{"id":"90b4b3ab-af11-4088-994f-0a17ef76a28c","arxiv_id":"2511.10773","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Fabrication refinements—lateral-etch compensation, stepped Parylene-C AR coatings, and sub-micrometer epoxy bonds—bring PRIMA FIRESS Band 1 and Band 4 lenslet arrays to specification.","lead":"This paper reports on making and testing the tiny silicon lens arrays that focus light onto the superconducting detectors of the planned PRIMA space telescope, including a corrected etching process, stepped anti-reflection coatings, and epoxy bonding. Generalist readers may care because it shows a key optical component for a future far-infrared observatory moving from design to flight-like hardware.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"5 K-to-120 mK extrapolation of Parylene-C and epoxy optical properties is the load-bearing unvalidated assumption for AR-coating and bond-loss requirements.","rationale":"The reader's weakest_assumption identifies the 5 K-to-120 mK optical-property extrapolation as a load-bearing concern, and this is indeed the single most load-bearing risk to the paper's central claim. The claim that the fabrication, AR coating, and bonding processes 'meet PRIMA requirements' is directly tied to the thickness thresholds, which are computed from material properties measured only at 5 K. If those properties change at 120 mK, the thresholds and the measured bond thicknesses may no longer satisfy the loss budget. This is a silent, unvalidated physical assumption, unlike the interpolation to Bands 2 and 3, which is scoped out of the central claim (the paper explicitly focuses on the extreme bands). The paper's own Section VI acknowledges that optical performance of bonded arrays is pending, so the authors are careful not to overclaim end-to-end performance; but the temperature extrapolation is presented as a settled assertion without supporting data. The concrete test — a cryogenic FTS measurement at sub-Kelvin temperature — would directly retire this risk. If it passes, the conditional verdict can move to accept; if it fails, the thickness budgets need re-examination. The reader's CONDITIONAL rating already accounts for this uncertainty, so I recommend no change to the verdict.","tokens_in":8337,"tokens_out":7729,"duration_ms":69927,"concrete_test":"Cool the same 30 µm Parylene-C and 77 µm Epo-Tek 301 free-standing samples to 120 mK (or the lowest temperature achievable in a ³He cryostat, e.g., ~300 mK) and measure FTS transmission over 20–300 µm using the same spectrometer as in Section IV. Fit the complex dielectric function and compare n and loss tangent to the 5 K values. If the loss tangent increases by more than the amount that would push the measured bond thicknesses above the 5% loss budget, or if n shifts by more than ~1%, the AR-coating and bond-thickness requirements must be revised. A null result would confirm the extrapolation and strengthen the central claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the AR-coated and bonded lenslet arrays meet PRIMA FIRESS requirements rests on the thickness budgets derived from FTS measurements at 5 K. In Section IV, the authors determine n(λ)=1.66–1.68 for Parylene-C at 5 K and state without evidence that this represents the material's properties at the 120 mK operating temperature. In Section V, the epoxy bond thickness thresholds (<1 µm for Band 1, <6 µm for Band 4) are calculated from the 5 K FTS of Epo-Tek 301, and the quoted bond thicknesses are then declared 'meeting the PRIMA requirements.' This extrapolation is not validated: amorphous polymers like Parylene-C and epoxy can exhibit temperature-dependent far-infrared properties due to two-level systems or phonon modes that persist down to sub-Kelvin temperatures. If the loss tangent or refractive index changes between 5 K and 120 mK, the quarter-wave AR thicknesses (5 µm and 28 µm) could become sub-optimal, and the measured 1–4 µm Band 4 bond layers might exceed the reflection/absorption loss budget. The paper gives no error bars on the FTS fits and defers the full dielectric model to an in-prep publication, so the margin is unknown. This is not an internal inconsistency but a correctness risk that directly threatens the 'meeting requirements' conclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports process refinements for the fabrication, anti-reflection (AR) coating, and bonding of monolithic silicon lenslet arrays for the PRIMA FIRESS instrument. The authors focus on the two extreme spectral bands (Band 1: 24–43 µm and Band 4: 130–235 µm). They describe a grayscale lithography and DRIE process with lateral etch compensation that enables deeper and wider lenses, including hexagonal-corner designs. For Band 1 they report profilometry residuals <0.9 µm; for Band 4 they show SEM/3D-profilometry images of the improved lens shape. They present FTS measurements at 5 K of Parylene-C and Epo-Tek 301, from which they derive AR-coating thicknesses (5 µm and 28 µm) and epoxy bond-thickness thresholds (<1 µm and <6 µm). Destructive bond metrology yields 1–4 µm for Band 4 and <1 µm for Band 1 dies, claimed to meet PRIMA requirements. Stepped-thickness AR coatings are demonstrated on a two-level test wafer, and a 28-µm Parylene-C coating survives 10 thermal cycles between 77 K and 323 K. The paper concludes that the optimized arrays for the two extreme bands are ready for bonding and cryogenic optical testing, with intermediate bands expected to follow by interpolation.","tokens_in":8594,"tokens_out":5690,"duration_ms":57553,"significance":"If the claims hold, this is a valuable engineering contribution: it shows a practical route to kilopixel, monolithic silicon lenslet arrays for background-limited far-infrared KID spectrometers, with quantified fabrication tolerances, material characterization, and bonding metrology. The paper provides concrete data on profilometry residuals, destructive bond-thickness measurements, thermal cycling, and FTS-derived material transmission. The lateral-etch compensation and hexagonal-corner lens design are plausible and potentially transferable to other FIR/THz instruments. However, two load-bearing elements are not directly demonstrated: the 14% optical-power gain from the hexagonal-corner lenses is asserted without simulation or measurement, and the extrapolation of 5-K FTS material data to the 120-mK operating temperature is unvalidated. These do not invalidate the fabrication results but weaken the 'meets PRIMA requirements' conclusion. The paper is well within the scope of IEEE TAS and will be of interest to the detector and instrumentation community.","major_comments":[{"comment":"The claim that the upgraded hexagonal-corner lenses 'direct ≈14% more optical power to the detectors' is presented as a quantitative result, but no derivation, simulation setup, ray-trace output, or measured throughput comparison is given. This number is load-bearing for the stated efficiency increase and for the value of the redesign. Please provide the supporting optical model (including the assumed focal-plane illumination, lens geometry, and absorber size) or a direct measurement, or adjust the claim to reflect an estimate based on geometric area only.","section":"Section III (last paragraph)"},{"comment":"The FTS measured at a 5 K bath temperature are stated (Section IV) to represent the material's optical properties at the PRIMA operating temperature of 120 mK. This extrapolation is not justified. The AR-coating thicknesses (5 µm and 28 µm) and the bond-loss thresholds (<1 µm and <6 µm) are directly derived from these data. If the complex dielectric function of Parylene-C or epoxy changes between 5 K and 120 mK (e.g., from two-level systems or low-frequency phonon absorption), the coating and bond thickness requirements could shift, and the conclusion that the measured bond thicknesses 'meet the PRIMA requirements' would be insecure. Please provide a concrete argument for temperature insensitivity, low-temperature validation data, or quantitative uncertainty bounds on the FTS fits and their propagation to the thickness budgets.","section":"Sections IV and V (FTS extrapolation)"},{"comment":"Quantitative profile residuals are reported only for Band 1 (RMS <0.9 µm in Figure 2). For Band 4, the improved lens shape is shown in SEM and 3D profilometry, but no numerical comparison of the etched profile to the design (e.g., RMS residual, maximum deviation) is given. Since the paper claims the fabrication process meets FIRESS requirements for the extreme bands, and Band 4 required the lateral-etch compensation, the absence of a quantitative Band 4 profile error leaves the requirement verification incomplete. Please include a measured residual map or table for the Band 4 lenses.","section":"Section III, Band 4 profile accuracy"},{"comment":"The derivation of the '5% loss' bond-thickness thresholds is not shown. The text states 'we calculated' and gives the resulting values (<1 µm for Band 1, <6 µm for Band 4) without presenting the model, the fringe- or absorption- loss formulation, or the material optical constants used. Because these thresholds are the basis for declaring the measured bond thicknesses compliant, please provide the calculation in sufficient detail to allow reproduction, or cite a publication containing it.","section":"Section V, bond-loss calculation"}],"minor_comments":[{"comment":"The statement that intermediate-band designs 'are expected to be relatively straightforward interpolations' is an assertion. Since the paper's title and abstract focus on PRIMA FIRESS generally, please either show a basis for this expectation (e.g., similarity of the two extreme designs) or explicitly cap the claim to Bands 1 and 4.","section":"Section II"},{"comment":"The complex dielectric function fit is deferred to 'Wollack et al., in prep.' which is not accessible for review. At minimum, give the retrieved values of n and k (or loss tangent) and the uncertainty, so that the thickness calculations are reproducible.","section":"Section IV"},{"comment":"No error bars or measurement uncertainties are shown on the FTS transmission spectra. Adding typical uncertainty envelopes (including thickness uncertainty of the free-standing samples) would strengthen the derived thickness budgets.","section":"Figure 4"},{"comment":"The abstract says 'demonstrate stepped-thickness AR-coatings to achieve high efficiency across broad wavelength ranges,' but the demonstration is a two-thickness test wafer, not a full four-step coating on a kilo-pixel array, and no optical efficiency measurement is reported. Please qualify the claim to 'process demonstration' and note that the final four-step arrays are under fabrication.","section":"Abstract and Section IV"},{"comment":"The destructive bond-thickness measurements are reported as ranges ('1–4 µm' for Band 4, '<1 µm' for Band 1) without stating the number of dies/locations measured or the measurement uncertainty. Adding this information would allow readers to judge the uniformity and reliability of the metrology.","section":"Section V"}],"recommendation":"major_revision","confidential_remarks":"The paper is well matched to IEEE TAS and the core fabrication/bonding metrology appears solid. The main concerns are the unquantified 14% optical-power claim and the 5 K-to-120 mK material-property extrapolation, both of which directly support the 'meeting PRIMA requirements' conclusion. These are fixable with additional simulations, measurements, or clearly framed limitations. The reliance on an in-prep reference for the dielectric model is also a weakness that should be addressed. No concerns about novelty or authorship practices."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the fabrication and bonding work is real, new, and backed by direct measurements. The optical claims (the 14% gain, the 'high efficiency' AR coating) are asserted rather than demonstrated, and the 5 K-to-120 mK extrapolation for Parylene-C and epoxy is a legitimate open question. None of that sinks the paper, but it does mean the 'meeting PRIMA requirements' conclusion is partly ahead of the evidence.\n\nWhat's genuinely new: the lateral-etch compensation model (65–75% isotropic etch fraction) that lets them etch deeper into the hexagonal corners, the hexagonal-corner Band 4 lens geometry, the demonstrated stepped-thickness Parylene-C coating, and the measured bond thicknesses for thick dies. The profilometry residuals (<0.9 µm) and the destructive bond measurements (<1 µm Band 1, 1–4 µm Band 4) are exactly the kind of direct evidence you want. Ten thermal cycles surviving is also a real data point. This is a clear advance over the group's prior work [4].\n\nWhere it's soft: the '≈14% more optical power' is just a number dropped in the text, with no ray trace, no throughput calculation, no measured optical efficiency behind it. For a paper that is otherwise careful about metrology, that one is out of place. The stepped AR coating is demonstrated as a two-thickness test wafer, not the four-step flight configuration; that's a fine process demo but not the final product. And the 5 K-to-120 mK extrapolation for the dielectric functions is asserted rather than justified. Amorphous polymers can have temperature-dependent loss from TLS or low-lying modes, so the bond-loss thresholds could shift. I don't think it's fatal—these materials tend to stabilize well above 120 mK—but the authors should at least state the assumption explicitly with some rationale, ideally with error bars on the FTS fits. The intermediate-band interpolation is a reasonable plan, not a result, and the paper does not oversell it.\n\nBottom line: the central engineering claims are solid and honestly scoped—Section VI says clearly that bonded-array optical testing is still pending. This deserves a serious referee. The referee should push for a derivation of the 14% number, error bars on the FTS fits, and a sentence on the temperature extrapolation. With those added, I'd be comfortable.","headline":"A solid, metrology-backed fabrication and bonding advance for PRIMA lenslets; the 14% optical gain claim and the 5 K-to-120 mK material extrapolation are the soft spots that need shoring up.","tokens_in":9192,"tokens_out":3076,"would_cite":true,"duration_ms":28911,"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 paper claims that optimized kilopixel silicon lenslet arrays for PRIMA's FIRESS spectrometer now meet fabrication, anti-reflection coating, and bonding requirements at the two extreme spectral bands, ready for cryogenic optical testing.","keywords":["PRIMA","FIRESS","kinetic inductance detectors","lenslet arrays","grayscale lithography","deep reactive ion etching","Parylene-C anti-reflection coating","epoxy bonding"],"falsifier":"Cool free-standing Parylene-C and Epo-Tek 301 films to 120 mK and remeasure their far-infrared transmission with an FTS; if the extracted refractive index or absorption differs from the 5 K values by more than the design tolerance, the reported AR-coating thicknesses and 1 µm/6 µm bond thresholds would no longer meet the 5 percent loss requirement.","tokens_in":8149,"feed_emoji":"🔭","tokens_out":6294,"duration_ms":55800,"temperature":0.7,"pith_summary":"Optimized silicon lenslet arrays for PRIMA's FIRESS spectrometer now meet the instrument's fabrication, anti-reflection coating, and bonding requirements at the two extreme wavelength bands. The key improvement is a fabrication model that accounts for 65–75 percent isotropic etching, allowing compensated resist profiles that yield deeper, wider lenses and hexagonal-corner shapes that collect about 14 percent more optical power. The paper also shows stepped-thickness Parylene-C coatings matched to the spectrometer dispersion and epoxy bond layers under 1 µm for the short-wavelength band and 1–4 µm for the long-wavelength band, keeping coupling losses below 5 percent. If these results hold, the kilopixel lenslet arrays are ready to bond to detector arrays and proceed to cryogenic optical testing, with the two intermediate FIRESS bands expected to follow by interpolation.","feed_headline":"Lenslet arrays meet PRIMA specs at both extreme FIRESS bands","feed_subtitle":"Hexagonal lenses and sub-micron epoxy bonds keep PRIMA's losses under 5 percent, clearing the way for cryogenic tests.","key_machinery":"The load-bearing mechanism is a modified grayscale-lithography and deep reactive ion etching process in which the etch is modeled as 65–75 percent isotropic, so an initial photoresist profile is deliberately over-deepened at larger radii to compensate for lateral etching. This correction is what allows deep lenses with hexagonal corners to be etched into thin dies. Around it sit two auxiliary mechanisms: quarter-wave Parylene-C anti-reflection coatings, stepped in thickness along the array to follow the spectrometer's dispersion, and an epoxy bonding process using a flip-chip bonder with modeled bond force to hold the adhesive layer thin enough to stay inside a 5 percent coupling-loss budget","core_discovery":"The central claim is that the combination of lateral-etch-compensated grayscale lithography, quarter-wave Parylene-C anti-reflection coatings, and thin epoxy bonding now satisfies PRIMA's requirements for both FIRESS Band 1 (24–43 µm) and Band 4 (130–235 µm). Profilometry shows lens-profile residuals below 0.9 µm for Band 1 dies, and the redesigned Band 4 lenses with hexagonal corners direct approximately 14 percent more optical power to the detectors than the previous circular lenses. FTS measurements at 5 K give a Parylene-C refractive index of 1.66–1.68, setting AR-coating thicknesses of about 5 µm and 28 µm for Bands 1 and 4, and the same measurements establish epoxy thickness thresholds","pith_inferences":["Inference: A direct measurement of Parylene-C and Epo-Tek 301 optical constants at 120 mK, not just 5 K, would test the main extrapolation; if the refractive index shifts by even a few percent, the optimum AR thicknesses and loss thresholds would move.","Inference: The 'straightforward interpolation' to FIRESS Bands 2 and 3 is asserted but not shown; a validation run with band-center test lenses would confirm whether the two extreme designs truly bound the intermediate ones.","Inference: The 14 percent optical-power gain is based on geometry; a before-and-after measurement of optical efficiency of circular versus hexagonal-corner lenslets on a bonded KID array would put the number on firmer footing.","Inference: If the 5 percent bond-loss budget is later tightened after system-level error budgeting, the demonstrated <1 µm Band 1 bond thickness leaves little margin, so the process may need re-optimization for any more stringent requirement."],"forward_implications":["Band 1 and Band 4 lenslet arrays can now be bonded to KID arrays and tested at cryogenic temperatures with the expectation that the as-fabricated optics meet PRIMA's requirements.","The same fabrication process can be applied to the remaining FIRESS bands and to PRIMA Imager lenslets, since the two extreme bands bound the parameter space.","The 14 percent gain in collected optical power from hexagonal-corner lenses, combined with lower stray light, directly improves spectrometer sensitivity if reproduced in bonded arrays.","Stepped-thickness AR coatings should let each FIRESS band maintain near-peak transmission across its full 1.8:1 wavelength range.","The measured bond thicknesses (<1 µm for Band 1, 1–4 µm for Band 4) make the 5 percent loss requirement achievable for both extreme bands."],"fun_headline_variants":["Hexagonal lenses lift PRIMA FIRESS coupling 14%","PRIMA lenslets pass specs for both FIRESS bands","Thin epoxy, thick AR coat ready PRIMA lenslets","Grayscale-etched lenslets clear PRIMA FIRESS specs","Optimized lenslets meet PRIMA's dual-band FIRESS needs"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole thickness budget rests on assuming that the optical properties of Parylene-C and Epo-Tek 301 measured at 5 K are identical at the 120 mK operating temperature, and that the two extreme-band designs interpolate cleanly to the two intermediate FIRESS bands; neither assumption is directly validated in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Hexagonal lenses lift PRIMA FIRESS coupling 14%","PRIMA lenslets pass specs for both FIRESS bands","Thin epoxy, thick AR coat ready PRIMA lenslets","Grayscale-etched lenslets clear PRIMA FIRESS specs","Optimized lenslets meet PRIMA's dual-band FIRESS needs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000634,"raw_usage":{"total_tokens":2807,"prompt_tokens":834,"completion_tokens":1973,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":578,"completion_tokens_details":{"reasoning_tokens":1896}},"tokens_in":578,"tokens_out":1973,"duration_ms":14137,"temperature":1.0,"reasoning_tokens":1896,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T22:20:37.050875+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cool free-standing Parylene-C and Epo-Tek 301 films to 120 mK and remeasure their far-infrared transmission with an FTS; if the extracted refractive index or absorption differs from the 5 K values by more than the design tolerance, the reported AR-coating thicknesses and 1 µm/6 µm bond thresholds would no longer meet the 5 percent loss requirement.","supporting_citations":[],"review_version":1}