{"id":"6b08daad-c950-40ad-83b6-45533958987c","arxiv_id":"2509.07138","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"On-sky tests quantify additional throughput loss, background elevation, fast sky variability, and a derotator-induced background pattern that all limit NIRC2 high-contrast imaging at 3-5 microns.","lead":"This paper measures three things that limit how well the Keck-NIRC2 camera can see faint planets and disks near bright stars. It finds the L/M vortex coronagraph dims images more than expected in the M-band, sky background changes in under 30 seconds, and the image derotator adds a rotating dust pattern.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Table 1 in/out comparisons are single-epoch sequential frames with no quoted uncertainty, and Section 3's own sky-drift result means the 4-point Lp 'extra background' excess (86% vs 82%) may be a drift artifact rather than coronagraphic emission.","rationale":"The reader's CONDITIONAL verdict is appropriate, and I agree with its focus on the in/out isolation assumption. My stress-test sharpens that concern into an internal inconsistency: Section 3, one of the strongest parts of the paper, provides the failure mode for Section 2. If backgrounds drift on sub-30 s timescales, sequential in/out and background frames cannot be compared as though the sky were constant, especially with no error bars. The Lp extra-background signal is only 4 percentage points above the throughput ratio, the same order as plausible sky drift, so the paper's own variability measurement undermines its least-secure quantitative claim. The Ms effect is larger and more likely robust, and the derotator and sky-variability findings are plausible, so I would not reject the paper. The central claim that Table 1 quantifies a coronagraphic throughput and background penalty should be conditioned on a time-resolved re-reduction, which leaves the reader's verdict unchanged. This is a good-faith concern with a concrete archival test using the same frames the authors say will be public.","tokens_in":11729,"tokens_out":9119,"duration_ms":89144,"concrete_test":"Extract FITS-header timestamps for the 24 KOA frames (n0197-n0220) and reconstruct the time offsets between vortex-in, vortex-out, and background frames. Using the Section 3 sky-drift time series and power spectra, Monte Carlo propagate the expected background difference at those offsets. Recompute Table 1 with the sky level interpolated at each PSF frame time instead of a single background frame per state, and add the propagated drift as a systematic uncertainty. If the Lp background ratio moves from 86% to at or below the 82% throughput ratio, the claimed extra coronagraphic background in Lp is not established; if the Ms ratio moves from 70% toward 57%, part of the Ms extra-background claim also disappears. A null result would instead confirm the throughput and background penalties as real.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing quantitative result is the Section 2.2/2.3 comparison of vortex-in and vortex-out frames summarized in Table 1, which yields on-sky throughput 82% (Lp) and 57% (Ms) and background ratios of 86% (Lp) and 70% (Ms), despite the table's 'no/with' label apparently denoting with/without. These numbers come from four temporally sequential image types per filter, PSF in, PSF out, background in, background out, with three frames each mean-combined. No uncertainties are quoted, no timestamps are given, and there is no check against temporal sky variability. This matters because Section 3 demonstrates that background counts in Lp and Ms drift on timescales shorter than 30 s and are inconsistent with Poisson statistics. The evidence for added coronagraphic background is only the difference between the measured background ratio and the throughput ratio: in Lp, 86% versus 82%, a 4-percentage-point excess. A sky drift of roughly 4% between the in and out background frames, well within the variability documented in Section 3, would erase the Lp excess and could also bias the Ms excess of 13 points. Thus the paper's own variability measurement undermines the isolation assumption in the exact comparison supporting the central claim of an added background penalty. In addition, the statement that these effects explain part of the 1.7-2.0 magnitude discrepancy is not quantitatively tied to Table 1; no error budget propagates the measured penalties into the discrepancy. The Ms throughput deficit is larger and more likely robust, and the sky-variability and derotator findings are plausible, but the least-secure quantitative anchor of the central claim is the Table 1 background excess.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports on-sky tests with Keck-NIRC2 to identify factors that limit 3–5 micron high-contrast imaging. The authors measure the L/M vector vortex coronagraph throughput in Lp and Ms, compare background levels with and without the coronagraph, use the new fast-readout electronics to sample sky background at sub-second cadence, and document a quasi-static background pattern attributed to the image derotator. They conclude that the vortex coronagraph reduces throughput more in Ms (57% by aperture photometry) than in Lp (82%), that it adds background flux above the throughput penalty, that the sky background drifts on timescales shorter than 30 s, and that the derotator contributes spatial non-uniformity. The paper closes with observing recommendations, including shorter exposures and optics cleaning.","tokens_in":12045,"tokens_out":4351,"duration_ms":40418,"significance":"If the quantitative claims survive scrutiny, this is a useful empirical characterization for observers planning NIRC2 high-contrast imaging and for the design of future mid-infrared instruments. The direct comparison of on-sky throughput to published lab measurements, the use of the upgraded readout to probe sub-second sky variability, and the decision to place frames in the Keck Observatory Archive are clear strengths. The paper addresses a real discrepancy between predicted and on-sky sensitivity limits and its recommendations are actionable.","major_comments":[{"comment":"The on-sky throughput and background-ratio numbers (82%, 57%, 86%, 70%) are quoted without uncertainties, and the measurement is based on a single sequence of three frames per configuration on one night. Since Section 3 demonstrates that the sky background in Lp and Ms drifts on timescales shorter than 30 s and is inconsistent with Poisson statistics, the in/out background comparison can be biased by a few percent to tens of percent. In Lp the claimed 'elevated background' is only a 4-percentage-point excess of the background ratio over the throughput ratio (86% vs 82%), well within the drift documented in Section 3; the Ms excess (70% vs 57%) could also be partly drift. Please provide uncertainty estimates, timestamps, or a control measurement that quantifies drift between the 'in' and 'out' background frames, or weaken the claim of coronagraph-added background accordingly.","section":"§2.2–2.3, Table 1"},{"comment":"The claim that these effects explain part of the 1.7–2.0 magnitude discrepancy between predicted and on-sky sensitivity is not quantitatively connected to the measured penalties. No error budget or propagation is given that converts throughput, background ratio, drift, and derotator non-uniformity into a magnitude offset. Please add a quantitative accounting, even approximate, showing how the measured factors sum to the discrepancy, or explicitly state that the paper does not attempt this.","section":"§1.1 and §5"},{"comment":"The claim that 30 s background sequences are inconsistent with Poisson statistics is supported only by illustrative plots and a qualitative statement; no statistical test is reported. The manuscript should quantify the excess variance (e.g., variance-to-mean ratio, reduced chi-squared, or autocorrelation timescale) for the sequences in Table 2 and state how many of the 36 sequences show the effect.","section":"§3, Fig. 8, Table 2"},{"comment":"The de-rotator dust map is descriptive: it is derived from a single median-combined de-rotated stack, with no test that the pattern rotates at the de-rotator rate, no comparison with a non-de-rotated reduction, and no estimate of its impact on contrast. Given the paper's stated goal of documenting the derotator contribution, please add a simple rotation test or explicitly mark this section as preliminary only; the conclusions currently present it as an established effect.","section":"§4, Fig. 10"}],"minor_comments":[{"comment":"The manuscript needs proofreading for typos such as 'utalize' (§1.1), 'of of' (§2.1), 'corisponding' (Fig. 8 caption), and 'hypothesis' used as a verb (§2.3).","section":"Throughout"},{"comment":"The row 'Measured ratio of bg counts no/withvortexlm' should state explicitly whether the ratio is (with vortex)/(without vortex) or its inverse; the current label is ambiguous.","section":"Table 1"},{"comment":"Please state whether the aperture photometry and background measurements in Section 2.2 were performed on flat/dark-corrected frames or on raw frames, since Fig. 3 is described as 'before flat/dark corrections were applied.'","section":"§2.2, Fig. 3"},{"comment":"The two throughput methods (max-pixel ratio and aperture photometry) are not equivalent because the vortex alters the PSF morphology; the paper should state which value is used in subsequent conclusions and why.","section":"§2.2"},{"comment":"For the sky background tests, please state the selection criteria for the sequences shown in Fig. 8 and report whether the remaining sequences in Table 2 show the same behavior.","section":"§3"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Useful instrument paper, but the 'extra background' from the vortex is not on solid ground. The on-sky throughput numbers and the sub-30s sky drift are real contributions; the Table 1 background excess is a knife-edge measurement that the paper's own sky-variability section cuts against.\n\nWhat's new: the first on-sky Lp/Ms throughput of the vortexlm. Aperture photometry gives 82% (Lp) and 57% (Ms), with max-pixel ratios 86%/63%—internal consistency suggests the measurement is honest. The Ms value is below the lab result but within reach of the ensemble average, so it's plausible. The sky-background drift at <30 s, seen with the upgraded readout, is well supported by the plotted time series and the fact that darks show no drift. The derotator dust map is descriptive and preliminary, but worth documenting.\n\nSoft spots, in order of importance. First, Table 1's background ratios have no uncertainties, and the in/out background frames are sequential and not interleaved. The claimed 'elevated background' in Lp rests on an 86% vs 82% difference—four percentage points, which is exactly the size of the temporal drift Section 3 says exists. The Ms excess (70% vs 57%) is larger and more likely real, but still lacks error bars. Second, the paper says these effects explain part of the 1.7–2.0 mag sensitivity gap, but never propagates the measured penalties into an error budget; that connection is asserted, not shown. Third, the sky variability claim is based on a small number of sequences without a formal statistical test; the PSDs look convincing, but this is a collection of examples, not a survey. Minor: Table 1's 'no/with' label appears backwards given the numbers, and the abstract has a typo ('3−5µm First').\n\nFor a SPIE proceeding, this is honest, useful work. It deserves a serious referee and a revision. The fix is concrete: report frame-to-frame scatter or repeat the in/out pairs, and either support the 'extra background' claim with a number that survives the sky-drift floor or drop it. The throughput and sky-variability results will be cited. I'd engage with it.","headline":"Useful on-sky numbers for NIRC2, but the 'extra coronagraph background' claim is not secure against the sky drift the paper itself measures.","tokens_in":12585,"tokens_out":3591,"would_cite":true,"duration_ms":33899,"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":"Keck-NIRC2's gap between predicted and actual infrared sensitivity is traced to three instrument effects: the vortex coronagraph's throughput loss and added background, sky background drift faster than 30 seconds, and a rotating derotator…","keywords":["high-contrast imaging","infrared","Keck NIRC2","vector vortex coronagraph","sky background drift","image derotator","throughput measurement","exoplanet imaging"],"falsifier":"A bench measurement of the same vortex mask at 4.6 microns under a controlled source giving about 70% throughput would show the on-sky Ms deficit (57%) comes from telescope-state differences rather than the optic itself; alternatively, repeating the on-sky in/out test while dithering the derotator would separate the rotator pattern from the coronagraph's added background.","tokens_in":11561,"feed_emoji":"🔭","tokens_out":5705,"duration_ms":45075,"temperature":0.7,"pith_summary":"The paper tests three suspected culprits for why Keck-NIRC2 high-contrast imaging at 3–5 microns reaches only a fraction of its predicted sensitivity. On-sky measurements show the L/M vector vortex coronagraph passes only 82% of Lp light and 57% of Ms light, and it adds extra background flux that is worse in Ms. Fast-readout data reveal the sky background drifts on timescales under 30 seconds, so a 30-second exposure integrates over a moving background. The image derotator imprints a quasi-static rotating dust pattern that standard PCA background subtraction cannot remove. Together these effects account for part of the 1.7–2.0 magnitude gap between predicted and observed sensitivity limits.","feed_headline":"Three hidden losses cut Keck-NIRC2 sensitivity by 2 magnitudes","feed_subtitle":"On-sky tests trace the gap to the vortex coronagraph, sky drift under 30 seconds, and a rotating derotator dust pattern.","key_machinery":"The central object is the vector vortex coronagraph, a diamond optic with an antireflective grating that suppresses on-axis starlight. The paper measures its throughput by moving it in and out of the beam with a star offset 1 arcsecond, comparing aperture photometry, and also compares background counts in a 100×100 pixel corner to separate throughput loss from added background. The other two mechanisms are the sky background's temporal drift, sampled at 0.01–0.05 seconds per coadd using the upgraded readout electronics, and the image derotator's quasi-static dust pattern, found by median-combining derotated frames.","core_discovery":"The central claim is that three instrumental effects degrade NIRC2's high-contrast sensitivity at Lp and Ms: the vortex coronagraph's throughput loss and self-emission, sub-30-second sky background drift, and a derotator-generated rotating background pattern. The paper quantifies the coronagraph's throughput and background ratios, documents background drift with sub-second coadds, and maps the derotator pattern. These findings explain part of the discrepancy between predicted and on-sky 5-sigma limits, which is 2.0 magnitudes in Lp and 1.7 magnitudes in Ms.","pith_inferences":["If the Ms throughput loss is caused by diamond absorption near 4.5–5 microns, other diamond-substrate coronagraphs on future ELT instruments may face the same wavelength-dependent penalty, so cold stops and coating choices should be tested at Ms before design freeze.","The sub-30-second sky drift could be exploited by a real-time background monitor: using the new high-speed coadds, observers could fit and subtract a time-varying sky level rather than a single median, potentially recovering a large fraction of the lost sensitivity.","The derotator's rotating pattern could be removed in post-processing if the derotator angle is recorded per frame: a de-rotation by half the drive angle before PCA would turn the pattern static, unlike current pipelines that assume a fixed sky.","A dedicated flat field taken with the vortex in place, combined with the derotator map, could correct the spatial non-uniformity, but the vortex center glow makes a simple flat inadequate; a two-step calibration might work."],"forward_implications":["Using the vortexlm in Ms costs more than in Lp: throughput drops from 82% to 57% while background flux ratios worsen, so Ms observations should budget for a larger sensitivity penalty.","Reducing exposure times to well under 30 seconds (ideally under 1 second) should improve sky-background subtraction because the sky drifts at those timescales and read noise is only 11–14 photons per pixel.","The derotator's rotating background pattern will not be removed by standard PCA sky subtraction, so new calibration or cleaning procedures are needed.","Regular cleaning of the derotator and vortex optic, plus measuring background before and after cleaning, should quantify and reduce the dust-related non-uniformity.","The vortex center glow adds 0.27–0.65 magnitudes of sensitivity loss at 2 lambda/D separations, affecting small inner working angle science."],"supporting_citations":[{"why":"Provides one of the published Ms-filter contrast curves whose plateau defines the on-sky 5-sigma background limit.","marker":"5"},{"why":"Supplies the other Ms contrast curve and modified signal-to-noise equations used with the vortexlm.","marker":"6"},{"why":"VICO contrast predictions give the Lp background limit used to compute the sensitivity discrepancy.","marker":"13"},{"why":"Documents diamond absorption overlapping the Ms bandpass, explaining the wavelength-dependent throughput loss.","marker":"14"},{"why":"Laboratory throughput measurements of the vortexlm mask serve as the baseline for the on-sky comparison.","marker":"15"},{"why":"Describes the NIRC2 electronics upgrade that enabled sub-second coadd sampling.","marker":"10"},{"why":"Models the vortex center glow, used to estimate its impact on sensitivity at small separations.","marker":"17"},{"why":"The PCA-based background subtraction method that fails on the rotating derotator pattern.","marker":"19"}],"fun_headline_variants":["Three on-sky effects cut Keck NIRC2 high-contrast sensitivity","Sub-30s sky drift, coronagraph loss, derotator pattern: Keck IR limits","Keck NIRC2 contrast gap from three on-sky culprits: coronagraph, sky, derotator","Three instrument effects explain Keck NIRC2's 2-mag contrast shortfall"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The in/out comparison is assumed to isolate the coronagraph's effect, with all other optics, alignment, and readout settings held fixed; if those shifted between frames, the reported throughput and background ratios would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Three on-sky effects cut Keck NIRC2 high-contrast sensitivity","Sub-30s sky drift, coronagraph loss, derotator pattern: Keck IR limits","Keck NIRC2 contrast gap from three on-sky culprits: coronagraph, sky, derotator","Three instrument effects explain Keck NIRC2's 2-mag contrast shortfall"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00101,"raw_usage":{"total_tokens":4247,"prompt_tokens":902,"completion_tokens":3345,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":518,"completion_tokens_details":{"reasoning_tokens":3246}},"tokens_in":518,"tokens_out":3345,"duration_ms":22204,"temperature":1.0,"reasoning_tokens":3246,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:11:58.280840+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A bench measurement of the same vortex mask at 4.6 microns under a controlled source giving about 70% throughput would show the on-sky Ms deficit (57%) comes from telescope-state differences rather than the optic itself; alternatively, repeating the on-sky in/out test while dithering the derotator would separate the rotator pattern from the coronagraph's added background.","supporting_citations":[{"cited_title":"Constraining the Orbit and Mass of epsilon Eridani b with Radial Velocities, Hipparcos IAD-Gaia DR2 Astrometry, and Multiepoch Vortex Coronagraphy Upper Limits,","cited_arxiv_id":null,"evidence_quote":"Provides one of the published Ms-filter contrast curves whose plateau defines the on-sky 5-sigma background limit."},{"cited_title":"A Wolf 359 in Sheep’s Clothing: Hunting for Substellar Companions in the Fifth-closest System Using Combined High-contrast Imaging and Radial Velocity Analysis,","cited_arxiv_id":null,"evidence_quote":"Supplies the other Ms contrast curve and modified signal-to-noise equations used with the vortexlm."},{"cited_title":"Characterizing the Performance of the NIRC2 Vortex Coronagraph at W. M. Keck Observatory,","cited_arxiv_id":null,"evidence_quote":"VICO contrast predictions give the Lp background limit used to compute the sensitivity discrepancy."},{"cited_title":"Intrinsic and extrinsic absorp- tion of chemical vapor deposition single-crystal diamond from the middle ultraviolet to the far infrared,","cited_arxiv_id":null,"evidence_quote":"Documents diamond absorption overlapping the Ms bandpass, explaining the wavelength-dependent throughput loss."},{"cited_title":"L- and M-band annular groove phase mask in lab performance assessment on the vortex optical demonstrator for coronagraphic applications,","cited_arxiv_id":null,"evidence_quote":"Laboratory throughput measurements of the vortexlm mask serve as the baseline for the on-sky comparison."},{"cited_title":"Upgrades to W. M. Keck observatory detector systems,","cited_arxiv_id":null,"evidence_quote":"Describes the NIRC2 electronics upgrade that enabled sub-second coadd sampling."},{"cited_title":"Modeling the vortex center glow in the ELT/METIS vortex coronagraph,","cited_arxiv_id":null,"evidence_quote":"Models the vortex center glow, used to estimate its impact on sensitivity at small separations."},{"cited_title":"PCA-based approach for subtracting thermal background emission in high-contrast imaging data,","cited_arxiv_id":null,"evidence_quote":"The PCA-based background subtraction method that fails on the rotating derotator pattern."}],"review_version":2}