{"id":"148fe94b-bdbc-4d52-a183-c750aed5f9a9","arxiv_id":"2607.22388","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The V36 design of the ELT's ANDES YJH spectrograph is simulated to achieve R≈100,000–123,000 with 2–3 pixel FWHM sampling across Y, J, H bands, pending tolerance verification.","lead":"This paper reports the current V36 design and performance simulations of ANDES YJH, a cryogenic near-infrared spectrograph for the Extremely Large Telescope, including a meter-scale echelle grating mosaic. It shows how the design is expected to meet the R=100,000 resolution and 2-pixel sampling requirements, but relies on an estimated 15% degradation margin rather than a full tolerance analysis.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 15% FWHM degradation margin in §4.2 is the load-bearing assumption for R≥100,000; it is not derived from a tolerance analysis, and the only mosaic misalignment simulation uses a different mosaic geometry and wavelength range. Sensitivity to larger or nonuniform degradation should be tested.","rationale":"The reader's weakest assumption correctly identifies the 15% degradation margin as the crux of the central claim. The paper is a credible design-status report with coherent optical modeling and detailed engineering, but the performance guarantee after tolerances is not demonstrated by a tolerance analysis. The §3.2.1 misalignment simulation is a partial and somewhat mismatched check: it models a different mosaic configuration and includes a wavelength outside the operational range, so it cannot underwrite the universal 15% margin. The conclusion explicitly defers the full tolerance analysis to future phases, so the conditional verdict is appropriate. My read does not introduce a new objection; it reinforces the reader's concern with the same evidence. A concrete tolerance-analysis check would settle whether the R≥100,000 claim survives realistic as-built errors, and until then the conditional verdict stands.","tokens_in":13374,"tokens_out":2444,"duration_ms":32711,"concrete_test":"Run a Monte Carlo tolerance analysis on the V36 optical design in Zemax, perturbing lens surface figures, air gaps, tilts, and grating mosaic segment piston/tip/tilt according to the as-built mechanical tolerances; compute the resulting FWHM of the fibre spread function for all orders and field positions in Y, J, and H. Then re-run the PyEchelle end-to-end model using these per-order FWHM maps instead of the uniform 15% degradation margin. Accept the R≥100,000 claim only if the 5th-percentile worst-case resolving power remains above 100,000 in all three bands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central performance claim—R between 100,000 and 123,000 across all YJH bands after manufacturing and alignment tolerances—rests on a single 'estimated 15 % degradation margin' applied uniformly in the end-to-end model (§4.2). This margin is justified only by 'prior experience with other cryogenic instruments,' not by a tolerance propagation. The only quantitative misalignment simulation (§3.2.1) does not support it: it models a 3×1 Zerodur mosaic at 0.5 µm and 1.8 µm, whereas the V36 baseline is a 4×1 Invar mosaic operating over 0.95–1.8 µm. The simulation qualitatively shows LSF distortion at 1.8 µm, but does not convert that distortion into a FWHM or resolving-power change, and the 0.5 µm case is outside the YS band. If the as-built FWHM degradation is larger than 15% or varies by band/field/order—for example, due to grating segment piston, lens despace/tilt, or detector focus errors—the claimed lower bound R≥100,000 may not hold. The paper itself acknowledges that 'preliminary results will be confirmed in the next phases' with a full tolerance analysis, so this is the weakest link in the headline performance claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the current V36 design of the YJH (YS) cryogenic echelle spectrograph for the ANDES instrument on the ELT, covering 0.95–1.8 μm at a target resolving power R ≈ 100,000. It describes the optical layout, the dual seeing-limited/IFU slit concept, the cold slit selector, the cryogenic/mechanical architecture, and the main technological risk: a metre-scale R4 echelle grating mosaic. The performance section presents end-to-end simulations (Zemax, PCGrate, PyEchelle) of throughput, spot diagrams, sampling, and resolving power. The central claim is that, after applying an estimated uniform 15% FWHM degradation to represent manufacturing and alignment tolerances, the as-built spectrometer will deliver FWHM between 2.0 and 3.0 pixels and resolving power between 100,000 and 123,000 across all three Y, J, H bands (§4.2, §5).","tokens_in":13750,"tokens_out":2535,"duration_ms":31581,"significance":"If the performance claim holds, the YS design demonstrates a feasible path to a large cryogenic high-resolution NIR spectrograph for the ELT, including a technology-critical 4×1 Invar echelle mosaic. The paper is valuable as a design-study status report, and the use of end-to-end simulations with PCGrate and PyEchelle is a strength. However, the headline result—R ≥ 100,000 after tolerances—rests on a single hand-assigned 15% FWHM degradation margin that is not derived from a tolerance analysis, and the only supporting misalignment simulation uses a different mosaic configuration and a wavelength outside the YS band. The claim is therefore not yet substantiated at the level required for a definitive performance guarantee, though the design itself appears plausible.","major_comments":[{"comment":"The central claim that the as-built instrument will maintain R between 100,000 and 123,000 is based entirely on 'an estimated 15% degradation margin' applied uniformly in the end-to-end model, justified only by 'prior experience with other cryogenic instruments.' No tolerance analysis, sensitivity study, or error budget is presented. The conclusion itself cautions that 'full tolerance analysis' will be needed in later phases. This makes the R ≥ 100,000 guarantee load-bearing but unsupported. Please provide either a tolerance propagation, a realistic bound on the degradation, or a sensitivity curve showing how R and FWHM vary with the degradation percentage (e.g., 10%, 15%, 20%, 25%) and with non-uniform degradation across field, band, and order.","section":"§4.2 and §5"},{"comment":"The only quantitative misalignment simulation used to inform the degradation margin is for a 3×1 Zerodur grating mosaic, evaluated at 0.5 μm and 1.8 μm. The V36 baseline, however, is a 4×1 Invar IC-DX mosaic (§3.3), and the YS operating band is 0.95–1.8 μm. The 0.5 μm case is explicitly outside the YS band, and the 3×1 versus 4×1 geometry, substrate, and gap spacing (4 mm vs 10 mm) affect both the number of phase steps and the width of the pupil distribution. Moreover, the simulation reports qualitative LSF distortion but does not convert that distortion into a FWHM or resolving-power change. Thus this simulation does not directly support the 15% margin applied to the baseline design. Please either rerun the misalignment study for the 4×1 Invar geometry over the full YS band, or clearly state that the existing simulation is only illustrative and provide an alternative quantitative basis","section":"§3.2.1 and §3.3"},{"comment":"The text states: 'At the shortest, visible wavelength of 0.5 μm of the YS...'. The YS band begins at 0.95 μm; 0.5 μm is not a wavelength of the YS. This is not merely a typo: it reveals that the simulation was designed for a different spectral range (likely the full ANDES visible arm) and was then applied to the NIR instrument. Please correct the wavelength attribution and ensure the simulation is relevant to the YS operating band.","section":"§3.2.1"}],"minor_comments":[{"comment":"The term 'sampling' is used interchangeably for 'FWHM of the collapsed FSF in pixels' and 'average sampling.' Define both formally at first use, and report the range of sampling across orders/bands, not only the average.","section":"§4.2"},{"comment":"The LSF plots would benefit from quantitative axes (intensity vs. pixel or angular scale) and from a statement of the piston offset magnitude (in nm or waves) used in the simulation. Without these, the reader cannot judge whether the applied offsets are representative of the as-built alignment tolerances.","section":"Figure 10"},{"comment":"The paper alternates between '0.4–1.8 μm' and '0.35–1.8 μm' for the full ANDES coverage; make the numbers consistent.","section":"§1 and abstract"},{"comment":"Reference [8] and [21] are duplicates (same Bouchy et al. paper). Also check author list formatting for Ref. [17]; 'for the for the' appears in the title.","section":"References"},{"comment":"Minor grammatical issue: 'which would result in the degrading and blurring the image' should be 'degrading and blurring the image'.","section":"§5"}],"recommendation":"major_revision","confidential_remarks":"The paper is a useful design-progress report for the instrumentation community, but the central performance guarantee (R≥100,000 across all bands after tolerances) needs a proper tolerance analysis or at least a sensitivity study. The mismatch between the misalignment simulation and the baseline mosaic is a serious validity issue, but it is fixable within the paper's scope by additional simulation or by rephrasing the claim as preliminary. I recommend major revision rather than rejection because the design itself is promising and the missing analysis is a well-defined, finite task."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Reader's take is fair. This is a credible preliminary design report for the ANDES YJH spectrograph, not a new scientific discovery. What's new: the V36 optical configuration, the 4×1 Invar echelle mosaic baseline, and the end-to-end simulated FWHM/resolving-power maps. The paper is honest that V35 was undersampled and describes how V36 addresses it. The engineering discussion of the grating mosaic trade-offs is genuinely useful, and the throughput model with a 5% contingency is a reasonable preliminary estimate.\n\nThe main weakness is the same one the stress test highlights: the 15% FWHM degradation margin used to claim R=100,000–123,000 after manufacturing and alignment tolerances is essentially an ansatz, justified by 'prior experience with other cryogenic instruments' rather than a tolerance analysis. The paper states this clearly and promises a full tolerance analysis next phase, which is good. But the only quantitative simulation of grating misalignment—Section 3.2.1—uses a 3×1 Zerodur mosaic and includes a 0.5 µm wavelength, which is outside the YS band; the baseline is a 4×1 Invar mosaic over 0.95–1.8 µm. That simulation does not translate LSF distortion into FWHM or resolving power, so it does not actually support the 15% figure.\n\nThere are also minor inconsistencies: the grating size is 250×1130 mm in Fig. 8 but 250×1330 mm in Table 2; the text calls 0.5 µm 'the shortest, visible wavelength of the YS', but YS starts at 0.95 µm. These are easy fixes.\n\nNone of this is fatal for a status paper. The authors are appropriately cautious and don't oversell. A serious referee would ask for the tolerance analysis or a softened claim, but the paper deserves referee time. I'd bring it to a reading group focused on instrumentation. I'd cite it if I were writing about cryogenic high-res spectrographs.","headline":"A credible, workmanlike design status paper for ANDES V36; the headline R≥100,000 rests on an honest but unverified 15% degradation margin, and the LSF simulation doesn't match the baseline mosaic, but these are typical growing pains for a preliminary design.","tokens_in":14490,"tokens_out":3401,"would_cite":true,"duration_ms":36607,"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":"The V36 design of the ANDES YJH spectrograph attains an average 2-pixel sampling, and with a 15% degradation margin the modelled resolving power stays between 100,000 and 123,000 across the Y, J and H bands.","keywords":["ANDES","YJH spectrograph","ELT instrumentation","echelle grating mosaic","cryogenic spectrograph","fibre-fed spectrograph","spectral resolution","near-infrared spectroscopy"],"falsifier":"A concrete check: once the V36 spectrograph is built, measure the collapsed fibre spread function FWHM at a grid of positions in the Y, J and H detectors and compute the resolving power from the delivered line profile; if any order falls below R=100,000, the paper's central claim is falsified. Before build, a full tolerance analysis that varies the 4x1 Invar mosaic segment piston, tip and tilt at 0.95–1.8 µm and propagates the resulting wavefront error through the end-to-end model would also settle whether the 15% margin is adequate.","tokens_in":13307,"feed_emoji":"🔭","tokens_out":6606,"duration_ms":70369,"temperature":0.7,"pith_summary":"The paper reports the current design and performance analysis of the YJH infrared spectrograph module for the ELT's ANDES instrument, covering 0.95–1.8 µm at a target spectral resolution of about 100,000. The central claim is that the redesigned V36 optical train achieves an average sampling of exactly 2 pixels across the fibre spread function, and that even after applying a 15% degradation margin to represent manufacturing and alignment errors, the resolving power at the detector remains between 100,000 and 123,000 in all three bands. This matters because the YJH module must combine ultra-stable wavelength calibration with high resolution on the largest telescope, to study exoplanet atmospheres, the intergalactic medium, and fundamental constants. The riskiest element is the primary disperser: a cryogenic R4 echelle grating mosaic over a metre long, which no existing instrument has built.","feed_headline":"One-metre echelle mosaic holds 100,000 resolution across three bands","feed_subtitle":"Cryogenic near-infrared spectrograph for the ELT keeps resolving power above 100,000 under a 15% degradation margin.","key_machinery":"The load-bearing component is the cryogenic R4 echelle grating mosaic: a roughly 1.3 metre long, 16 lines-per-mm grating on an Invar substrate, assembled from four segments because no single grating of that size can be manufactured, and mounted with same-material kinematics to keep alignment from room temperature to 100 K. Around it, the optical system uses a 75-fibre reformatted slit with microlenses, an anamorphic slit module, a non-Littrow off-axis collimator, dichroic splitters into Y/J/H channels, and three five-lens cameras each feeding an H4RG-15 detector. The performance claim is carried by the end-to-end model and the PyEchelle simulator that convert the fibre spread function FWHM i","core_discovery":"On the paper's own terms, the discovery is a design validation rather than a new astrophysical result: the V36 YJH spectrograph layout, after moving the anamorphic slit module and rebalancing the three cameras, produces a collapsed fibre spread function with a FWHM averaging 2.0 pixels across the Y, J and H orders. When the end-to-end model applies an estimated 15% FWHM degradation as a stand-in for as-built tolerances, the FWHM falls between 2.0 and 3.0 pixels, the inter-order gaps exceed 6 pixels, and the resolving power stays between 100,000 and 123,000 at every wavelength. The paper stresses that these are preliminary numbers to be confirmed by a matured optical model, a full tolerance a","pith_inferences":["The paper leaves implicit that the 15% degradation margin is a placeholder rather than a measured budget; a full tolerance analysis could show the real FWHM growth varies across the field, which would make the R≥100,000 claim wavelength-dependent.","The misalignment simulation in the paper uses a 3x1 Zerodur mosaic at 0.5 µm, while the baseline is a 4x1 Invar mosaic operating at 0.95–1.8 µm; repeating the piston-shift study at the baseline configuration would test whether the H-band LSF distortion seen in the simulation actually matters for the science cases.","The tight link between the R4 mosaic's supply and the resolution requirement suggests that if the supplier cannot deliver the 16 lines-per-mm R4 grating with the assumed efficiency, the whole 100,000-resolving-power case would need to be reworked.","The interchangeable seeing-limited and diffraction-limited modes place different illumination patterns on the same slit and detector; the paper's FWHM metric is based on a particular fibre format, so the IFU mode may need its own verification of sampling and LSF."],"forward_implications":["If the V36 performance holds in the as-built instrument, a single fibre-fed cryogenic spectrograph can cover the full 0.95–1.8 µm range at R≥100,000 without being oversampled, meeting the baseline ANDES resolution requirement.","The greater-than-10^14 suppression of thermal background from the cold slit selector at 1.8 µm means the H band, the hardest for warm background, is usable for high-resolution science.","Because the 4x1 Invar mosaic keeps the line spread function between 2.0 and 3.0 pixels, the detector sampling stays in the regime needed for precise radial-velocity work, supporting the 1 m/s stability goal over 24 hours.","The grating mosaic's gaps and segment misalignments produce LSF wing distortion at infrared wavelengths but, at the simulated levels, do not push resolving power below 100,000; this defines how much grating segment alignment error the science can tolerate."],"fun_headline_variants":["Metre-long echelle mosaic keeps ELT's YJH resolving power above 100k","Cryogenic YJH spectrograph passes design test at 100,000 resolution","Design of largest cryogenic spectrograph holds 100k resolution","With 15% tolerance, ELT's YJH spectrograph keeps 100k resolution"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The central claim rests on the assumption, stated in Section 4.2, that a single uniform 15% FWHM degradation margin chosen from prior experience with cryogenic instruments represents the combined manufacturing and alignment errors of this as-built spectrograph; if the real degradation is larger or nonuniform, the R≥100,000 result may not hold.","fun_headline_variants_meta":{"raw":{"variants":["Metre-long echelle mosaic keeps ELT's YJH resolving power above 100k","Cryogenic YJH spectrograph passes design test at 100,000 resolution","Design of largest cryogenic spectrograph holds 100k resolution","With 15% tolerance, ELT's YJH spectrograph keeps 100k resolution"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000351,"raw_usage":{"total_tokens":1788,"prompt_tokens":815,"completion_tokens":973,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":559,"completion_tokens_details":{"reasoning_tokens":892}},"tokens_in":559,"tokens_out":973,"duration_ms":11457,"temperature":1.0,"reasoning_tokens":892,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T04:53:07.634270+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: once the V36 spectrograph is built, measure the collapsed fibre spread function FWHM at a grid of positions in the Y, J and H detectors and compute the resolving power from the delivered line profile; if any order falls below R=100,000, the paper's central claim is falsified. Before build, a full tolerance analysis that varies the 4x1 Invar mosaic segment piston, tip and tilt at 0.95–1.8 µm and propagates the resulting wavefront error through the end-to-end model would also settle whether the 15% margin is adequate.","supporting_citations":[],"review_version":1}