{"id":"7ebe99de-20e7-41c6-88c9-c8dd8b48fd97","arxiv_id":"2501.01570","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Jewel Optics uses stacked wedged windows to create multiple non-redundant aperture masks in one pupil, boosting throughput by factors of 3 to 7 while preserving non-redundant Fizeau beam combination.","lead":"This paper presents a way to split a telescope pupil into several non-redundant patterns, each deflected to its own camera region by a phase wedge, so that much more of the collected light is used than with a traditional aperture mask. If it works on sky, it could make high-resolution interferometric imaging sensitive enough for fainter targets.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Wedge-orientation manufacturing error (7°) already distorts the 4-pattern layout and is admitted to be detrimental for larger pattern counts; no evidence is given that required placement accuracy can be achieved for 7-8 interferogram designs.","rationale":"After reading the paper, I agree with the reader that manufacturing tolerances are the key uncertainty, but I would sharpen the specific concern. The central claim is that multiple non-redundant interferograms can be formed on separate detector regions. The prototype's interferogram layout deviated from the designed square to a 75° parallelogram, and Section 7.1 attributes this to ~7° errors in the manufacturer's marking of wedge orientation. The authors themselves state that such errors 'would be detrimental for future designs incorporating a larger number of patterns.' This is a direct admission that the demonstrated process does not meet the accuracy needed for the scale-up that makes the concept attractive (Nint = 7 or 8). The OPD concern raised by the reader is necessary but not the immediate failure mode: the achieved alignment errors (50 µm, 0.6°) produce OPD variations of only a few nanometres for the small wedge slopes used, so coherence is likely maintained; rather, the wedge-orientation error visibly alters the pattern and threatens crosstalk if patterns are packed closer. A simulation using measured orientation errors for an 8-pattern design would settle whether the current manufacturing tolerance is adequate. If the simulation shows overlap, the paper's promise of high-throughput multiplexing is conditional on a manufacturing improvement that is not yet demonstrated; if it shows adequate separation, the concern is resolved. Thus the verdict remains CONDITIONAL, as the reader concluded, but for a more precisely identified reason.","tokens_in":13436,"tokens_out":21558,"duration_ms":207021,"concrete_test":"Simulation-based check: using the optical rig described in Section 7.1, measure the actual wedge orientations of the two manufactured VAMPIRES windows. Then draw Monte Carlo orientation errors from the same manufacturing distribution (±7°) and simulate the detector layout of a 3-window, 8-interferogram Jewel design from Table 1 (e.g., Fig. 13 or Fig. 18). If for any pair of interferograms the centroid separation falls below the sum of their radii (or below the designed guard band), the current manufacturing tolerance cannot support the claimed scale-up, and the concept remains conditional on a process improvement not yet demonstrated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires multiplexing several non-redundant patterns onto distinct detector regions. The prototype's four interferograms were designed to sit on a square but instead formed a parallelogram with angle 75°, traced to ~7° errors in the manufacturer's marking of wedge orientation (Section 7.1). The authors state that such errors 'would be detrimental for future designs incorporating a larger number of patterns.' This is the load-bearing issue: the demonstrated manufacturing tolerance already visibly distorts the 4-pattern layout, and the paper offers no evidence that the required orientation accuracy can be met for the 7- or 8-pattern designs, whose interferograms must be packed more closely in the same field of view. If adjacent interferograms overlap or cannot be unambiguously separated, the throughput gain of Nint cannot be realized and the scalability promise fails. The reader's OPD-coherence concern is related but less directly grounded: the observed 50 µm offset and 0.6° rotation produce negligible OPD variations at the small wedge angles used, whereas the wedge-orientation error directly changes the deflection direction and is the demonstrated failure mode.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces Jewel Optics, a pupil-plane optical element that partitions a telescope pupil into several sparse non-redundant sub-aperture patterns, each pattern carrying a distinct phase wedge so that its interferogram is deflected to a separate region of the detector. The authors describe a direct-binary-search algorithm for discovering such tilings, report 14 designs for circular apertures, LBT, and GMT, and present a two-wedge MgF2 prototype for VAMPIRES with four non-redundant four-hole patterns. Laboratory images at 633 nm show four interferograms at approximately the designed locations, with power spectra resembling simulations. The paper also gives geometric throughput estimates (54% for the prototype, 69% for a future seven-pattern optic) and proposes an achromatic doublet wedge design for large apertures such as the LBT.","tokens_in":13581,"tokens_out":6477,"duration_ms":72670,"significance":"If the concept holds, it addresses a real limitation of aperture masking: the drastic throughput loss caused by non-redundant pupil masks. The idea of multiplexing several non-redundant patterns in one optic is attractive and the laboratory demonstration goes beyond a purely theoretical proposal. Strengths of the paper include the separation between the tiling-search merit function and the subsequent throughput/Fourier-coverage claims, the absence of fitted parameters in the comparison between measured interferograms and independent simulations, the availability of code and data, and the explicit exploration of limiting regimes (background, field of view, chromatic dispersion). The main open risk is manufacturing scalability to seven or eight interferograms, where wedge-orientation errors already visible in the four-pattern prototype could prevent reliable separation of the interferograms.","major_comments":[{"comment":"The reported ~7° wedge-orientation marking error is load-bearing for the scalability claim. The manufactured optic produced a parallelogram with angle 75° instead of the designed square, and the text admits that such errors 'would be detrimental for future designs incorporating a larger number of patterns.' Since the central promised gain is the simultaneous use of Nint = 7–8 interferograms on one detector, the paper should quantify the wedge-orientation tolerance required to keep adjacent interferograms separated for, say, the hexagonally arranged seven-pattern designs of Table 1 and Figures 12–13, and either demonstrate that the newly constructed optical rig achieves that accuracy or provide a calibration/alignment strategy that would. Without this, the experimental support for the throughput and Fourier-coverage claims is limited to Nint = 4.","section":"7.1"},{"comment":"The claim of 'excellent agreement between design and lab results' is not quantified. The comparison appears to be visual or based on the locations of the interferograms and power-spectrum peaks; no numerical figure of merit (e.g., normalized mean-square difference between simulated and measured power spectra, measured fringe visibility per baseline, or closure-phase residuals) is provided. Because the method's scientific value rests on the fringes carrying correct complex-visibility information, the laboratory validation should include at least one quantitative comparison of measured power-spectrum peaks or visibility amplitudes/phases with the independent simulation.","section":"7.1, Figure 6"},{"comment":"The central throughput advantage of 3.1× and 6.5× is computed from geometric area and assumed edge losses, not measured for the assembled prototype. This is acceptable for an instrument-concept paper, but the assumptions (±70 µm versus ±20 µm edge defects, air-gapped uncoated MgF2 surfaces) strongly influence the quoted gains, and no sensitivity estimate is given. The paper should either report a measured transmission of the prototype or state clearly that the 3.1×/6.5× factors are modeled predictions and provide a simple uncertainty range reflecting the stated manufacturing tolerances.","section":"6"}],"minor_comments":[{"comment":"The abstract states that 'all individual sets are fully non-redundant,' but Table 1 contains three designs with nonzero redundant baselines (Figures 9, 14, and 15). The abstract should be qualified to refer to the solutions that achieve full non-redundancy, or the redundancy in those designs should be discussed more prominently.","section":"Abstract / Table 1"},{"comment":"The code and data link is given only as '/github-square', which appears to be a placeholder. A working URL should be provided.","section":"Code availability"},{"comment":"There is a typo in the first sentence: 'All lengths are n ow in meters' should read 'All lengths are now in meters.'","section":"A.2"},{"comment":"The reported alignment errors (50 µm linear offset, 0.6° rotation) are not translated into an estimate of differential piston between sub-apertures of the same pattern. A one-sentence calculation of the induced OPD relative to λ would help the reader judge the robustness of the coherence assumption, even though fringes were visible in the current lab test.","section":"7.1"},{"comment":"The figure would be clearer with explicit labels matching the four patterns and a quantitative residual map or error metric between the simulated and measured power spectra; the current side-by-side presentation is suggestive but does not allow the reader to assess the agreement independently.","section":"Figure 6"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nJewel Optics is worth a serious look: the idea is a clean, practical evolution of the Keck segment-tilting experiment, and the authors have done the legwork to make it a drop-in transmissive optic. The new tiling catalog (11 circular-aperture designs plus LBT and GMT) and the direct-binary-search approach are genuine contributions. The lab demonstration with two wedged windows is credible: the interferograms appear at the predicted locations, the power spectra match the simulation, and the reported throughput numbers come from straightforward geometric and loss accounting rather than a fit. The paper also deserves credit for flagging its own limitations—non-redundancy failures in some published tilings, chromatic dispersion, and the field-of-view restriction.\n\nThe soft spot is the manufacturing tolerance story. The prototype shows a 7° error in wedge orientation: the four designed interferograms land as a parallelogram with 75° instead of a square. The authors say this is not a problem for four patterns but would be \"detrimental\" for more. That is exactly where the scalability claim lives. The seven- and eight-interferogram designs need tighter packing on the detector, and the paper offers no evidence that the orientation accuracy can be improved to the required level. I see this as a bigger deal than the OPD-coherence concern the reader raised; the 50 µm offset and 0.6° rotation are negligible at these wedge angles, whereas the orientation error is directly observed and affects the interferogram layout. It is not fatal—they have built an optical rig to re-measure wedge orientation, and a future manufacturing run with better fiducials plausibly fixes it—but it is the load-bearing risk, and the paper should address it with explicit tolerance budgets for the larger designs.\n\nAlso, the code/data link is a placeholder (\"/github-square\"), which blocks independent reproduction of the tiling search. That is a small fix but important for a paper whose catalog is a key result.\n\nOverall, this is a solid instrumentation paper for the aperture-masking community. The proof-of-concept is genuine, the limitations are mostly acknowledged, and the remaining question is manufacturing precision rather than fundamental physics. It deserves a serious referee. I would want the tolerance analysis and the real code link addressed before acceptance, but I would not desk-reject it.\n\nGive it a careful review.","headline":"A credible and honest proof-of-concept for multiplexed non-redundant masking, but the manufacturing tolerance needed for scaling is the unaddressed risk.","tokens_in":14148,"tokens_out":2176,"would_cite":true,"duration_ms":21524,"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":"A single optic made of stacked wedged windows can tile a telescope pupil into several sparse non-redundant patterns, each deflected to its own detector region, so that non-redundant Fizeau interferometry gains roughly the number of…","keywords":["aperture masking","interferometry","non-redundant masking","high angular resolution","pupil remapping","Fizeau beam combination","phase wedge","throughput"],"falsifier":"A decisive test would be to build a three-window Jewel optic with eight interferograms, illuminate it with a monochromatic point source, and compare the measured fringe visibility on each baseline to the design: if wavefront pistons or wedge orientations exceed a fraction of a wavelength, fringe contrast will drop and the recovered visibilities will deviate from the model, contradicting the claim that all patterns stay fully non-redundant and coherent.","tokens_in":13203,"feed_emoji":"🔭","tokens_out":6338,"duration_ms":61429,"temperature":0.7,"pith_summary":"This paper tries to establish that non-redundant aperture masking, which normally throws away most of the light, can be done at high throughput by interlacing several independent non-redundant patterns across the same pupil. Each pattern is given its own small phase wedge, sending its fringes to a separate part of the detector, and the authors show by search that such tilings exist with zero redundant baselines while covering the full available area. A two-wedge prototype for a visible-light instrument transmits 54 percent of the light, about 3.1 times more than one of its constituent masks alone, and its laboratory interferograms match simulation. The same idea, built from two materials with opposite wedge angles, is shown in simulation to be nearly achromatic, which would matter for large telescopes and short wavelengths. If correct, the technique turns aperture masking from a throughput-starved niche into a drop-in, high-sensitivity tool for existing cameras.","feed_headline":"One wedge stack makes four full-pupil interferometry masks","feed_subtitle":"A prototype tiles the pupil with four sparse patterns, cutting the throughput penalty of aperture masking by about threefold.","key_machinery":"The central object is the phase-wedge stack: a set of thin, perforated glass wedges whose through-holes pass light undeflected and whose wedge material adds a tilt that directs each pattern's beam to its own detector region. Non-redundant means that every pair of sub-apertures within a pattern creates a unique baseline frequency, so its fringes are isolated in the Fourier plane. The design machinery is a direct binary search on a hexagonal grid that swaps sub-aperture assignments between patterns and accepts swaps that lower the number of redundant baselines, following the non-redundant point-array idea of compact autocorrelations.","core_discovery":"The central claim is that a single pupil optic can fragment a telescope pupil into several sparse-array non-redundant patterns, each encoded onto a unique phase wedge, with all patterns fully non-redundant while together tiling the available pupil area. The implementation stacks perforated wedged windows, where each window gives a binary choice of hole or glass for every sub-aperture, so N_w windows support up to 2^N_w interferograms; the prototype uses two windows to make four patterns of four sub-apertures. The authors report that the searched tilings cover the pupil with no redundant baselines, and that the manufactured prototype shows excellent agreement between design and laboratory results, with throughput 3.1 times that of a comparable single aperture mask. They further show that an achromatic doublet version reduces chromatic smearing by a factor of about 150, which would remove a barrier to short wavelengths and large apertures.","pith_inferences":["A natural extension would be to augment the direct binary search merit function to include Fourier-coverage depth or sky-background, letting a specific science case trade between more independent samples of a few baselines and broader coverage; the paper flags this trade but does not implement it.","If wedge-orientation metrology is hardened beyond the prototype's roughly seven-degree marking error, the same stack construction should scale to the full 2^N_w patterns, such as three windows for eight interferograms, without new optical principles.","The wavefront-sensing readout, combined with the fact that fringes are measured at the science sensor, suggests a testable route to calibrating pupil-discontinuity modes that conventional wavefront sensors miss."],"forward_implications":["Throughput scales roughly with the number of interferograms: a Jewel optic with N_int patterns passes about N_int times more light than a conventional mask with the same sub-apertures, or 3.1 times for the four-pattern prototype and potentially 6.5 times for a seven-pattern design with improved manufacturing.","Fourier coverage is denser because multiple interferograms measure many spatial frequencies simultaneously, and several of the published tilings achieve zero redundant baselines while sampling the available grid.","The mask is a drop-in optic for standard filter-wheel slots, so existing adaptive-optics cameras can gain aperture-masking capability without changing observatory infrastructure, in contrast to primary-mirror segment tilting.","An achromatic doublet version keeps fringe smearing below roughly lambda/D even at large aperture and mid-infrared wavelengths, opening a path toward extremely large telescopes and shorter wavelength bands.","Because baselines map uniquely to pupil phasing, the same optic can act as a science imager and a wavefront sensor at the science focal plane, avoiding non-common-path errors."],"supporting_citations":[{"why":"Establishes the non-redundant masking technique whose throughput penalty this paper addresses.","marker":"[1]"},{"why":"Describes the predecessor experiment that tilted mirror segments to form multiple interferograms; Jewel Optics generalises it from segments to a drop-in optic.","marker":"[5]"},{"why":"Supplies the non-redundant point-array design criterion and the hexagonal-grid geometry used in the search.","marker":"[18]"},{"why":"Defines the instrument whose pupil, filters, and detector constrain the prototype design.","marker":"[23]"},{"why":"Provides the multi-band upgrade that the paper combines with Jewel Optics to reach 32 simultaneous interferograms.","marker":"[24]"}],"fun_headline_variants":["Jewel Optics: four full-pupil masks from one wedge stack","Four non-redundant masks from one optic, threefold throughput","Jewel Optics: full-pupil masks without apodisation loss","Threefold throughput from a single Jewel Optics wedge","Jewel Optics: four masks, one optic, no throughput penalty"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the stacked wedged windows keep optical path differences between sub-apertures of the same pattern well below one wavelength after machining, alignment, and gluing; the prototype's offsets and wedge-orientation errors were tolerable here, but the paper says such errors would be detrimental if more patterns are added.","fun_headline_variants_meta":{"raw":{"variants":["Jewel Optics: four full-pupil masks from one wedge stack","Four non-redundant masks from one optic, threefold throughput","Jewel Optics: full-pupil masks without apodisation loss","Threefold throughput from a single Jewel Optics wedge","Jewel Optics: four masks, one optic, no throughput penalty"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001185,"raw_usage":{"total_tokens":4914,"prompt_tokens":986,"completion_tokens":3928,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":602,"completion_tokens_details":{"reasoning_tokens":3838}},"tokens_in":602,"tokens_out":3928,"duration_ms":26388,"temperature":1.0,"reasoning_tokens":3838,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:24:51.315386+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be to build a three-window Jewel optic with eight interferograms, illuminate it with a monochromatic point source, and compare the measured fringe visibility on each baseline to the design: if wavefront pistons or wedge orientations exceed a fraction of a wavelength, fringe contrast will drop and the recovered visibilities will deviate from the model, contradicting the claim that all patterns stay fully non-redundant and coherent.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the non-redundant masking technique whose throughput penalty this paper addresses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the predecessor experiment that tilted mirror segments to form multiple interferograms; Jewel Optics generalises it from segments to a drop-in optic."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the non-redundant point-array design criterion and the hexagonal-grid geometry used in the search."},{"cited_title":"Norris, G","cited_arxiv_id":null,"evidence_quote":"Defines the instrument whose pupil, filters, and detector constrain the prototype design."},{"cited_title":"Lucas, B","cited_arxiv_id":null,"evidence_quote":"Provides the multi-band upgrade that the paper combines with Jewel Optics to reach 32 simultaneous interferograms."}],"review_version":1}