{"id":"bbc3186c-46c2-40cd-a206-288daa7c688d","arxiv_id":"2501.03897","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Segment-to-segment coating thickness variations on TMT, ELT, and GMT add only about 1e-8 raw contrast variation in high-contrast coronagraphy, negligible compared to nominal polarization aberrations and AO residuals.","lead":"This paper simulates how small thickness differences in the mirror coatings of the planned TMT, ELT, and GMT telescopes affect their ability to image planets and measure polarized light. It finds that these coating variations add only about one part in a hundred million to the coronagraphic background, far below other error sources, so they are not a major concern for instrument design.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Polarimetric half of the central claim is not isolated: Section 5 compares the varying-coating simulation to the input disk model, not to a uniform-coating baseline, so 'negligible above and beyond nominal' is not demonstrated.","rationale":"The paper's central claim has two legs: coronagraphic contrast and polarimetric performance. The coronagraphic leg is well supported: 25 random realizations per case, explicit subtraction of the uniform-coating image, multiple perfect-coronagraph orders, and a worst-case RMS of 4.3e-8 in Appendix A show the segment-to-segment contribution is orders of magnitude below AO residuals. I also credit the released code and transparent parameterization of amplitudes. The polarimetric leg, however, is not supported by the reported experiment because the observable in Section 5 is never differenced against a uniform-coating baseline. The paper's own language moves from 'difference ... with respect to the disk model' (Fig. 11) to the abstract's 'above and beyond nominal polarization aberration' without the controlled comparison that would justify that inference. This is an internal gap rather than a debate over external assumptions. The spatial-structure uncertainty flagged by the reader is real and acknowledged in Section 6 point 3, but the missing baseline is more load-bearing for the stated claim: it can be settled by a rerun of the existing pipeline, and it directly gates the polarimetric conclusion. The reader's verdict of CONDITIONAL remains appropriate; the condition should include demonstrating the polarimetric baseline, not only validating coating maps.","tokens_in":15290,"tokens_out":5084,"duration_ms":51476,"concrete_test":"Re-run Section 5's debris-disk simulation with the pristine uniform-coating Jones pupil from Anche et al. (2023), using identical Mcoro library sampling, AO phase correction, disk model, and normalization. Compute Delta_s_var = s_with_variations - smodel and Delta_s_nom = s_uniform - smodel. If the norm of (Delta_s_var - Delta_s_nom) is below the segment-induced signal level (e.g., below ~1e-3 in q/u across the disk), the polarimetric conclusion stands; if the difference is comparable to Delta_s_nom, the claim that segment-to-segment variations are negligible for polarimetry is unsupported and the abstract/summary should be revised to report only total polarization aberration effects.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5 builds Mcoro from the Jones pupil that includes the segment-to-segment variations, propagates the disk model through it (Eq. 5), and then reports the difference between the normalized output sconv and the input smodel (Eq. 6, Fig. 11). That difference contains the full nominal polarization aberration content (diattenuation, retardance, crosstalk) already present in Anche et al. (2023), so the 0.1-0.3 level features in Fig. 11 cannot be attributed to segment-to-segment variations. The abstract and Summary point 7 claim the segment variations are negligible 'above and beyond' nominal polarization aberration, but no uniform-coating Mcoro_uniform baseline is computed or subtracted anywhere in the polarimetric analysis. The coronagraphic section does this subtraction explicitly (Section 4, Fig. 5), which highlights the omission. As reported, the polarimetric experiment can only show that total GSMT polarization aberrations affect disk polarimetry, not that the segment-to-segment component is small.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper extends the authors' previous polarization aberration modeling of the three GSMTs (TMT, ELT, GMT) by adding segment-to-segment coating thickness variations. Using the Poke polarization ray tracing package to compute Jones pupils with spatially varying coatings, and HCIPy to propagate them through perfect coronagraph models, the authors simulate the impact on high-contrast imaging residuals and on polarimetric imaging of a debris disk. They report that the coronagraphic contrast variation due to segment variations is at or below roughly 2e-8 to 4e-8 in I-band, orders of magnitude below the AO residual targets, and conclude that segment-to-segment coating variations are not a substantive error term for high-contrast imaging or polarimetry.","tokens_in":15541,"tokens_out":3930,"duration_ms":33919,"significance":"If the conclusions hold, this is a valuable result for the design of GSMT high-contrast instruments: it removes coating nonuniformity from the critical error budget, provided the assumed spatial structure is realistic. The paper's strengths include a forward-modeling pipeline built on open-source packages (Poke, HCIPy), a released code repository (Ashcraft 2024), explicit statistical sampling over 25 trials per case and three amplitude cases, and a clear subtraction of the uniform-coating baseline in the coronagraphic analysis (Section 4). The coronagraphic claim is well supported by the presented experiments. The polarimetric claim, however, is not yet supported by the analysis as written, because the experiment does not isolate the segment-to-segment contribution from the nominal polarization aberrations.","major_comments":[{"comment":"The polarimetric experiment does not isolate the effect of segment-to-segment variations. The comparison is between Sconv (which includes all nominal polarization aberrations plus segment variations) and Smodel (the input disk model). The abstract and Summary point 7 claim that segment-to-segment variations are negligible 'above and beyond the impact of nominal polarization aberration,' but no baseline Mcoro computed with a perfectly uniform coating is presented or subtracted anywhere in Section 5. Without that baseline, the 0.1-0.3 level features in Fig. 11 could be dominated by the nominal polarization aberrations already reported in Anche et al. (2023), and the segment-to-segment component remains unquantified. I recommend computing Mcoro for the uniform-coating case and subtracting it (or directly differencing the segment-varying and uniform cases) to support the 'above and beyond' claim.","section":"Section 5, Eqs. (5)-(6), Fig. 11"},{"comment":"The conclusion that segment-to-segment variations are negligible rests on the assumed spatial structure of the coating variations: low-order Zernike piston/tilt/focus maps for TMT/ELT and a power-law PSD with negative index for the GMT's Al2O3 layer. The paper acknowledges in Section 6 point 3 that actual meter-scale segment coating maps are unmeasured, yet the parameter space explored does not include high-order spatial-frequency ripple or large segment-to-segment steps at high spatial frequencies. Under such structures, the contrast contribution could scale differently with coronagraph order and could potentially exceed the reported values. The summary claim in Section 6 point 7 ('does not contribute a substantive error term') is therefore conditional on the low-order spatial structure assumption; the paper should either test realizations with high-order spatial content or temper the summary claim to explicitly state this dependence.","section":"Section 3 (modeling approach), Table 1, Section 6 point 3"}],"minor_comments":[{"comment":"The instrument name 'SPHERE/IRIDIS' appears in the Introduction and 'SPHERE-IRIDIS' in Section 5; the correct name is 'SPHERE-IRDIS' (see also the references). Please correct throughout.","section":"Sections 1 and 5"},{"comment":"The reference list contains apparent duplicates: 'van Holstein, R. G., Girard, J. H., de Boer, J., et al. 2020, A&A, 633, A64' appears twice, and 'van Holstein et al. 2023a' and '2023b' both list A&A 677, A150 with the same title and nearly identical author lists. Please merge or correct these entries.","section":"References"},{"comment":"The field-dependent transform in Eq. (5) writes Sconv(x,y) as a sum of Mcoro, j,k(x,y;θj,k) Smodel(x,y), but the notation is ambiguous about how the focal-plane coordinates (x,y) relate to the field positions θj,k. Please clarify that each Mcoro is a spatially-varying Mueller matrix evaluated at the focal-plane coordinate for a given field angle, and specify the interpolation or summation convention.","section":"Section 5, Eq. (5)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for A&A as a modeling paper. The coronagraphic result appears well-supported and is a useful contribution. The polarimetric claim is fixable by adding a uniform-coating baseline subtraction, and the spatial-structure limitation could be addressed either by additional simulations or by tempering the summary wording. I would not reject; the requested changes are substantive but localized."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the coronagraphic answer is solid, and the polarimetric answer is probably right but not actually demonstrated. Worth engaging with seriously, with one clear fix requested.\n\nThe paper extends Paper I by adding spatially varying segment-to-segment coating thicknesses to the Jones-pupil models of TMT, ELT, and GMT, then propagates those through perfect coronagraphs with HCIPy. The coronagraphic analysis is done cleanly: they subtract the uniform-coating coronagraphic image from the varying-coating image and report the differential contrast, which peaks around 2e-8 in I-band for the worst case. That is a legitimate forward simulation, with code released, and it supports the claim that coating nonuniformity is not a driver for high-contrast imaging on these telescopes. Credit where due: the pipeline is a real step toward higher-fidelity error budgets.\n\nThe soft spot is the polarimetric section. The text and abstract claim that segment-to-segment variations are negligible \"above and beyond\" nominal polarization aberrations. But the difference images in Figure 11 are computed by propagating the disk model through a Mcoro that includes both the nominal aberrations and the segment variations, then subtracting the input disk model. There is no uniform-coating baseline subtracted. So the 0.1-0.3 level features shown are mostly the nominal polarization aberrations from Paper I, not the additional contribution from segment-to-segment variations. The conclusion may well be true, but the experiment as reported does not isolate it. This is a moderate gap, not a fatal one, and it can be fixed by adding a baseline run without the segment variations, exactly as Section 4 does for coronagraphy.\n\nThe second caveat is the assumed spatial structure of the coating variations. As the authors admit, actual meter-scale segment coating maps are unmeasured; they use low-order Zernike maps for TMT/ELT and a power-law PSD for GMT, scaled from a 4 cm witness sample. They test plausible amplitudes and the conclusion appears robust to amplitude, but if real variations have high spatial-frequency ripple or large segment-to-segment steps, the contrast contribution could scale differently. That is a caveat, not a killer.\n\nOverall, the paper is honest, clearly written, and reproducible. The coronagraphic result is well-supported; the polarimetric claim needs a baseline subtraction before it earns the current wording. I would send this to a serious referee. The reading group might enjoy the discussion of how to isolate effects in forward simulations, but it is not a must-read.","headline":"Good forward simulation of coating variations; the coronagraphic conclusion holds, but the polarimetric claim needs a uniform-coating baseline before it is demonstrated.","tokens_in":16097,"tokens_out":1916,"would_cite":true,"duration_ms":19283,"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":"Segment-to-segment coating thickness differences do not add a substantive error to the GSMTs' high-contrast imaging or polarimetry, adding at most about $2\\times10^{-8}$ in I-band contrast.","keywords":["polarization aberrations","segment-to-segment coating variations","giant segmented mirror telescopes","high-contrast imaging","coronagraphy","polarimetry","Jones pupil","Mueller point-response matrix"],"falsifier":"Measure coating thickness maps across full-size segments of a GSMT primary (for example with ellipsometry or interferometry on witness segments) and run those maps through the same Jones-pupil and coronagraph pipeline; if the RMS contrast residual in I-band for a second-order perfect coronagraph exceeds about $1\\times10^{-7}$, the paper's central conclusion would be overturned.","tokens_in":15047,"feed_emoji":"🔭","tokens_out":10030,"duration_ms":82492,"temperature":0.7,"pith_summary":"Direct imaging of Earth-like exoplanets from the ground demands contrasts near $10^{-7}$, but the next-generation giant segmented telescopes (TMT, ELT, GMT) already carry polarization aberrations that limit them to $10^{-5}$–$10^{-6}$ in the infrared. This paper asks whether the next level of realism—coating thickness differences from segment to segment—makes that limit worse. Using polarization ray tracing to build Jones pupils with spatially varying coatings and propagating them through perfect coronagraph models, the paper finds that segment-to-segment variations add at most about $2\\times10^{-8}$ RMS contrast in I-band for the most sensitive coronagraph and worst-case telescope, with worst-case peaks near $4.3\\times10^{-8}$. That is 2–3 orders of magnitude below the target contrast and well below the adaptive-optics residual floor, so coating nonuniformity does not add a substantive error term to high-contrast imaging or to polarimetric imaging of debris disks.","feed_headline":"Uneven mirror coatings won't derail exoplanet imaging","feed_subtitle":"New simulations show segment-to-segment coating variations add under 2e-8 contrast, far below adaptive-optics noise.","key_machinery":"The argument is carried by the Jones pupil—the $2\\times2$ complex polarization response map across the exit pupil—computed with polarization ray tracing on telescope models whose segment overcoat thickness varies as low-order Zernike polynomials (TMT, ELT) or a power-law PSD (GMT). These pupils are fed through ideal 'perfect coronagraph' models (order 2, 4, 6) that remove spatial modes of the electric field; comparing the residual image to the uniform-coating case isolates the segment-variation contribution. For polarimetry, the amplitude response matrix $A_{\\mathrm{coro}}$ is converted into a Mueller point-response matrix $M_{\\mathrm{coro}} = U(A_{\\mathrm{coro}}\\otimes A_{\\mathrm{coro}})U^{-1}$, which maps the incoming Stokes vector to the coronagraphic focal-plane Stokes image and is used to propagate a debris disk model through the field.","core_discovery":"The central discovery is that spatially varying coating thickness across the segmented primaries of TMT, ELT, and GMT adds only a minor perturbation to the polarization aberrations those telescopes already produce. Modeling the ELT/TMT overcoat as low-order Zernike piston/tilt/focus variations of 10–50% peak-to-valley and the GMT oxide layer as a power-law PSD with $\\pm0.08$ nm peak-to-valley, the paper simulates 25 random realizations per case, applies wavefront control from an ideal AO system, and subtracts the uniform-coating coronagraphic image. The RMS contrast variation peaks at $1.7\\times10^{-8}$ for TMT behind a second-order perfect coronagraph in I-band (worst case $4.3\\times10^{-8}$), and order-6 coronagraphs fall below $10^{-10}$. In polarimetry, the Mueller point-response matrix shows polarized structure near the inner working angle, but when a debris disk model is propagated through it, the changes in normalized Stokes parameters are dominated by the telescopes' nominal instrumental polarization and crosstalk, not by segment-to-segment variations. The paper concludes that coating thickness nonuniformity is not a substantive error term for high-contrast detection or polarimetry on these observatories.","pith_inferences":["A consequence the authors leave implicit: if real segment coatings have high-order thickness ripple or sharp steps between segments rather than the assumed low-order shapes, the contrast contribution could be larger than $2\\times10^{-8}$, so measured meter-scale coating maps would settle the margin.","The same Jones-pupil and Mueller-matrix pipeline could set segment-coating tolerance specifications for a future space observatory aiming at $10^{-10}$ contrast.","The polarimetry finding suggests that calibration of the telescope's static Mueller matrix will buy more disk-science accuracy than tightening coating uniformity.","A testable extension is to inject thickness discontinuities at segment boundaries into the same simulation and check whether the contrast residual stays below the AO-limited floor."],"forward_implications":["Segment-to-segment coating thickness variations can be dropped from the first-order error budget for GSMT high-contrast imaging in I-band, because the added residuals are orders of magnitude below the AO-limited floor.","Coating uniformity requirements for TMT, ELT, and GMT primary segments need not be driven by polarization aberration concerns, potentially relaxing coating tolerances.","Debris disk polarimetry on these telescopes will be limited by the calibrated Mueller matrix of the telescope (instrumental polarization and crosstalk), not by segment-level coating variations.","For a future space observatory aiming at $10^{-10}$ contrast, the same segment-variation analysis should be repeated, since the residuals found here are still orders of magnitude above that target.","Higher-order coronagraphs (6th order) suppress the segment-variation residual below $10^{-10}$, so they are insensitive to coating nonuniformity."],"supporting_citations":[{"why":"Prior paper in the series; supplies the nominal GSMT Jones pupils and polarization aberration baseline that this study extends with segment-to-segment variations.","marker":"Anche et al. (2023)"},{"why":"Gemini witness-sample ellipsometry; provides the 15% (1.3 nm) peak-to-valley coating thickness variation amplitude used to scale the ELT/TMT models.","marker":"Schneider et al. (2016b)"},{"why":"Measurement of the Al2O3 layer on aluminum; anchors the GMT oxide-layer thickness and peak-to-valley uncertainty.","marker":"van Harten et al. (2009)"},{"why":"Private communication on TMT coating plans; source of the 10–20% expected coating thickness variation for the TMT cases.","marker":"Skidmore et al. (2023)"},{"why":"Perfect coronagraph model used to compute the coronagraphic residuals and contrast from polarization aberrations.","marker":"Guyon et al. (2006)"},{"why":"Companion perfect coronagraph model used alongside Guyon to define the order-2/4/6 coronagraph response.","marker":"Cavarroc et al. (2006)"},{"why":"Physical optics propagation package used to simulate the diffraction response of the coronagraphs and build the Mueller point-response matrices.","marker":"Por et al. (2018)"}],"fun_headline_variants":["Coating flaws barely dent GSMT coronagraph contrast","Segment coating variations: negligible for exo-Earth imaging","Coating nonuniformity won't limit GSMT high-contrast imaging","Coating variations add tiny contrast noise for GSMTs","Segment coating flaws: negligible for planet imaging"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the assumed spatial structure of coating thickness variations—low-order Zernike shapes for TMT/ELT and a power-law PSD for the GMT oxide layer, with amplitudes scaled from small witness samples—represents what real meter-scale segment coatings do; actual segment coating maps have not been measured.","fun_headline_variants_meta":{"raw":{"variants":["Coating flaws barely dent GSMT coronagraph contrast","Segment coating variations: negligible for exo-Earth imaging","Coating nonuniformity won't limit GSMT high-contrast imaging","Coating variations add tiny contrast noise for GSMTs","Segment coating flaws: negligible for planet imaging"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00092,"raw_usage":{"total_tokens":4023,"prompt_tokens":1096,"completion_tokens":2927,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":712,"completion_tokens_details":{"reasoning_tokens":2846}},"tokens_in":712,"tokens_out":2927,"duration_ms":21909,"temperature":1.0,"reasoning_tokens":2846,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:44:40.879290+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure coating thickness maps across full-size segments of a GSMT primary (for example with ellipsometry or interferometry on witness segments) and run those maps through the same Jones-pupil and coronagraph pipeline; if the RMS contrast residual in I-band for a second-order perfect coronagraph exceeds about $1\\times10^{-7}$, the paper's central conclusion would be overturned.","supporting_citations":[{"cited_title":"M., Ashcraft, J","cited_arxiv_id":null,"evidence_quote":"Prior paper in the series; supplies the nominal GSMT Jones pupils and polarization aberration baseline that this study extends with segment-to-segment variations."},{"cited_title":"S., Gallagher, B., & Hansen, E","cited_arxiv_id":null,"evidence_quote":"Private communication on TMT coating plans; source of the 10–20% expected coating thickness variation for the TMT cases."},{"cited_title":"A., Kuchner, M","cited_arxiv_id":null,"evidence_quote":"Perfect coronagraph model used to compute the coronagraphic residuals and contrast from polarization aberrations."},{"cited_title":"H., Haffert, S","cited_arxiv_id":null,"evidence_quote":"Physical optics propagation package used to simulate the diffraction response of the coronagraphs and build the Mueller point-response matrices."}],"review_version":1}