{"id":"94ec86f6-914c-4adb-b21d-373054cc288a","arxiv_id":"1908.02585","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A new coronagraph design, the phase-apodized Lyot coronagraph, can block starlight deeply while keeping high throughput on complex telescope pupils.","lead":"This paper shows that combining a phase-shaping glass plate with a Lyot-stop coronagraph can block starlight more efficiently and let telescopes see planets closer to their host star. It matters because such designs could improve future space missions and ground-based instruments looking for Earth-like planets.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline 1.4λ/D at 10^-10 rests on a monochromatic one-sided parameter study (§4.1); the proposed achromatization (§4.2) adds only a tilt and does not account for the physical phase plate's (λ0/λ) phase scaling, so broadband 10^-10 contrast is unverified and likely not achieved.","rationale":"The concern is load-bearing because the abstract's numbers (1.4λ/D, 10^-10, >75%) are all supplied by Figure 6, and Figure 6 is explicitly monochromatic. A coronagraph for a real instrument must work over a band; a phase plate is a dispersive optical element, so the monochromatic optimum cannot be assumed to remain optimal or even meet contrast at other wavelengths. The proposed achromatization in Section 4.2 is not a proof: it addresses only the position of the PSF relative to the knife edge, not the wavelength-dependent shape of the phase-generated dark zone. This is a correctness risk, not merely a missing demonstration, because the physical phase term (λ0/λ)φ0 enters the Fourier propagation at every wavelength. The reader's verdict was CONDITIONAL and I agree with the conditional recommendation; I do not think the paper needs to be rejected, since the method and monochromatic results appear sound and the limitations are partly acknowledged. However, the abstract should either state the monochromatic basis or the broadband claim must be supported by an actual optimization over the spectral band. The phase-only relaxation is a related concern, but I view the broadband transfer as the decisive one: even a perfectly phase-only design would still face the chromatic issue.","tokens_in":16881,"tokens_out":13831,"duration_ms":162556,"concrete_test":"Re-evaluate the LUVOIR-A one-sided design (Section 5.2) with a physical phase plate: set the apodizer phase at wavelength λ to (λ0/λ)φ0(x), center the knife-edge mask (fedge=0), add the tilt prescribed in Section 4.2, and propagate λ0, 0.9λ0, and 1.1λ0 using the same focal-plane and Lyot masks. Compute Cdesign over the D-shaped dark zone per Eq. 3 and IWA per Eq. 4 for the 20% band. If Cdesign > 10^-10 or IWA > 2.2λ/D, the 'completely achromatic' statement fails and the abstract's broadband extrapolation is unsupported; if it stays below, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim as stated in the abstract is that knife-edge PAPLCs reach 1.4λ/D at 10^-10 with >75% throughput. The numbers in Figure 6 come from a parameter study that Section 4.1 says is monochromatic: 'All masks were calculated for a single wavelength only.' Section 4.2 tries to lift this to broadband by centering the knife-edge mask and adding a phase tilt, arguing that the PSF scales with wavelength and the mask edge stays fixed relative to the rescaled PSF. That argument ignores the chromaticity of the phase plate itself: a plate fabricated for λ0 has phase (λ0/λ)φ0(x) at wavelength λ, so the dark-zone-generating phase structure and the tilt both rescale. The paper never evaluates Cdesign of Eq. 3 over λ0±Δλ for a one-sided design, and Figure 7's broadband panel is qualitative, not a 10^-10 contrast verification. A secondary, related gap is Section 2.2's assertion that the relaxed |X+iY|≤1 solutions are phase-only; no proof or amplitude map is provided, and the VLT/SPHERE design blocks dead-actuator light 'at the apodizer' (Section 5.1), an amplitude operation. Both gaps make the headline performance hard to translate into a physical broadband phase-only coronagraph.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces a convex optimization framework for designing phase-apodized-pupil Lyot coronagraphs (PAPLCs), in which a phase-only pre-apodizer is paired with a focal-plane mask and a Lyot stop. The method is applied to two configurations: annular focal-plane masks producing point-symmetric dark zones, and knife-edge masks producing one-sided dark zones. Parameter studies with simplified circular obstructed pupils show that point-symmetric PAPLCs perform comparably to APLCs, while one-sided designs achieve smaller inner working angles (down to 1.4 lambda/D) and higher throughput (up to about 75-80%) at design contrasts down to 10^-10, based on monochromatic optimizations. Case studies for VLT/SPHERE and LUVOIR-A demonstrate the approach on segmented, strutted, and non-circular pupils. Section 4.2 argues that the one-sided designs can be made achromatic by centering the knife-edge mask and adding a phase tilt, and Section 5.3 evaluates tip-tilt robustness. The paper also states that the relaxed optimization problem from phase-only to sub-unity complex amplitude in practice returns phase-only solutions, and that PAPLCs always perform as well as or better than APLCs.","tokens_in":17254,"tokens_out":9769,"duration_ms":100895,"significance":"If the headline performance holds under broadband illumination and with physically realizable phase-only apodizers, the knife-edge PAPLC would be an important new coronagraph option for high-contrast imaging on obscured and segmented apertures, offering substantially better throughput and smaller inner working angles than APLCs. The convex optimization methodology is clearly presented, builds on prior work by the same author, and the use of the open-source HCIPy package supports reproducibility. The point-symmetric PAPLC analysis and the VLT/SPHERE and LUVOIR-A case studies are valuable contributions. However, the central claims currently rest on two unverified assumptions: that the relaxed optimization solutions are exactly phase-only, and that the tilt-based achromatization actually delivers the advertised broadband contrast. These gaps make the significance contingent on additional verification.","major_comments":[{"comment":"The relaxation from the phase-only constraint X^2+Y^2=1 to the inequality X^2+Y^2<=1 is asserted to produce phase-only solutions in practice, but no evidence is provided: there are no amplitude maps, histograms of |X+iY|, or statistics across the parameter studies. This is load-bearing because the physical coronagraph uses a phase-only apodizer; if any optimized solution has regions with |X+iY|<1, the phase-only projection is not guaranteed to meet the contrast and throughput claims. Moreover, the VLT/SPHERE design in Section 5.1 explicitly blocks light at dead deformable-mirror actuator locations 'at the apodizer,' which is an amplitude operation and directly contradicts the phase-only assertion. The authors should demonstrate that all final designs (including the VLT/SPHERE case) saturate the norm constraint, or they should treat amplitude as part of the physical model and re-derive the affected performance numbers.","section":"Section 2.2, Eqs. (10)-(11)"},{"comment":"The headline values in the abstract - inner working angle as close as 1.4 lambda/D at contrasts of 10^-10 and maximum post-coronagraphic throughput of >75% - come from the parameter study in Section 4.1, which states explicitly that 'All masks were calculated for a single wavelength only: we presume monochromatic light.' The abstract and conclusions present these numbers without the monochromatic qualifier. Section 4.2's achromatization argument is qualitative and does not quantitatively account for the chromatic scaling of the phase plate's non-tilt phase components; no evaluation of Cdesign as defined in Eq. (3) over a finite band is presented for any one-sided design. The authors should either run a broadband end-to-end simulation of the 1.4 lambda/D design over the intended band (e.g., 10% or 20% bandwidth) and report the achieved Cdesign, or clearly state in the abstract that the demonstrated 10^-10 contrast is monochromatic and that broadband achromatization is a proposal that remains to be verified.","section":"Section 4.1 and Abstract"},{"comment":"The statement that 'a PAPLC will always perform the same or better than an APLC for a given telescope pupil, dark zone geometry and design contrast' is not justified as a universal claim. The APLC solution space is indeed a subspace of the relaxed complex-amplitude problem, but the actual PAPLC apodizer is constrained to be phase-only. If the relaxed optimum has |X+iY|<1, the phase-only projection may have lower throughput or worse contrast. The empirical comparisons in Sections 3 and 4 are reasonable evidence for the specific cases studied, but the unconditional 'always' wording should be qualified by the tightness of the relaxation, or replaced by a statement about the specific designs presented.","section":"Section 2.2, paragraph after Eq. (11)"},{"comment":"The comparison between PAPLC and APLC throughput in Table 1 is a central demonstration of the PAPLC's advantage, but the APLC designs are described only as a 'preliminary solution' for VLT/SPHERE and 'a part of a coronagraph design study' for LUVOIR-A, courtesy of other groups. These APLC designs may not have been optimized with the same hyperparameter search (focal-plane mask size, Lyot stop geometry) used for the PAPLCs. To make the factor-of-two-to-three throughput comparison fair, the authors should either optimize the APLC baselines with the same procedure described in Section 3, or provide the design parameters and optimization effort for both families so that the reader can judge the comparison.","section":"Table 1 and Section 5.3"}],"minor_comments":[{"comment":"In Eq. (8), 'X(x) + iY(y)' should presumably read 'X(x) + iY(x)'.","section":"Eq. (8)"},{"comment":"The phrase 'the solutions space for APLCs' contains a typo; it should be 'the solution space for APLCs'.","section":"Section 2.2"},{"comment":"In the rendered figure, the y-axis label 'Throughput' is missing from the CO=20% and CO=30% panels; please ensure all panels have axis labels.","section":"Figure 6"},{"comment":"The sentence 'This makes the one-sided PAPLC completely achromatic in theory (barring experimental effects)' is too strong given that the chromaticity of the phase plate's non-tilt phase components is not analyzed. A more cautious wording such as 'approximately achromatic to first order, subject to verification' would better match the evidence presented.","section":"Section 4.2"},{"comment":"For reproducibility, the authors should report numerical details such as the pupil grid size, the number of optimization variables, and the solver settings (e.g., Gurobi version and optimality tolerances).","section":"Appendix A and Section 2.1"}],"recommendation":"major_revision","confidential_remarks":"This is an interesting design paper with a sound convex-optimization core, and the case studies are useful. The two load-bearing gaps - the unproven phase-only tightness of the relaxation and the unverified broadband achromatization - are addressable with additional simulations and should be fixed before the paper is accepted. I do not see an inherent contradiction that would force rejection; the issues are about completing the verification of the central claims. The author is an early-career researcher, and the referee comments should be framed constructively to help strengthen the work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Emiel Por's PAPLC paper is a serious piece of design work. The architecture is new: combine an APP-style phase apodizer with a Lyot stop and optimize the phase with a convex program. The one-sided knife-edge case is the real contribution, with parameter studies showing ~1.4λ/D IWA at 10^-10 contrast on obstructed pupils, and the VLT/SPHERE and LUVOIR-A case studies are genuinely impressive. The point-symmetric case is basically a phase-only APLC, and the paper says so; that negative result is honest and useful. The optimization problem is clearly posed, the linearization of contrast constraints is standard, and the use of HCIPy means the designs are reproducible. The author also states plainly that the designs are not robust to aberrations or Lyot-stop misalignment.\n\nTwo soft spots matter, and one is load-bearing. First, the headline 1.4λ/D at 10^-10 comes from a parameter study that is explicitly monochromatic (Section 4.1). The achromatization in Section 4.2 argues that adding a tilt to the phase plate makes the design achromatic, but that ignores the physical plate's chromaticity: a plate made for λ0 has phase (λ0/λ)φ0 at wavelength λ. A tilt then produces a physical shift that is independent of λ, not one that scales with the PSF size. So the mask edge moves relative to the rescaled PSF, and 10^-10 contrast is not established over band. Figure 7's broadband panel is qualitative, not a contrast curve. If the author can demonstrate broadband 10^-10 with the actual chromatic phase, fine; as written, the abstract overstates the result.\n\nSecond, the convex relaxation to |X+iY|≤1 is asserted to yield phase-only solutions, but no proof or amplitude maps are shown. The VLT/SPHERE design's statement that dead-actuator light is 'blocked at the apodizer' sounds like an amplitude operation. If the optimized solutions are not phase-only, the comparison to APLCs is not apples-to-apples and the 'always beats APLC' argument collapses. This should be easy to check and needs to be in the paper.\n\nThe circularity note about design contrast being a constraint is fair but minor. The citation pattern looks fine, and HCIPy use is reproducible.\n\nRecommendation: send to peer review. The method is novel, the parameter study useful, the case studies valuable, but the paper needs a revision that presents broadband contrast verification with a real chromatic phase plate (or weakens the abstract) and proves or demonstrates phase-only convergence, e.g., by showing amplitude maps and the performance of a projected phase-only solution.","headline":"A promising coronagraph design method with an honest parameter study, but the headline broadband performance rests on a suspect achromatization argument and an unproven phase-only relaxation.","tokens_in":17788,"tokens_out":6410,"would_cite":false,"duration_ms":63854,"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":"Phase-only aperture apodization plus a Lyot stage—the PAPLC—reaches inner working angles of 1.4λ/D at 10⁻¹⁰ contrast with >75% throughput, including on realistic segmented and obstructed telescope pupils up to 30% central obscuration.","keywords":["coronagraphy","high-contrast imaging","exoplanet direct imaging","apodizing phase plate","Lyot coronagraph","convex optimization","phase apodization","knife-edge focal-plane mask"],"falsifier":"Run the paper's relaxed convex optimization problem for the LUVOIR-A pupil and check whether the optimal complex transmission has any region with $|X+iY|<1$; if it does, the physical phase-only apodizer cannot match the optimized contrast. Independently, simulate the broadband LUVOIR-A knife-edge design at 2.2λ/D over a 10% band with the wavelength-dependent tilt applied and look for any dark-zone pixel above 10⁻¹⁰ raw contrast.","tokens_in":16653,"feed_emoji":"🔭","tokens_out":9671,"duration_ms":91962,"temperature":0.7,"pith_summary":"The paper sets out to show that replacing the amplitude apodizer of an apodized-pupil Lyot coronagraph (APLC) with a phase-only apodizer—the phase-apodized-pupil Lyot coronagraph (PAPLC)—yields a coronagraph that matches the APLC for point-symmetric dark zones and substantially outperforms it for one-sided dark zones. The central result is that with a knife-edge focal-plane mask, the PAPLC reaches inner working angles as small as 1.4λ/D at a raw contrast of 10⁻¹⁰ with post-coronagraphic throughput above 75%, for telescope apertures with central obscurations up to 30%. A sympathetic reader would care because deep-contrast exoplanet imaging is currently limited by the trade-off between inner working angle, throughput, and robustness to real telescope pupils; the PAPLC appears to relax that trade-off while still handling struts, segments, and non-circular edges. The paper also presents concrete designs for the VLT/SPHERE instrument and the LUVOIR-A aperture.","feed_headline":"Phase-only Lyot coronagraph reaches 1.4 λ/D at 1e-10 contrast","feed_subtitle":"Knife-edge design keeps over 75% throughput on pupils with up to 30% central obscuration, even segmented ones.","key_machinery":"The load-bearing device is the convex optimization of the complex transmission of the pre-apodizer. The apodizer is written as $A(x)(X(x)+iY(x))$, the phase-only constraint is relaxed to $|X(x)+iY(x)|\\le 1$, and the objective is the real part of the on-axis non-coronagraphic field, which removes the piston degeneracy while keeping the problem convex. Because the electric field in the post-coronagraphic focal plane is a linear map of the apodizer transmission (Fourier propagation through the pupil, knife-edge mask, and Lyot stop), the stellar-leakage contrast constraints become convex inequalities; a linearized tip-tilt constraint prevents the optimizer from sneaking a centroid shift into the phase solution. The same machinery yields the point-symmetric designs as a real-valued special case. For one-sided dark zones, a wavelength-dependent phase tilt added to the apodizer centers the growing PSF on the knife-edge over the band, giving theoretical achromatization.","core_discovery":"The central discovery is that the PAPLC design problem can be made convex by writing the apodizer as a complex transmission $X(x)+iY(x)$ with the phase-only constraint relaxed to $|X(x)+iY(x)|\\le 1$, and that in practice the optimum is still phase-only. The post-coronagraphic electric field is a linear function of that transmission, so the $10^{-c}$ stellar-leakage constraints are convex, and the piston ambiguity is removed by maximizing the real part of the on-axis field. Since the APLC's solution space (real nonnegative apodizer) is a subset of this complex-amplitude space, the PAPLC is guaranteed to match or beat the APLC for the same masks, dark zone, and contrast. The one-sided PAPLC uses an offset knife-edge focal-plane mask; the apodizer creates the one-sided dark zone in the focal plane, and the Lyot stop deepens the contrast by several orders of magnitude, reaching 1.4λ/D at 10⁻¹⁰ contrast and up to about 80% throughput for central obscurations up to 30%. Point-symmetric annular-mask designs are found as a real-valued special case and show only marginal gains over the APLC.","pith_inferences":["Inference (not in the paper): the one-sided dark zone covers about half the field of view, so survey-mode mapping of a full circumstellar disk would require observing at several roll angles, and the throughput penalty of that mode is not included in the headline >75% figure.","Inference (not in the paper): the proposed achromatization assumes the wavelength-dependent tilt is implemented exactly; a laboratory test across a 10–20% band could confirm whether the 10⁻¹⁰ contrast floor survives real dispersion, and would be the natural next experiment.","Inference (not in the paper): because the same convex machinery was used to harden APLCs against low-order aberrations, a similar inclusion of aberration modes should produce robust PAPLC designs; the paper explicitly leaves that to future work."],"forward_implications":["Because the PAPLC design space contains the APLC design space, any PAPLC design is guaranteed to match or beat the best APLC for the same pupil, focal-plane mask, Lyot stop, dark zone, and design contrast.","For one-sided dark zones, the knife-edge PAPLC reaches 1.4λ/D inner working angle at 10⁻¹⁰ contrast with throughput above 75% on pupils with up to 30% central obscuration, a combination not available from the corresponding APLC at these parameters.","The LUVOIR-A knife-edge PAPLC design (IWA 2.2λ/D, maximum throughput 78%) roughly triples the maximum throughput of the APLC comparison design on the same aperture, while cutting the inner working angle by 1.5λ/D.","Alignment of residual atmospheric dispersion along the knife edge makes the coronagraph effectively insensitive to that dispersion, relaxing the atmospheric dispersion corrector requirement to about 1λ/D instead of a few tenths or hundredths of λ/D.","The design method is not restricted to circular or unobstructed pupils: the VLT/SPHERE and LUVOIR-A case studies show it handles struts, segments, dead deformable-mirror actuators, and non-circular outer edges."],"supporting_citations":[{"why":"Supplies the convex-optimization method for phase apodizers that the paper adapts to the Lyot architecture, including the result that a globally optimal complex-amplitude apodizer is phase-only.","marker":"Por (2017)"},{"why":"Defines the APLC optimization problem whose solution space is a subset of the PAPLC problem, providing the direct comparison baseline for point-symmetric designs.","marker":"Zimmerman et al. (2016)"},{"why":"Establishes the apodized-pupil Lyot coronagraph architecture—a pupil apodizer combined with a Lyot stage—that the PAPLC extends to phase-only apodization.","marker":"Soummer (2004)"},{"why":"Introduces the apodizing phase plate concept, the phase-only pupil apodizer that the PAPLC's pre-apodizer is based on.","marker":"Codona et al. (2006)"},{"why":"Provides the aperture-photometry definitions of throughput, raw contrast, and inner working angle used to evaluate all designs.","marker":"Ruane et al. (2018)"},{"why":"Supplies the hybrid shaped-pupil/APLC design procedure used for the VLT/SPHERE and LUVOIR-A comparison APLC designs.","marker":"N’Diaye et al. (2016)"},{"why":"The existing ALC2 Lyot mask of VLT/SPHERE, used as a fixed Lyot-stop hyperparameter in the SPHERE case study.","marker":"Guerri et al. (2011)"},{"why":"Shows that APLCs can be made robust to aberrations by including them in the optimization, which motivates the paper's proposed future robustness extension.","marker":"N’Diaye et al. (2015)"}],"fun_headline_variants":["Phase-only coronagraph: 1.4 λ/D at 1e-10 contrast","Knife-edge PAPLC achieves 1.4 λ/D at 10^-10 contrast","Phase apodizer: >75% throughput even with 30% obscuration","Optimized phase mask reaches 1.4 λ/D at 1e-10 contrast"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The relaxed optimization problem allows the apodizer amplitude to dip below unity, and the paper's results rely on the assertion—not a proof—that every optimal solution is still phase-only, plus the assumption that the monochromatic one-sided designs hold their 10⁻¹⁰ contrast after the proposed achromatization.","fun_headline_variants_meta":{"raw":{"variants":["Phase-only coronagraph: 1.4 λ/D at 1e-10 contrast","Knife-edge PAPLC achieves 1.4 λ/D at 10^-10 contrast","Phase apodizer: >75% throughput even with 30% obscuration","Optimized phase mask reaches 1.4 λ/D at 1e-10 contrast"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001074,"raw_usage":{"total_tokens":4541,"prompt_tokens":1036,"completion_tokens":3505,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":3412}},"tokens_in":652,"tokens_out":3505,"duration_ms":26055,"temperature":1.0,"reasoning_tokens":3412,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:40:27.050632+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the paper's relaxed convex optimization problem for the LUVOIR-A pupil and check whether the optimal complex transmission has any region with $|X+iY|<1$; if it does, the physical phase-only apodizer cannot match the optimized contrast. Independently, simulate the broadband LUVOIR-A knife-edge design at 2.2λ/D over a 10% band with the wavelength-dependent tilt applied and look for any dark-zone pixel above 10⁻¹⁰ raw contrast.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the convex-optimization method for phase apodizers that the paper adapts to the Lyot architecture, including the result that a globally optimal complex-amplitude apodizer is phase-only."},{"cited_title":"T., Eldorado Riggs , A","cited_arxiv_id":null,"evidence_quote":"Defines the APLC optimization problem whose solution space is a subset of the PAPLC problem, providing the direct comparison baseline for point-symmetric designs."},{"cited_title":"2004, The Astrophysical Journal Letters, 618, L161","cited_arxiv_id":null,"evidence_quote":"Establishes the apodized-pupil Lyot coronagraph architecture—a pupil apodizer combined with a Lyot stage—that the PAPLC extends to phase-only apodization."},{"cited_title":"2006, in SPIE Astronomical Telescopes+ Instrumentation, International Society for Optics and Photonics, 62691N--62691N","cited_arxiv_id":null,"evidence_quote":"Introduces the apodizing phase plate concept, the phase-only pupil apodizer that the PAPLC's pre-apodizer is based on."},{"cited_title":"2018, in Space Telescopes and Instrumentation 2018: Optical, Infrared, and Millimeter Wave, Vol","cited_arxiv_id":null,"evidence_quote":"Provides the aperture-photometry definitions of throughput, raw contrast, and inner working angle used to evaluate all designs."},{"cited_title":"2016, , 818, 163","cited_arxiv_id":null,"evidence_quote":"Supplies the hybrid shaped-pupil/APLC design procedure used for the VLT/SPHERE and LUVOIR-A comparison APLC designs."},{"cited_title":"2011, Experimental Astronomy, 30, 59","cited_arxiv_id":null,"evidence_quote":"The existing ALC2 Lyot mask of VLT/SPHERE, used as a fixed Lyot-stop hyperparameter in the SPHERE case study."},{"cited_title":"2015, The Astrophysical Journal, 799, 225","cited_arxiv_id":null,"evidence_quote":"Shows that APLCs can be made robust to aberrations by including them in the optimization, which motivates the paper's proposed future robustness extension."}],"review_version":1}