{"id":"7cd50013-9c22-4681-a0df-9f0f20265979","arxiv_id":"2412.10189","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Commissioning of the MIRAC-5 mid-infrared camera on the MMT yields roughly 10% total throughput, near-diffraction-limited N-band images, and calibrated sensitivity estimates for exoplanet science.","lead":"MIRAC-5, a new mid-infrared camera on the MMT telescope, completed commissioning observations that measure its total throughput at about 10% and show near-diffraction-limited image quality in the N band. The paper reports calibrated sensitivity limits and an exposure time calculator for future exoplanet and disk observations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Future-dichroic limiting magnitudes rest on an ETC calibrated against the same May 2024 Alpha Boo data, so 18.0/15.6/12.6 are extrapolations; the paper's own Nov 2024 data is the missing external test.","rationale":"The reader's weakest_assumption is close to mine, but I put less weight on the 1/sqrt(t) long-integration scaling and more on the circular calibration chain: the ETC's empirical corrections are fitted to the same Alpha Boo data that produces the throughput and background tables, so the Table 8 limits are not externally validated. I also note that the Table 8 limits assume an ideal Airy PSF, whereas the current measured encircled energies within the FWHM are only 8.3%, 6.8%, and 19.9% in L', M', and N', respectively; the quoted magnitudes therefore also depend on unverified AO performance. Both concerns support the reader's CONDITIONAL verdict rather than pushing it to REJECT, because the directly measured current performance (throughput near 10%, background levels, 1/f behavior, N-band PSF) is supported by a transparent pipeline and the authors explicitly flag the future-sensitivity assumptions. The concrete test uses the paper's own independent Nov 2024 observations, currently relegated to a footnote, to break the circularity. If that test passes, the concern is resolved; if it fails, the central competitive-sensitivity claim would need to be reworded as an expectation. No change to the reader's verdict is needed.","tokens_in":21968,"tokens_out":8684,"duration_ms":90504,"concrete_test":"External validation with the Nov 2024 data: run the already-calibrated ETC for the Beta Gem / Beta Peg / Beta And L' and N' observations using the measured frame rates, sky backgrounds, and PSF/encircled-energy inputs with the dichroic removed, and compare predicted source count rates and background levels to measured values without refitting any per-bandpass correction. If the model reproduces both bands to better than ~20% and the implied 8-hour limiting magnitudes shift by less than ~0.3 mag, the calibration transfers; if not, the Table 8 future-dichroic magnitudes should be labeled provisional rather than headline results.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline limiting magnitudes (18.0/15.6/12.6 in L', M', N') are generated by the exposure-time calculator in Section 4.2, which applies per-bandpass empirical corrections to both throughput and telescope/instrument background. These corrections are fitted to the same May 2024 Alpha Boo dataset used for the throughputs in Table 7 and backgrounds in Table 6, so the quoted agreement 'within 10%' is consistency with calibration data rather than independent validation. Table 7 reports no uncertainties, and the future-dichroic entries in Table 8 are obtained by scaling current measured throughputs and backgrounds using filter-table dichroic transmissions (Table 3) plus a modeled dichroic emission. The transferability of per-bandpass empirical corrections to a different dichroic and to 8-hour integrations is therefore untested. The independent Nov 2024 no-dichroic L' and N' measurements are mentioned only in a footnote and used neither to validate the ETC nor to update the Table 8 predictions. The paper's own Section 5.1 caveat that 'several hour integrations may not continuously scale as 1/sqrt(t) owing to noise sources with non-Poisson distributions' reinforces that the headline sensitivity is an extrapolation. The directly measured current performance is credible, but the specific competitive-sensitivity claim is not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports commissioning results for MIRAC-5, a ground-based mid-infrared imager on the 6.5-m MMT, built around a GeoSnap HgCdTe detector and supported by the MAPS adaptive optics system. The authors characterize the detector dark current, read noise, 1/f noise, pixel-to-pixel QE variations, effective sky/telescope/instrument backgrounds, on-sky throughputs, and delivered image quality using Alpha Boo data from May 2024 and additional data from other runs. They present an exposure-time calculator (ETC) calibrated to these on-sky measurements and use it to predict background-limited sensitivities. The central quantitative claims are a total current throughput of roughly 10%, a future throughput of about 20% after a dichroic upgrade, and SNR=5, 8-hour limiting magnitudes of 18.0 (L'), 15.6 (M'), and 12.6 (N') for the future configuration.","tokens_in":22170,"tokens_out":6119,"duration_ms":60443,"significance":"If the measured throughput, background, and image-quality characterizations are reliable, this paper provides valuable, quantitative information for the ground-based mid-infrared community, particularly regarding the GeoSnap detector's on-sky behavior and the practical mitigation of 1/f noise through chop/nod scheduling. The public exposure-time calculator is a useful tool for planning observations, and the paper makes falsifiable predictions for the planned dichroic upgrade. The strongest measured results—the backgrounds, 1/f scaling over 83 minutes, and N-band image quality—are supported by a clear reduction pipeline and on-sky data. However, the headline sensitivity numbers are ETC model outputs, not direct measurements, and their dependence on empirical corrections fitted to the same calibration dataset and on untested 8-hour noise scaling means the competitive-sensitivity claim is not yet fully established.","major_comments":[{"comment":"The future-dichroic limiting magnitudes in Table 8 are produced by the exposure-time calculator using per-bandpass empirical throughput and background corrections fitted to the May 2024 Alpha Boo data (Section 4.2). The quoted 'within 10%' agreement therefore tests internal consistency with the calibration dataset, not independent predictive accuracy. The Nov 2024 no-dichroic L' and N' data, which could serve as an external check, appear only in a footnote to Table 7 and are not used to validate the ETC or update the Table 8 predictions; the throughputs reported there (19.5% L', 12.3% N') lie below the future-dichroic predictions (22% and 17%). Please report uncertainties on the limiting magnitudes and either validate the ETC against the Nov 2024 data or present the Table 8 values explicitly as model projections that remain to be verified on sky.","section":"§4.2 and Table 8"},{"comment":"The 8-hour limiting magnitudes assume that noise scales as 1/sqrt(N) out to 8 hours, but the longest on-sky noise-scaling measurement shown is 83 minutes (Figure 5). The authors themselves caution in Section 5.1 that 'several hour integrations may not continuously scale as 1/sqrt(t) owing to noise sources with non-Poisson distributions,' and Figure 5 demonstrates an early departure from 1/sqrt(N) when data are temporally co-added before subtraction. The abstract's headline limiting magnitudes are therefore extrapolations. Please add a quantitative discussion of how a non-Poisson noise floor or systematic background drift would affect the Table 8 limiting magnitudes, or explicitly label these numbers as optimistic projections.","section":"§5.1 and Figure 5"},{"comment":"The throughput values in Table 7 are quoted without uncertainties, despite being derived from aperture photometry of Alpha Boo with known magnitude uncertainties (0.01-0.03 mag), a PSF with extended wings, and nod-pair subtraction with frame rejection. The abstract's central claim of 'approximately 10%' throughput and the future 20% estimate scale directly from these numbers. Please provide at least the statistical photometric uncertainties and propagate the standard-star magnitude uncertainties so that the measured throughputs can be assessed quantitatively.","section":"Table 7"}],"minor_comments":[{"comment":"The abstract states that MIRAC-5 was used on 'six engineering observing runs,' while Section 1 says 'five observing runs'; please harmonize the count.","section":"Abstract and §1"},{"comment":"When reporting that effective background levels changed 'at only about a 2.5% level,' please clarify whether this is an RMS scatter or a peak-to-peak variation across the night.","section":"§3.2"},{"comment":"The text says the limiting magnitude calculation assumes 'an 8 hour observing window at 100% efficiency,' but Table 8 lists observing efficiencies between 91.5% and 97%. Please clarify that the 8 hours refers to science exposure time, with overheads accounted separately.","section":"§5.1 and Table 8"},{"comment":"Typo: 'close-loop heater control' should be 'closed-loop heater control.'","section":"§2.1"},{"comment":"The future dichroic transmission for H-band is 0.004, which is a dramatic change from the current value; a brief note explaining that the new dichroic is not intended for H-band operation would remove potential confusion.","section":"Table 3"}],"recommendation":"major_revision","confidential_remarks":"The measured performance characterization (detector noise, backgrounds, throughputs, image quality) is credible and well documented, and the paper is a useful contribution to the ground-based mid-IR literature. My recommendation of major revision is driven by the need to separate measured results from model-based projections. The authors have independent Nov 2024 data that could partially validate the ETC, and they should use it or clearly reclassify the Table 8 sensitivities as forward-looking estimates with no quoted uncertainty. The internal inconsistency in the number of observing runs is easy to fix but should be caught in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The useful part of this paper is the measured core: first on-sky commissioning numbers for MIRAC-5's GeoSnap detector, including total throughput near 10% in the current configuration, per-band effective backgrounds, 1/f noise behavior, pixel QE variations, and N-band image quality. The reduction pipeline is transparent, the data are presented in enough detail to follow, and the 1/f noise analysis with pair-subtracted frames plus the chopping-efficiency discussion is genuinely practical. That part deserves to be in the literature.\n\nThe soft spot is exactly what the stress-test note flags. The Table 8 limiting magnitudes for the upgraded dichroic (18.0, 15.6, 12.6 in L', M', N') come from an exposure-time calculator that is calibrated with per-bandpass empirical corrections fitted to the same May 2024 Alpha Boo data used for the measured throughputs and backgrounds. So the \"within 10%\" agreement with the data is a consistency check, not an independent validation. The paper itself admits in Section 5.1 that several-hour integrations may not follow 1/sqrt(t) scaling because of non-Poisson noise sources, and the November 2024 no-dichroic data are relegated to a footnote instead of being used to test the model. The future-dichroic numbers are therefore predictions, not measurements. That does not make them worthless, but they should be labeled as extrapolations, and Table 7 needs error bars on the throughputs.\n\nMinor issues: the notebook is mentioned but no repository link is given, and the JWST-comparable contrast claim in the abstract is forward-looking, not demonstrated. None of this undermines the measured commissioning results.\n\nThis is a paper for instrument builders and for observers planning ground-based mid-IR programs. It is not a breakthrough, but it is a careful, honest characterization of a new detector and instrument. I would send it to peer review, with referees asked to push the authors to separate measured from modeled numbers and to release the data and code. A solid commissioning paper with one clearly labeled extrapolation is worth referee time.","headline":"Solid measured on-sky performance, but the headline future-dichroic sensitivities are model extrapolations, not yet validated independently.","tokens_in":22839,"tokens_out":2203,"would_cite":true,"duration_ms":22678,"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":"MIRAC-5, a new mid-infrared camera on the MMT, reaches about 10 percent end-to-end throughput and, after a dichroic swap, should reach background-limited magnitudes of 18.0, 15.6, and 12.6 in L', M', and N' for 8-hour SNR=5 observations.","keywords":["mid-infrared astronomy","instrumentation commissioning","HgCdTe detector","GeoSnap","1/f noise","adaptive optics","throughput measurement","exposure time calculator"],"falsifier":"Take an 8-hour N' integration with the new dichroic installed, reduce it without temporal co-adding, and compare the median per-pixel MAD against the 1/sqrt(N) trend from Figure 5; if the noise floor departs from that trend, as the paper itself warns is possible, the quoted limiting magnitudes would need to be revised. A shorter check would be to compare the exposure time calculator's predicted background and throughput for L', M', and N' against a single night of photometry of a standard star taken after the dichroic swap.","tokens_in":21687,"feed_emoji":"🔭","tokens_out":6068,"duration_ms":53712,"temperature":0.7,"pith_summary":"The paper reports results from six engineering runs of MIRAC-5, a mid-infrared camera on the 6.5-m MMT that uses a new GeoSnap HgCdTe detector with adaptive-optics support from the MAPS system. It claims the current end-to-end throughput, including telescope, atmosphere, instrument, and detector, is approximately 10 percent, and that replacing a temporary low-transmission dichroic will raise that to roughly 20 percent. If those numbers hold, the instrument should reach background-limited magnitudes of 18.0, 15.6, and 12.6 in L', M', and N' bands for an 8-hour SNR=5 observation, placing it alongside the best past ground-based mid-IR imagers and, after a coronagraph upgrade, in the contrast regime now occupied by JWST. The paper also shows that the detector's 1/f noise can be suppressed below ten percent of the Poisson background noise by chopping and nodding, and it provides a calibrated exposure-time calculator for observers.","feed_headline":"MIRAC-5 upgrade promises JWST-class mid-IR contrast","feed_subtitle":"With a new dichroic, the imager should reach SNR=5 limits of 18.0 in L', 15.6 in M', and 12.6 in N' over eight hours.","key_machinery":"The load-bearing element is the GeoSnap 1024x1024 HgCdTe detector, sensitive from 2 to 13 microns with 65 percent average quantum efficiency, deep wells, and readout rates up to 85 Hz. Its main liability, 1/f noise, is handled by the instrument's internal pupil-plane chopper and telescope nodding: pair-subtracting nearby frames reduces the 1/f term below a tenth of the Poisson shot noise of the sky and telescope background, and the paper models this with an empirical power-law term, $sigma^{2}$_1/f = $g^{2}$ k_f (nu_chopper / nu_detector)^$\\alpha$, in the SNR equation. The second key component is the temporary 50:50 dichroic that currently limits throughput and adds background; the paper's projected sensitivities are computed for a planned replacement with greater than 90 percent transmission.","core_discovery":"On the paper's own terms, the central discovery is that a ground-based mid-IR imager built around the GeoSnap detector can be made to work on a 6.5-m telescope: the measured system throughput is roughly 5 to 13 percent across the L' through N' bands, the N' point-spread function already has a 0.75 Strehl compared to an ideal Airy pattern, and the noise in chopped-and-nodded data follows 1/sqrt(N) for at least 83 minutes of integration. The same data calibrate an exposure time calculator whose predictions agree with the May 2024 Alpha Boo observations to within 10 percent. The paper's projected performance, namely L', M', and N' limiting magnitudes of 18.0, 15.6, and 12.6 at SNR=5 in 8 hours after the dichroic swap, is the quantitative statement that would make MIRAC-5 competitive for warm-companion and disk science.","pith_inferences":["If the 8-hour 1/sqrt(N) extrapolation fails only mildly, MIRAC-5 would still outperform VISIR without adaptive optics, but it might fall short of the quoted 18.0 L' limit; a dedicated long-integration test would settle this.","The same 1/f suppression by rapid modulation should carry over to GeoSnap-based instruments on extremely large telescopes, where chopper overhead may favor slower chop frequencies than naive noise arguments suggest.","The measurement that telescope and instrument emission dominates over atmospheric emission in the N-band implies that reducing dichroic emissivity is more valuable for this instrument than observing at a drier site.","A testable extension is that, because 1/f noise scales with number of frames rather than elapsed time, the optimal chop frequency depends on detector frame rate; future instruments could use the paper's equations to pick frequencies that maximize real SNR per wall-clock hour."],"forward_implications":["If the throughput and noise scaling hold, MIRAC-5 can observe warm wide-orbit companions now, before the adaptive optics and coronagraph are fully commissioned.","After the dichroic replacement, L' and M' observations will be limited by ambient dark current and telescope and instrument emission, with limiting magnitudes of 18.0 and 15.6.","With the AGPM coronagraph and MAPS adaptive optics, the instrument should reach contrast-limited performance comparable to JWST's MIRI for close-in companions, enabling searches for ammonia at 10.6 microns.","The calibrated exposure time calculator lets observers optimize chopper frequency and nod timing, since a 1 Hz chopper gives lower total observing time than higher frequencies for N' half-well data.","Observing efficiency in chop and nod mode is around 90 percent, and avoiding temporal co-adding preserves the 1/sqrt(N) noise scaling."],"supporting_citations":[{"why":"Provides the GeoSnap detector characterization, including quantum efficiency, gain, read noise, and 1/f noise behavior, that the paper's noise model and throughput calculations build on.","marker":"Leisenring et al. 2023"},{"why":"Defines the MIRAC-5 cryostat optical layout, filter set, and planned ammonia science that this commissioning paper updates.","marker":"Bowens et al. 2022"},{"why":"Supplies the Alpha Boo spectral-model magnitudes used to turn aperture photometry into per-band throughputs and zero-magnitude references.","marker":"Cohen et al. 1995"},{"why":"Provides the NEAR background-limited and contrast performance baseline used to argue MIRAC-5's competitiveness.","marker":"Wagner et al. 2021"},{"why":"Gives the NIRC2 M-band limiting magnitude used for comparison and the caution that multi-hour integrations may deviate from 1/sqrt(t) scaling.","marker":"Bowens-Rubin et al. 2023"},{"why":"Provides the PYNPOINT routines used for PSF fitting, frame alignment, and image combination in the photometric analysis.","marker":"Stolker et al. 2019"},{"why":"Underlie the ESO SkyCalc atmospheric emission and transmission model used by the exposure time calculator.","marker":"Noll et al. 2012; Jones et al. 2013"},{"why":"Describes the MAPS adaptive secondary system whose commissioning status drives the paper's image-quality expectations.","marker":"Morzinski et al. 2020"}],"fun_headline_variants":["MIRAC-5 shows close to diffraction-limited N-band","Measured 10% system throughput for MIRAC-5 on MMT","Chopping and nodding beat GeoSnap 1/f noise","MIRAC-5 exposure calculator matches on-sky data"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the noise continues to scale as 1/sqrt(N) from the measured 83-minute baseline out to 8-hour integrations, and that the empirical per-band corrections fitted to one night of Alpha Boo data remain valid for the new dichroic and for other observing conditions.","fun_headline_variants_meta":{"raw":{"variants":["MIRAC-5 shows close to diffraction-limited N-band","Measured 10% system throughput for MIRAC-5 on MMT","Chopping and nodding beat GeoSnap 1/f noise","MIRAC-5 exposure calculator matches on-sky data"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1778,"prompt_tokens":1123,"completion_tokens":655,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":739,"completion_tokens_details":{"reasoning_tokens":580}},"tokens_in":739,"tokens_out":655,"duration_ms":6767,"temperature":1.0,"reasoning_tokens":580,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:14:27.276348+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take an 8-hour N' integration with the new dichroic installed, reduce it without temporal co-adding, and compare the median per-pixel MAD against the 1/sqrt(N) trend from Figure 5; if the noise floor departs from that trend, as the paper itself warns is possible, the quoted limiting magnitudes would need to be revised. A shorter check would be to compare the exposure time calculator's predicted background and throughput for L', M', and N' against a single night of photometry of a standard star taken after the dichroic swap.","supporting_citations":[{"cited_title":"M., Atkinson, D., Bowens, R., et al","cited_arxiv_id":null,"evidence_quote":"Provides the GeoSnap detector characterization, including quantum efficiency, gain, read noise, and 1/f noise behavior, that the paper's noise model and throughput calculations build on."},{"cited_title":"2022, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol","cited_arxiv_id":null,"evidence_quote":"Defines the MIRAC-5 cryostat optical layout, filter set, and planned ammonia science that this commissioning paper updates."},{"cited_title":"C., Walker, R","cited_arxiv_id":null,"evidence_quote":"Supplies the Alpha Boo spectral-model magnitudes used to turn aperture photometry into per-band throughputs and zero-magnitude references."},{"cited_title":"M., Montoya, M., Fellows, C., et al","cited_arxiv_id":null,"evidence_quote":"Describes the MAPS adaptive secondary system whose commissioning status drives the paper's image-quality expectations."}],"review_version":1}