{"id":"679889b3-69d3-4ae6-8ca2-e9152e2ca75c","arxiv_id":"2411.08740","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"CE-65v2 MAPS test structures in 65 nm TPSCo show about 5% gain uniformity, around 1.5 to 2 micrometer spatial resolution at low thresholds, and over 99% efficiency up to roughly 180 electron seed threshold for the Modified with Gap process.","lead":"Scientists tested a next-generation test chip for the ALICE particle tracker using X-rays and particle beams, and found it has uniform pixel gains, high efficiency, and very good spatial resolution. The results help decide which 65 nm silicon process and pixel pitch to use in the upcoming ITS3 upgrade.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"A quoted 1.5 um DUT resolution from a telescope with 2.2 um resolution is not reproducible because the residual-deconvolution step is never described; the central sub-2 um claim depends on it.","rationale":"The reader's weakest-assumption analysis correctly identifies the missing deconvolution procedure as the most load-bearing gap. The paper otherwise presents first-characterization results that are plausible and partially supported: gain uniformity is demonstrated with 55Fe data, efficiency curves are shown for four chip variants, and charge-sharing ratios are quantified. However, the central sub-2 um spatial-resolution claim is the key performance figure for the ITS3 validation narrative, and the text does not specify how the DUT resolution was separated from the telescope's 2.2 um resolution. Because the quoted DUT value is below the telescope resolution, the extraction is inherently sensitive to the telescope model; without the residual estimator and covariance treatment, the claim cannot be independently checked. This is a reproducibility concern rather than evidence of a wrong result, so the appropriate verdict remains CONDITIONAL: the quantitative conclusions should be accepted only after the analysis method is documented and the abstract/conclusion resolution discrepancy is reconciled. My recommendation therefore leaves the reader's verdict unchanged.","tokens_in":848,"tokens_out":858,"duration_ms":33612,"concrete_test":"Request from the authors the exact Corryvreckan analysis configuration and the residual-width extraction for the lowest-threshold points in Figure 5a: state whether the quoted values are fitted sigma/RMS of track-to-cluster residuals, unbiased residuals, or pull widths, and specify how the telescope resolution was subtracted (e.g., sigma_DUT^2 = sigma_resid^2 - sigma_tel^2 with the 2.2 um value). Independently recompute the 70 e- point for the 15 um Standard chip from the raw test-beam ntuple: if the observed residual width is not consistent with sqrt(2.2^2 + 1.5^2) ~ 2.66 um, the reported DUT resolution is internally inconsistent. Also require a single sentence mapping the abstract '<2 um' claim to the conclusion's '<3 um' statement, identifying the threshold and pitch used.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline claim, 'spatial resolution of under 2 um during beam tests,' rests on the spatial-resolution extraction in Section 3. The telescope resolution is quoted as 2.2 um using a telescope optimizer, and Figure 5a reports DUT resolutions as low as ~1.5 um for the 15 um Standard chip at 70 e- threshold. No residual-based method is described that would allow the DUT resolution to be separated from the telescope contribution: the text gives no residual estimator (e.g., unbiased track residuals, pull distributions, or fitted residual sigma), no covariance propagation from the telescope, and no formula such as sigma_DUT^2 = sigma_resid^2 - sigma_tel^2. For Gaussian residuals, a measured width of ~2.66 um would be needed to extract a 1.5 um DUT resolution from a 2.2 um telescope; even small errors in the assumed telescope resolution therefore translate into large relative errors in the extracted DUT value. This directly affects the central quantitative claim used to validate the 65 nm TPSCo process for ITS3-style applications. The inconsistency between the abstract (<2 um) and the conclusion (<3 um) reinforces the concern: the reader cannot tell which resolution estimator, threshold, or operating point is being claimed. The paper does provide useful efficiency, charge-sharing, and gain-uniformity data, but the sub-2 um resolution figure is not reproducible from the text as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports the first characterisation results of the CE-65v2 analogue MAPS test structures fabricated in the 65 nm TPSCo CMOS process, developed in the context of the ALICE ITS3 upgrade. Four chip variants were studied: 15 um and 22.5 um pixel pitch, in the Standard and Modified-with-Gap process versions. Lab measurements with a 55Fe source establish matrix gain uniformity at the O(5%) level, and test-beam measurements at the CERN SPS provide efficiency versus seed threshold, spatial resolution versus threshold, and charge-sharing distributions. The paper reports over 99% efficiency up to about 130-150 e- for the Standard process and about 180 e- for the Modified-with-Gap process, spatial resolution as low as about 1.5 um at low threshold for the 15 um Standard chip, and a wider operating range with faster charge collection for the Modified-with-Gap process. The conclusion argues that these results validate the 65 nm TPSCo process for ITS3 and inform design choices for future detectors.","tokens_in":6687,"tokens_out":5976,"duration_ms":53459,"significance":"If the quantitative claims hold, this is a valuable direct measurement for the 65 nm TPSCo MAPS R&D programme, with direct relevance to ALICE ITS3 and to future experiments such as FCC-ee. The paper benefits from the use of external telescope tracks, energy calibration from the known 55Fe K_alpha line, a systematic comparison of process and pitch variants, and a clear charge-sharing diagnostic. The main quantitative claim, however, is the spatial resolution of 'under 2 um', and the manuscript currently does not describe how that number was extracted from telescope residuals; the abstract also disagrees with the conclusion's 'sub 3 um' wording. These issues must be resolved before the central claim can be accepted as reproducible.","major_comments":[{"comment":"The spatial-resolution extraction is not described. After stating that 'The telescope resolution was estimated to be 2.2 um using a telescope optimizer', the manuscript reports DUT resolutions as low as ~1.5 um in Figure 5a, but gives no residual estimator (unbiased track residuals, pull distributions, or fitted residual width), no covariance propagation from the telescope, and no deconvolution formula such as sigma_DUT^2 = sigma_resid^2 - sigma_tel^2. For Gaussian residuals, extracting a 1.5 um DUT resolution from a 2.2 um telescope requires a residual width of about 2.66 um, so modest errors in the assumed telescope resolution translate into large relative errors in the DUT value. Without an explicit description of the residual-based method, the abstract's headline 'under 2 um' claim is not reproducible from the text.","section":"Section 3"},{"comment":"The quantitative headline is internally inconsistent. The abstract claims 'a spatial resolution of under 2 um during beam tests', while the conclusion says 'sub 3 um spatial resolution obtained in the Standard process for both pitches' and Figure 5a shows the resolution rising to roughly 3-4 um at higher thresholds. The manuscript should state explicitly which estimator, seed threshold, and operating point each claim refers to, and quote a single headline resolution with its operating conditions. As written, the reader cannot determine whether the intended result is 1.5 um at 70 e-, under 2 um somewhere in the threshold scan, or sub 3 um over a range.","section":"Abstract and Section 4"},{"comment":"The test-beam efficiency and resolution curves are presented without error bars or numerical uncertainties. The comparisons between variants, such as 'over 99% efficiency up to ~130 e- and ~150 e-' for the Standard process and '~180 e-' for the Modified-with-Gap process, and the resolution values quoted in Figure 5, are therefore not quantitatively supported. Please include statistical uncertainties (at least) on the efficiency and resolution points, and state how the quoted threshold values and their uncertainties are derived.","section":"Section 3, Figures 4 and 5"}],"minor_comments":[{"comment":"There is a typo in the first sentence: 'Monolitihic' should be 'Monolithic'.","section":"Introduction"},{"comment":"The process name 'TPSco' appears once and should be 'TPSCo' to match the rest of the text.","section":"Introduction"},{"comment":"The sentence 'Figure 2b depicts the main K_alpha peak position for the 15 um Standard process chip' is confusing; Figure 2b appears to show a spectrum, not a peak position. Please clarify what is plotted.","section":"Section 2"},{"comment":"The terms 'electron threshold' and 'seed threshold' are used somewhat interchangeably. Please define both once and state explicitly how the x-axis thresholds of Figures 4 and 5 relate to the 100 e- seed threshold used in cluster building.","section":"Section 3"},{"comment":"The accumulated-charge-ratio metric can exceed 100% due to negative noise contributions after frame subtraction, as acknowledged in the text. Please state how negative pixel values are treated in the charge-ordering and normalization, because this affects the reported central-pixel charge fractions.","section":"Section 3, Figure 6"},{"comment":"Reference [12] is a web page; please provide the version or configuration of the telescope optimizer used for the quoted 2.2 um telescope resolution.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"For the editor: this is a solid characterisation study from an established collaboration, and the missing residual-deconvolution description in Section 3 appears to be an analysis-description omission rather than evidence of a flawed measurement. The main risk is that the paper's most prominent quantitative claim, the sub-2 um spatial resolution, is not reproducible as written. Please require the authors to describe the residual extraction, reconcile the abstract/conclusion wording, and add uncertainties before acceptance. I see no circularity concern in the use of the 55Fe calibration or the external telescope, and no inappropriate citation behavior."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is the first results paper for the CE-65v2 test structure in the 65 nm TPSCo process, and it gives useful efficiency, gain-uniformity, and charge-sharing data. The soft spot is the spatial resolution claim, which is under-supported because the residual deconvolution is never described.\n\nThe new thing here is the CE-65v2 chip itself, not in the cited prior papers. The paper provides a systematic comparison of Standard vs Modified-with-Gap process variants at 15 and 22.5 micron pitch. The gain uniformity at the O(5%) level is clean, the efficiency curves are sensible, and the accumulated-charge-ratio plots clearly show the difference between diffusion- and drift-dominated collection. That part is genuine and useful. The 55Fe calibration and test beam analysis use standard methods, which is appropriate for a characterisation paper.\n\nThe main weakness is the spatial resolution extraction. The telescope resolution is quoted as 2.2 microns, and Figure 5a shows DUT resolutions down to ~1.5 microns for the 15 micron Standard chip at 70 e- threshold. The text gives no residual estimator, no formula for subtracting the telescope contribution, and no covariance propagation. Without that, the headline 'under 2 micron' claim is not reproducible from the text. The abstract also says 'under 2 micron' while the conclusion says 'sub 3 micron', which is confusing. There are no error bars on efficiency or resolution either. These are real reporting gaps, and the resolution claim is important, so they need fixing.\n\nThat said, I don't think this is a fatal flaw. The qualitative trends are consistent with the expected charge-collection physics, and the efficiency and charge-sharing data do not depend on the deconvolution. The missing method description looks like an oversight rather than a bad measurement. A referee can ask for the extraction details and error bars without throwing out the paper.\n\nWho is this for? Detector instrumentation people working on MAPS for ITS3, FCC-ee, or similar. It deserves a serious peer review; the authors should be asked to clarify the resolution extraction and resolve the abstract/conclusion inconsistency. I would not cite this version in my own work, but I would read the revised version.","headline":"First CE-65v2 characterisation data for 65 nm TPSCo MAPS, with a real but fixable gap in how the sub-2 micron resolution is extracted.","tokens_in":7324,"tokens_out":2145,"would_cite":false,"duration_ms":19668,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Gx","29.40.Wk"],"model":"deepseek-v4-flash","headline":"A monolithic pixel sensor prototype is shown to track particles with sub-2 µm spatial resolution while keeping over 99% efficiency up to a 180-electron seed threshold.","keywords":["monolithic active pixel sensor","MAPS","65 nm CMOS","charge sharing","spatial resolution","tracking detector","test beam","iron-55 calibration"],"falsifier":"Reanalyse the stored test-beam data with an explicit, documented deconvolution of the 2.2 µm telescope resolution, for example subtracting the telescope contribution from the residual width in quadrature or fitting the residual distribution with the telescope response folded in, and check whether the extracted device resolution stays below 2 µm at a 70-electron seed threshold. If it does not, the paper's headline claim fails.","tokens_in":6244,"feed_emoji":"🎯","tokens_out":6453,"duration_ms":53451,"temperature":0.7,"pith_summary":"This paper reports first characterisation results for a family of small monolithic active pixel sensor test chips built in a 65 nm CMOS imaging process, aimed at the next upgrade of a large experiment's inner tracking system. The authors show that at a 70-electron seed threshold, the standard-process 15 µm-pitch chip reaches about 1.5 µm spatial resolution and the 22.5 µm-pitch chip about 2 µm, while both chips exceed 99% efficiency up to roughly 130 to 150 electrons. A process modification that adds a deep low-dose n-type implant with gaps at pixel edges gives drift-dominated charge collection, raising the 99%-efficiency threshold to about 180 electrons for both pitches at the cost of somewhat worse spatial resolution. Matrix gain uniformity at the few-percent level is demonstrated by iron-55 X-ray spectra. If these results hold, the process is a validated candidate for the planned tracker upgrade and the resolution-versus-operating-range tradeoff becomes a practical design handle.","feed_headline":"65 nm pixel chip tracks charged particles to under 2 µm","feed_subtitle":"Four chip variants exceed 99% efficiency up to a 180-electron seed threshold, widening the detector's safe operating range.","key_machinery":"The central object is the CE-65v2, a monolithic active pixel sensor test matrix of 1152 pixels with variants spanning two pixel pitches (15 and 22.5 µm) and two collection-layer processes, named Standard and Modified with Gap. The load-bearing mechanism is the difference in charge collection: the Standard process depletes a balloon-shaped region around each n-well electrode, leaving lateral charge to diffuse slowly and share widely; the Modified with Gap process adds a deep low-dose n-type implant with gaps at pixel edges, letting the depletion region extend laterally so collection is drift-dominated and charge stays concentrated. The paper measures the consequences through cluster charge-ordering, giving an accumulated charge ratio, and through efficiency and centre-of-mass resolution as functions of electron threshold.","core_discovery":"The central claim is that the CE-65v2 sensor variants meet the performance targets of the planned inner-tracker upgrade: spatial resolution below 2 µm at low seed threshold in the Standard process, and over 99% efficiency up to about 180 electrons in the Modified with Gap process. The controlling mechanism is charge sharing: the Standard process, with diffusion-dominated collection, produces large clusters that give finer centre-of-mass position but lose efficiency quickly as threshold rises; the Modified with Gap process adds a deep n-type implant with edge gaps, making collection drift-dominated, concentrating charge in one or two pixels, stabilising efficiency but degrading resolution. The paper concludes that these measurements validate the 65 nm CMOS process and motivate choosing pitch and process variant according to the radiation and occupancy environment.","pith_inferences":["If the same tradeoff holds after irradiation, the Modified with Gap process's concentrated charge collection should make it more resilient to charge trapping; the paper announces radiation studies, so this can be checked directly.","A smaller-pitch Modified with Gap chip, for example one with a 10 µm pitch, might recover much of the lost spatial resolution while keeping the wide operating range, since the resolution loss appears dominated by reduced cluster size rather than by pitch.","The 18 µm pitch and hexagonal-staggered geometry variants under analysis could interpolate the 15 to 22.5 µm behaviour and test whether a hexagonal arrangement reduces cluster-edge effects.","The reported sub-2 µm device resolutions sit below the quoted telescope resolution, so a fuller statement of the resolution-deconvolution procedure would let other groups reproduce the headline number from the same residual distributions."],"forward_implications":["Sub-2 µm spatial resolution in a 15 µm-pitch Standard-process chip at a 70-electron seed threshold satisfies the resolution target for the inner-tracker upgrade and for future lepton-collider tracking.","The 99%-efficiency operating range extends from roughly 130 to 150 electrons in the Standard process to about 180 electrons in the Modified with Gap process at both pitches, meaning the modified process tolerates higher noise and threshold settings.","The Standard process's resolution degrades quickly between 70 and 250 electrons before plateauing near 3 to 4 µm, while the Modified with Gap process trades the best resolution for a flatter, more stable response over threshold.","Matrix gain uniformity at the few-percent level in all four variants means full-matrix spectra can be used for energy calibration and pixel-to-pixel corrections are small.","The demonstrated performance validates the 65 nm CMOS imaging process as a candidate for wafer-scale bent sensors in the planned tracker upgrade, with the Modified with Gap variant favoured for high-radiation environments.","The charge-ordering method gives a direct empirical handle on cluster-size distributions and could be used to tune the reset voltage for optimal charge sharing, a knob the authors say they are exploring.","A smaller-pitch Modified with Gap chip, for example one with a 10 µm pitch, might recover much of the lost spatial resolution while keeping the wide operating range, since the resolution loss is dominated by reduced cluster size rather than by pitch.","The 18 µm pitch and hexagonal-staggered geometry variants under analysis could interpolate the 15 to 22.5 µm behaviour and test whether a hexagonal arrangement reduces cluster-edge effects."],"supporting_citations":[{"why":"Defines the process and the spatial-resolution and material-budget targets for the inner-tracker upgrade that the CE-65v2 is meant to validate.","marker":"[5]"},{"why":"Describes the process modification, a deep low-dose n-type implant with gaps, that creates drift-dominated charge collection and underlies the Modified with Gap variant.","marker":"[9]"},{"why":"Supplies the six reference planes of the test telescope used for track reconstruction in the beam-test measurements.","marker":"[11]"},{"why":"Provides the telescope-optimiser estimate of 2.2 µm total telescope resolution, the scale against which the device resolution is judged.","marker":"[12]"},{"why":"Track-reconstruction and analysis software used to build straight-line tracks, associate clusters, and produce the efficiency and resolution curves.","marker":"[13]"},{"why":"States the spatial-resolution requirement for a future lepton-collider tracker that the Standard-process results are said to satisfy.","marker":"[4]"},{"why":"Prior analogue pixel test structure in the same 65 nm process whose validation the CE-65v2 results supplement.","marker":"[6]"}],"fun_headline_variants":["CE-65v2: sub-2 µm precision, >99% efficiency","65 nm MAPS: <2 µm resolution, robust to 180 e- threshold","Charge-sharing trade-off: sub-2 µm vs >99% efficiency","65 nm TPSCo MAPS validate ALICE ITS3 upgrade","Sub-2 µm and >99%: 65 nm MAPS for ITS3"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline sub-2 µm spatial resolution rests on knowing the test telescope's 2.2 µm resolution and on an undescribed deconvolution of the device resolution from the track residual distribution; if that deconvolution is wrong or unaccounted for, the sub-2 µm figure is not reproducible.","fun_headline_variants_meta":{"raw":{"variants":["CE-65v2: sub-2 µm precision, >99% efficiency","65 nm MAPS: <2 µm resolution, robust to 180 e- threshold","Charge-sharing trade-off: sub-2 µm vs >99% efficiency","65 nm TPSCo MAPS validate ALICE ITS3 upgrade","Sub-2 µm and >99%: 65 nm MAPS for ITS3"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0011,"raw_usage":{"total_tokens":4607,"prompt_tokens":983,"completion_tokens":3624,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":3535}},"tokens_in":599,"tokens_out":3624,"duration_ms":25882,"temperature":1.0,"reasoning_tokens":3535,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T21:24:21.165295+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reanalyse the stored test-beam data with an explicit, documented deconvolution of the 2.2 µm telescope resolution, for example subtracting the telescope contribution from the residual width in quadrature or fitting the residual distribution with the telescope response folded in, and check whether the extracted device resolution stays below 2 µm at a 70-electron seed threshold. If it does not, the paper's headline claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the process and the spatial-resolution and material-budget targets for the inner-tracker upgrade that the CE-65v2 is meant to validate."},{"cited_title":"Snoeys, G","cited_arxiv_id":null,"evidence_quote":"Describes the process modification, a deep low-dose n-type implant with gaps, that creates drift-dominated charge collection and underlies the Modified with Gap variant."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the six reference planes of the test telescope used for track reconstruction in the beam-test measurements."},{"cited_title":"Telescope optimizer","cited_arxiv_id":null,"evidence_quote":"Provides the telescope-optimiser estimate of 2.2 µm total telescope resolution, the scale against which the device resolution is judged."},{"cited_title":"Characterisation of analogue Monolithic Active Pixel Sensor test structures implemented in a 65 nm CMOS imaging process","cited_arxiv_id":"2403.08952","evidence_quote":"Prior analogue pixel test structure in the same 65 nm process whose validation the CE-65v2 results supplement."}],"review_version":1}