{"id":"76941965-b01d-4b02-ba0e-8b55e9aaa22b","arxiv_id":"2607.18419","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A 4th-magnitude star saturates a ~150-pixel-diameter region in Roman's WFI after 170 s, and persistence from deeply saturated point sources decays to detector background within ~20 minutes.","lead":"Tests on the Roman Space Telescope's camera show that very bright stars can saturate a patch about 150 pixels wide, and the lingering signal fades to background in about 20 minutes. The measurements give survey planners concrete numbers for masking bright stars and scheduling observations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SORC-to-flight PSF/spectrum fidelity is the load-bearing assumption for the 150-pixel saturation diameter; a direct PSF-transfer test is needed before using it for flight masking.","rationale":"The reader's CONDITIONAL verdict and weakest-assumption identification are sound. My independent reading confirms the central persistence result has credible in-campaign support: the flat-field persistence comparison (Figs 23-24, 35-36) provides a check not contaminated by the SORC fiber leak, and the SCA 11 decay curves are broadly consistent. The saturation-diameter claim, however, rests entirely on the ground-stimulus-to-flight transfer, and the manuscript's own figures flag SORC-only artifacts (Fig 6 arcs) and a truncated spectrum (Fig 4). Because the saturation mask is formed from the same frames those artifacts affect, the ~150-pixel number has an unquantified systematic tied to the projector. This does not invalidate the paper for its stated risk-reduction purpose, but it does mean the headline masking number should be used with a margin until the SORC PSF is validated against stpsf or a flight-like model. I therefore keep the reader's CONDITIONAL verdict unchanged. No ad hominem is intended; the concern is about external validity of a lab stimulus, not the integrity of the measurements.","tokens_in":22684,"tokens_out":6611,"duration_ms":61402,"concrete_test":"Use the unsaturated mag 18 SORC PSF (SCA 11 final frame) after gain and linearity correction to measure the delivered PSF and encircled energy. Compare it to stpsf WFI F146 PSFs computed with (a) the SORC SPLIT-IR spectrum and (b) the full F146 bandpass with a representative stellar SED. For each PSF, simulate a mag 4 exposure with the Table 1 total flux, ~80,000 e- full well, and 170 s, applying the paper's 100,000-130,000 e- saturation mask to get the saturated-region diameter. If models (a) and (b) differ by >20%, or if masking the Fig 6 SORC arcs changes the mag 4 diameter by >30 pixels, the 150-px number needs a correction or a stated systematic margin before flight use.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline most used by survey planners is the ~150-pixel saturated-region diameter for a mag 4 source. That number is only transferable to orbit if the SORC's delivered PSF, stray-light environment, and spectrum match flight F146 illumination. The manuscript itself documents two deviations: Fig 4 shows the SPLIT-IR spectrum cuts on at 1.2 um, omitting 0.927-1.2 um of F146, and Fig 6 shows SORC-specific stray-light arcs that 'are not expected from the Roman telescope assembly in flight.' Since the saturation mask (and hence diameter) is defined by pixels reaching 100,000-130,000 e- in the final frame, any PSF broadening, redder spectrum, or extra projector stray light can enlarge the measured saturated region. The mag 4 diameter is reported without error bars or a PSF-model comparison, and the SORC characterization reference (Wake & Lyons, in prep) is not public. The persistence-decay claim is less exposed because it is independently compared with flat-field persistence from the same TVAC2 campaign; the unresolved issue is specifically the saturation-diameter/PSF transfer.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a bright point-source saturation and persistence test performed during the Roman WFI Thermal Vacuum campaign TVAC2. Using the SORC telescope simulator, nine in-focus point sources approximating stellar magnitudes 4–18 were projected through the F146 filter onto SCAs 4 and 11. The authors analyze up-the-ramp data to measure the growth of saturated regions, characterize non-linear slope behavior near the saturation front, and derive persistence decay curves from interleaved darks. The headline results are that a ~4 mag source produces a saturated region ~150 pixels in diameter after ~170 s (vs. ~15 pixels for ~12 mag), and that persistence from all tested sources decays to detector background levels (≲0.05 e− s−1) within about 20 minutes. The paper also reports a flux halo with defect shadows around the brightest source and makes its analysis products available via Zenodo and STScI repositories.","tokens_in":1715,"tokens_out":1715,"duration_ms":65783,"significance":"If correct, these measurements provide the first direct pre-flight characterization of deep point-source saturation and persistence for Roman WFI, with immediate value for survey planning, masking, and scheduling in the GBTDS and other bright-star fields. The paper is unusually detailed in its data processing: IRRC reference correction, superbias subtraction, per-pixel gains, classical non-linearity correction, explicit saturation-mask definitions, bad-pixel masks, dark-current subtraction, and a documented fiber-contamination subtraction procedure. The persistence comparison with flat-field illumination from the same TVAC2 campaign is a strong internal consistency check, and the public data products will enable further analysis. The principal weakness is the transferability of SORC-based measurements, especially the saturation-region diameter, to the flight telescope.","major_comments":[{"comment":"The headline saturation-region diameter (~150 px for mag 4, ~15 px for mag 12) is measured under SORC illumination whose spectrum (SPLIT-IR, 1.2–1.9 µm) is markedly narrower than F146 (0.927–2.0 µm) and whose optics produce stray-light arcs (Fig. 6) that are acknowledged to be absent in flight. The diameter is reported without error bars or a comparison with a modeled PSF (e.g., stpsf) that would support transfer to flight conditions. Since the saturation mask is defined by a 100,000–130,000 e− threshold in the final frame, any PSF broadening or spectral shift could change the measured diameter. Please provide a PSF-fidelity assessment (e.g., comparing a sub-saturated SORC PSF to stpsf predictions and propagating the uncertainty) or explicitly frame the 150-pixel value as a SORC-specific measurement with a corresponding caveat in the abstract and conclusions.","section":"§2.1, Fig. 4, §3.1"},{"comment":"The central persistence claim — decay to ≲0.05 e− s−1 within ~20 min — is not supported by a quantitative criterion. The abstract quotes ≲0.05 e− s−1, but the text does not show how this background threshold is derived, nor is it overlaid on the decay curves in Figs. 19 and 21. Several data points are explicitly labeled as oversubtracted or biased by stray light (mag 4 first three points; last points for SCA 4), and the first interleaved dark values for SCA 4 do not agree within the error bars. Please define 'detector background level,' show the threshold on the decay curves, and fit or tabulate the time at which each magnitude/SCA reaches the threshold (or provide another statistical summary). Without this, the 20-minute claim is not crisply evidenced.","section":"§2.2.3, §3.2"},{"comment":"The fiber-contamination subtraction uses an empirical flux threshold (0.62 e−/s for SCA 11, 0.75 e−/s for SCA 4) and a scaling factor for the 'lower' component that is tuned to force the lowest residual in the mask to zero. These are free parameters, and the paper itself reports oversubtraction for the mag 4 source in the first three interleaved darks. The sensitivity of the persistence decay curves to reasonable variations in the threshold and scaling factor should be quantified, or the correction validated using a source geometry where the fiber leak lies outside the ROI. This would establish that the persistence-decay conclusion is robust to the correction.","section":"§2.2.3, steps 1–4"}],"minor_comments":[{"comment":"The SORC characterization reference (Wake and Lyons et al., in prep) is not public. If possible, include additional details of the SORC PSF and its measured quality, or a preprint/technical note.","section":"§2.1"},{"comment":"The IRRC algorithm is cited as 'Rauscher et al., in prep' in addition to the STScI report; a public reference or more algorithmic detail would help reproducibility.","section":"§2.2.1"},{"comment":"The statement that the saturation region 'grows to about 150 pixels in diameter' should specify how the diameter is defined (e.g., equivalent circular diameter from the mask area) and should include an uncertainty estimate.","section":"§3.1"},{"comment":"The error bars are described as 'the average of the persistence interquartile range for all pixels within a given saturation mask.' This is unusual; clarify the statistical choice (e.g., why not the median absolute deviation or bootstrap uncertainty).","section":"§3.2"},{"comment":"For mag 17, only 6 pixels are in the saturation mask and all are affected by fiber contamination. This context should be stated in the figure caption or main text when interpreting the mag 17 point.","section":"Fig. 20 caption"},{"comment":"The flux-halo/defect-shadow explanation rests on 'internal discussions with detector experts' and a private communication with T. Brandt. This is speculative; label it clearly as a hypothesis rather than an established mechanism, and cite the NIRCam in-flight observation once the analysis is public.","section":"§4.1"},{"comment":"The statement 'We make our analysis products publicly available' is qualified by 'the Zenodo repository will be published after paper acceptance.' Please clarify the current availability status in the submitted version.","section":"§4.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is careful and transparent, and the persistence results are bolstered by the flat-field comparison. The main concern is the SORC-to-flight transfer of the saturation-region diameter: if the authors add a PSF-model comparison or an explicit caveat, and tighten the quantitative definition of the 20-minute decay threshold, I would be supportive of publication. The flux-halo section is speculative but clearly flagged; it does not affect the central claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is a solid, much-needed measurement paper. It fills a real gap: Roman WFI's H4RG-10 SCAs had never been characterized under deep point-source saturation before — prior testing was flat-field only up to about 15-16 mag equivalent. The authors used the SORC telescope simulator to project ~4 to ~18 mag point sources onto two SCAs during TVAC2, with careful up-the-ramp processing. The result that will get used is the persistence number: after deep saturation, persistence decays to background (<0.05 e-/s) in roughly 20 minutes, across magnitudes, on both SCAs, and it is consistent with flat-field persistence from the same campaign. That internal consistency is the right kind of check, and it makes the persistence claim credible. The charge-leakage slope enhancement at the saturation front is also a clear, useful observation, consistent with earlier H2RG work.\n\nThe main caveat is the SORC. The paper's own figures show the SPLIT-IR spectrum cuts on at 1.2 um while F146 starts at 0.927 um, and the SORC produces structured stray-light arcs that won't be in flight. That means the absolute saturation diameter (~150 pixels for a mag 4 source) is the headline number most likely to be taken up by survey planners, and it is exactly the number that depends on the projector's PSF and spectrum. The paper reports it without error bars and without a PSF-model comparison. I don't think this sinks the persistence conclusion, because that is anchored by the flat-field comparison, but the 150-pixel figure needs to be labeled as SORC-dependent, or better, tested with a model of the projector.\n\nOther issues are more minor. The abstract says the first-post-illumination persistence is \"broadly consistent\" across magnitude; that is true for SCA 11, but not for SCA 4, where points disagree within error bars — the paper says so itself in Section 3.2, so the abstract should be qualified. The fiber-contamination subtraction is partly tuned to the persistence data, and the mag 4 decay curve shows oversubtraction in the first few points; the authors disclose this, but it limits what you can say about the earliest decay. The flux-halo discussion in Section 4.1 is speculative and rests on a private communication; it's fine as a flagged observation, but it should not be treated as established. And the Zenodo data are promised, not yet public.\n\nBottom line: this is a careful, mostly transparent instrument-characterization paper that the Roman survey planning community needs. I'd send it to a serious referee, asking for the abstract fix, public data, and a clear statement of the SORC transfer caveat. I'd bring it to reading group and would cite it for the persistence measurements.","headline":"A careful, much-needed measurement of Roman WFI deep point-source saturation and persistence; the 20-minute persistence decay is solid, but the 150-pixel saturation diameter carries an unquantified SORC-to-flight transfer caveat.","tokens_in":23535,"tokens_out":3957,"would_cite":true,"duration_ms":39010,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Deeply saturated bright stars in Roman survey fields will create large masked regions and short-lived persistence, with a 4th-magnitude source engulfing ~150 pixels in a ~170 s exposure and persistence fading to background within ~20 minute","keywords":["saturation","persistence","HgCdTe detectors","H4RG-10","Roman Space Telescope","thermal vacuum testing","charge leakage","band-edge emission"],"falsifier":"Measure, on orbit, the saturated-region diameter and persistence decay for a known ~4th-magnitude star observed through F146: if the saturated region differs substantially from ~150 pixels at ~170 s (beyond PSF differences) or if persistence remains above the detector background for more than ~20–30 minutes at the flight operating temperature, the central claims would be refuted. On the ground, repeating the test with the full F146 spectrum (extending the projector's cut-on from 1.2 µm down to 0.927 µm) would show whether the narrower test bandpass altered the saturation growth or the persiste","tokens_in":22522,"feed_emoji":"🌟","tokens_out":6720,"duration_ms":69600,"temperature":0.7,"pith_summary":"Roman's surveys will point at hundreds of thousands of stars bright enough to saturate the WFI detectors, yet prior ground characterization never pushed a point source deep into saturation. This paper reports a dedicated thermal-vacuum test in which nine point sources from ~4th to ~18th magnitude were projected onto two of the flight-like sensor chips. The central measurements: a ~4th-magnitude source saturates a region roughly 150 pixels in diameter after ~170 s (versus ~15 pixels at ~12th magnitude), pixels at the expanding saturation front show a dynamic charge-leakage non-linearity, and the persistence signal left after illumination decays to the detector background (≲0.05 e−/s) within about 20 minutes, largely independent of source brightness. If these numbers hold on orbit, survey planners can mask bright-star cores with known sizes and schedule repeat observations knowing that persistence will not linger beyond roughly one visitor.","feed_headline":"Mag 4 star saturates a 150-pixel region of Roman's WFI","feed_subtitle":"Persistence from saturated stars decays to detector background within ~20 minutes, regardless of magnitude.","key_machinery":"The analysis rests on a 56-frame up-the-ramp exposure sequence (reset frame plus 55 science frames) corrected with the IRRC reference correction, superbias subtraction, and per-pixel gain maps. A saturation mask flags pixels reaching 100,000–130,000 e− in the final frame; the saturation front is characterized by the ratio of instantaneous slope (difference between consecutive frames) to mean slope (average of pre-saturation frames) per pixel. Persistence is measured as the slope of interleaved dark exposures after subtracting a thermal dark-current frame and empirically constructed fiber-contamination masks, and the decay curves are the median persistence within each saturation mask versus t","core_discovery":"We measured the saturation response and persistence of two Roman WFI sensor chip assemblies by projecting nine in-focus point sources through the F146 filter with the SORC telescope simulator, at fluxes approximating stellar magnitudes ~4 to ~18, in ~170 s up-the-ramp exposures. The saturated region grows to ~150 pixels in diameter for a ~4 mag source versus ~15 pixels for ~12 mag, and pixels adjacent to the advancing saturation front exhibit a dynamic increase in instantaneous slope relative to their pre-saturation mean slope, consistent with charge leakage from saturated neighbors. In interleaved dark exposures, the median persistence within the saturated pixels is broadly consistent acros","pith_inferences":["If the trap population responsible for persistence is already filled at the fluxes tested, even deeper saturation (say a mag 2–3 source or a ~400 s exposure) should leave the ~20-minute recovery unchanged; a follow-up test could check that directly.","The near-magnitude independence of persistence suggests a simple scheduling rule for Roman operations: after any exposure containing a saturated star, wait roughly 25–30 minutes before using the same detector region for faint science, making persistence a fixed overhead.","The observed dynamic charge leakage could be modeled as a spreading-front wave triggered by neighbor saturation; if such a model were developed, it might recover reliable photometry for pixels that are currently discarded in crowded fields.","The flux halo may be mistaken for telescope stray light or PSF wings in ground test data; on-orbit observations of bright stars in touchstone fields should separate the internal band-edge halo from the flight telescope's scattered light."],"forward_implications":["A ~4 mag star in a ~170 s exposure will saturate a ~150-pixel-diameter region, so masking tools for GBTDS and other surveys can use this as a first-order footprint for the brightest targets.","Persistence from deeply saturated point sources decays to the detector background within about 20 minutes, so repeat visits to dense stellar fields can be planned with roughly that recovery timescale in mind.","Because the first post-illumination persistence is nearly independent of source magnitude, a single persistence correction curve may apply across a wide brightness range.","The charge leakage at the saturation front is time-dependent and scales with the pixel's own brightness, so photometry of pixels bordering saturated cores requires a model with a time-varying count-rate boost.","The band-edge halo redistributes flux laterally across the detector and casts defect shadows, so high-accuracy photometry of the brightest saturated sources must account for internal radiative emission."],"fun_headline_variants":["Mag-4 star saturates 150-pixel region in Roman WFI","Roman WFI: saturated stars show leakage, 20-min persistence","Saturation size and persistence measured for Roman WFI","Bright star saturation on Roman WFI: growth and recovery","Roman WFI tests: 150-pixel saturation, 20-minute persistence decay"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The quantitative results transfer to orbit only if the projector's point spread function, spectral passband (1.2–1.9 µm versus the F146 band 0.927–2.0 µm), and stray light faithfully mimic how the flight telescope will illuminate the detectors.","fun_headline_variants_meta":{"raw":{"variants":["Mag-4 star saturates 150-pixel region in Roman WFI","Roman WFI: saturated stars show leakage, 20-min persistence","Saturation size and persistence measured for Roman WFI","Bright star saturation on Roman WFI: growth and recovery","Roman WFI tests: 150-pixel saturation, 20-minute persistence decay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000801,"raw_usage":{"total_tokens":3423,"prompt_tokens":876,"completion_tokens":2547,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":620,"completion_tokens_details":{"reasoning_tokens":2455}},"tokens_in":620,"tokens_out":2547,"duration_ms":16305,"temperature":1.0,"reasoning_tokens":2455,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T15:28:16.196185+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure, on orbit, the saturated-region diameter and persistence decay for a known ~4th-magnitude star observed through F146: if the saturated region differs substantially from ~150 pixels at ~170 s (beyond PSF differences) or if persistence remains above the detector background for more than ~20–30 minutes at the flight operating temperature, the central claims would be refuted. On the ground, repeating the test with the full F146 spectrum (extending the projector's cut-on from 1.2 µm down to 0.927 µm) would show whether the narrower test bandpass altered the saturation growth or the persiste","supporting_citations":[],"review_version":1}