control, before and after irradiation, with hysteresis checks, at -20 C, and the dose is cross-checked with foil activation. Table 1 shows, for every device, an overlap between the least-irradiated pixel's breakdown-limited upper voltage and the most-irradiated pixel's depletion knee. That is the empirical content, and it is a useful data point. The k-factor acceptor-removal comparison to uniform irradiations is a consistency check, not a circular fit. The reference list covers the relevant uniform-irradiation literature and the detector context.\n\nThe soft spots match the conditional verdict. The operating range is defined by a derivative-of-current threshold (50% of the initial dI/dV peak), which is not a gain or timing metric. The stress-test concern is fair: at 1e16 p/cm2 their own Fig. 10 shows substantial acceptor removal, and a timing detector needs enough gain-layer field. A bias above the IV knee may not restore the gain layer, so the headline \"common working point\" should be read as \"common bias voltage from an IV/CV standpoint.\" The authors do hedge this in Section 7, but the abstract and conclusions could easily over-read. Also soft: no raw data or code; quantitative operating ranges have only step-size uncertainties, with derivative-threshold choices not propagated; pixel dose is estimated from beamspot, not per-pixel dosimetry; the central [2,2] pixel anomaly is explained by ungrounded neighbors but not tested. These are limitations, not fatal flaws.\n\nWho is this for: people designing LGAD-based forward proton timing at HL-LHC and anyone tracking radiation tolerance of carbon-infused UFSD4. It deserves a serious referee. I would send it to review, and expect a referee to either require timing/efficiency data or make the authors narrow the claim to IV/CV compatibility.","headline":"First non-uniform-irradiation LGAD IV/CV data with an honestly hedged common-bias conclusion; the timing question is deferred, so treat the headline as a depletion-level result, not a timing-certified one.","tokens_in":11363,"tokens_out":3064,"would_cite":true,"duration_ms":30696,"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":"Carbon-infused silicon LGAD timing detectors can still be operated at a single bias voltage after non-uniform proton irradiation with a tenfold dose gradient and peak fluence of 1e16 p/cm2, based on current and capacitance measurements.","keywords":["Low Gain Avalanche Diode","LGAD","non-uniform irradiation","proton irradiation","acceptor removal","IV characteristics","CV characteristics","forward proton timing"],"falsifier":"Measure per-pixel gain, detection efficiency, and time resolution on the same devices after the same non-uniform irradiation. If the pixel exposed to 1e16 p/cm2 requires a bias voltage above the least-irradiated pixel's breakdown threshold to deliver the required timing resolution, the claimed single working point fails.","tokens_in":10371,"feed_emoji":"⚡","tokens_out":5197,"duration_ms":48904,"temperature":0.7,"pith_summary":"The paper reports the first current-voltage and capacitance-voltage measurements of carbon-infused silicon LGAD timing detectors (low-gain avalanche diodes, silicon sensors with modest internal gain) after a deliberately non-uniform 24 GeV/c proton irradiation, with peak fluence up to 1e16 p/cm2 and a factor-of-ten gradient across the sensor. It argues that even after such damage, every pixel on a single device can be operated at one common bias voltage: the most irradiated pixels reach their operating knee at or below 90 V, while the least irradiated pixels do not approach breakdown until above 200 V. This matters because forward proton timing detectors at the High-Luminosity LHC must sit millimeters from the beam, where radiation dose varies by an order of magnitude over a few centimeters, and the sensor design has only a single high-voltage connection. The conclusion is drawn from IV/CV behavior only; gain, efficiency, and timing resolution after irradiation are not measured here.","feed_headline":"One bias voltage covers factor-10 radiation spread in LGADs","feed_subtitle":"New IV/CV data show carbon-infused timing sensors can run at a single high voltage despite a tenfold dose spread.","key_machinery":"The analysis rests on the shape of the IV curve. A cubic-spline interpolation of the current versus voltage data is differentiated; the operating range of each pixel is taken from the initial peak of dI/dV (full depletion of the bulk) to the point where dI/dV again reaches 50% of that peak (onset of breakdown). The same derivative, weighted by V/I, defines the k-factor, whose first maximum gives the gain-layer voltage VGL used to quantify acceptor removal. This derivative-based definition of the operating voltage range is the mechanism that lets the paper compare pixels with very different dose on a common scale, and it is also the premise that would need to be validated against timing measurements.","core_discovery":"On its own terms, the central finding is that a common high-voltage working point exists for all pixels in a single carbon-infused LGAD despite roughly a factor of ten in local proton fluence, up to a peak of 1e16 p/cm2. After irradiation at -20 C, every pixel studied reaches its operational voltage at or below 90 V, with leakage currents near 1 microampere, while the least irradiated pixels show the onset of breakdown only above about 200 V. For all four irradiated devices, the maximum operating voltage of the least irradiated pixel lies above the minimum operating voltage of the most irradiated pixel, which is the condition for a single bias setting. The same devices show an acceptor-removal fraction that decreases with dose and is higher than in some earlier uniform-irradiation studies, attributed to the carbon-infused, radiation-hard wafer choice. The paper frames this as an encouraging first step for using LGADs in forward proton timing, and lists efficiency and time-resolution tests as future work.","pith_inferences":["If timing performance later confirms the IV-derived ranges, the single-common-voltage conclusion implies that per-pixel bias trimming is unnecessary; but if the most irradiated pixels need extra overdepletion for gain, per-pixel voltage control or stricter dose limits would be required.","The three measured pixels trace one diagonal; a full two-dimensional map would reveal whether the guard-ring and inter-pad effects seen in the central pixel also appear elsewhere, which matters for detectors with larger pixel arrays.","The higher surviving acceptor fraction than in earlier uniform-irradiation results suggests carbon-infused, radiation-hard wafers could tolerate even steeper dose gradients, which a dedicated uniform-dose comparison of the same wafer would test.","Because annealing changes acceptor removal over time, the common working point found immediately after irradiation may drift during High-Luminosity LHC operation; storing and re-measuring at controlled annealing steps would show how much margin remains."],"forward_implications":["Forward proton timing detectors at the High-Luminosity LHC could run non-uniformly irradiated LGADs with a single bias voltage, simplifying power and control systems.","After doses up to 1e16 p/cm2, the irradiated pixels still show a defined operating knee below 90 V at -20 C, so the sensors are not immediately unusable after a year at peak dose.","The least irradiated pixels retain margin before breakdown (onset above about 200 V), so the common operating point is not set by the undamaged corner.","Acceptor removal, quantified through the gain-layer voltage, trends consistently with dose across different devices, suggesting the damage mechanism is the same as in uniform irradiation.","CV measurements indicate that pixels receiving up to about 5e15 p/cm2 reach a similar capacitance as before irradiation, about 5-10 V higher, which supports the IV-based operating range for those pixels."],"supporting_citations":[{"why":"Supplies the sensor design and wafer selection that the paper identifies as the most radiation-hard of the production run.","marker":"[12]"},{"why":"Provides the forward-proton-detector radiation environment, including the non-uniform dose profile and peak fluence targets that motivate the study.","marker":"[14]"},{"why":"Supports the claim that carbon enrichment mitigates acceptor removal and supplies a comparison for the measured behavior.","marker":"[16]"},{"why":"Gives the acceptor-removal and operating-voltage expectations from uniform irradiation that the non-uniform results are compared against.","marker":"[20]"},{"why":"Defines the k-factor method used to extract the gain-layer voltage and acceptor-removal fraction.","marker":"[25]"},{"why":"Describes the irradiation facility and beam conditions used to deliver the 24 GeV/c proton dose.","marker":"[23, 24]"}],"fun_headline_variants":["Single bias voltage works across 10x proton dose in LGADs","LGADs run at one voltage despite 10x radiation spread","Carbon-infused LGADs tolerate non-uniform proton irradiation","One voltage fits all: LGADs survive factor-10 radiation gradient","LGAD timing sensors handle 10x dose with common bias"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire common-working-point conclusion rests on treating the current-voltage curve's knee and breakdown rise as the true usable operating window, even though the paper measures neither gain nor timing performance after irradiation.","fun_headline_variants_meta":{"raw":{"variants":["Single bias voltage works across 10x proton dose in LGADs","LGADs run at one voltage despite 10x radiation spread","Carbon-infused LGADs tolerate non-uniform proton irradiation","One voltage fits all: LGADs survive factor-10 radiation gradient","LGAD timing sensors handle 10x dose with common bias"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000628,"raw_usage":{"total_tokens":2860,"prompt_tokens":855,"completion_tokens":2005,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":1915}},"tokens_in":471,"tokens_out":2005,"duration_ms":13371,"temperature":1.0,"reasoning_tokens":1915,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:47:59.673610+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure per-pixel gain, detection efficiency, and time resolution on the same devices after the same non-uniform irradiation. If the pixel exposed to 1e16 p/cm2 requires a bias voltage above the least-irradiated pixel's breakdown threshold to deliver the required timing resolution, the claimed single working point fails.","supporting_citations":[{"cited_title":"Tornago, ”Detector optimization and physics performance of the CMS Phase-2 Endcap Timing Layer”, Thesis CERN-THESIS-2023-004, 2023","cited_arxiv_id":null,"evidence_quote":"Supplies the sensor design and wafer selection that the paper identifies as the most radiation-hard of the production run."},{"cited_title":"The CMS Precision Proton Spectrometer at the HL-LHC -- Expression of Interest","cited_arxiv_id":"2103.02752","evidence_quote":"Provides the forward-proton-detector radiation environment, including the non-uniform dose profile and peak fluence targets that motivate the study."},{"cited_title":"Effect of deep gain layer and Carbon infusion on LGAD radiation hardness","cited_arxiv_id":"2004.05260","evidence_quote":"Gives the acceptor-removal and operating-voltage expectations from uniform irradiation that the non-uniform results are compared against."},{"cited_title":"Bacchetta, D","cited_arxiv_id":null,"evidence_quote":"Defines the k-factor method used to extract the gain-layer voltage and acceptor-removal fraction."}],"review_version":1}