{"id":"c077b7b6-c490-4931-80bf-282f2b1a073e","arxiv_id":"2608.03789","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A three-pixel on-chip spectrometer array was deployed on the 50 m Large Millimeter Telescope, detecting CO(2-1) in NGC253 and mapping Orion KL, with 106 mK√s on-sky noise.","lead":"Researchers installed a three-pixel array of SuperSpec on-chip millimeter-wave spectrometers into the MUSCAT camera at the Large Millimeter Telescope, and took first on-sky data in summer 2025. They detected the CO(2-1) line in the galaxy NGC253, mapped the Orion KL region, and measured a per-detector noise of about 106 mK√s, a pathfinder step toward large line-intensity-mapping focal planes.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported NET and line-strength calibration rest on an unvalidated beam-dilution correction (Sec. 4.2 footnote); this is the load-bearing weak point.","rationale":"The reader's conditional verdict is appropriate. I focused on the calibration chain because it is the only place where all quantitative outputs—sensitivity and line intensities—depend on a correction that is asserted, not derived. The paper's independent evidence (hot/cold load responsivity, beam maps, known spectral line positions) shows the instrument works qualitatively, but does not constrain the beam-dilution factor. I considered the lack of error bars on the NGC253 spectrum; that is real but secondary: it affects the significance of one detection, whereas the calibration issue affects the entire quantitative payload. I also considered that only the central pixel is analyzed; that is a scope limitation, not a flaw in the demonstrated physics. The requested check would settle whether the beam-dilution footnote is a minor omission or a source of substantial bias.","tokens_in":13243,"tokens_out":11389,"duration_ms":140067,"concrete_test":"Recompute G_k for at least 10 representative channels from the raw Neptune observations using a per-channel beam solid angle obtained from the 2D Gaussian fits in Fig. 9 plus a Neptune disk model, applying the dilution correction explicitly as Ω_beam/Ω_source (or its correct form). Compare the resulting median NET (timestream and map) with the reported 106/115 mK√s. If the values shift by more than ~20% or the required correction varies by more than a factor of ~2 across the 198–296 GHz band, the quoted sensitivity and line intensities are not robust. Also verify that the same correction is used in the map-based NET; if the correction is absent there, the two values agreeing is not evidence of correctness.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative core of the paper—the on-sky NET (106/115 mK√s) and the calibrated line strengths—is set by the per-detector gain G_k derived from Neptune point-source observations (Sec. 4.2). The only information given about this calibration is the footnote: \"Since the gains are derived from point-source observations, a beam dilution correction must be applied.\" No formula, no per-channel beam model, and no validation is provided. This matters because Fig. 9 shows 5–10 arcsec FWHMs that vary with channel, while Neptune is ~2 arcsec and NGC253/Orion structures are extended. A single scalar dilution factor—or one derived without per-channel beam solid angles—would propagate directly into R_k = G_k^{-1} in Eq. 1 and into the map calibration, biasing both NET estimates. The stated consistency between timestream and map NET (106 vs 115 mK√s) cannot validate the absolute scale, since both use the same G_k. The CO line detection itself may survive a calibration error, but the paper's quantitative performance and line-intensity claims do not. Since the reduction is deferred to Lapuente et al. (2026), this preprint is not self-contained at exactly the point where the headline numbers are set.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the summer 2025 deployment and on-sky commissioning of three SuperSpec on-chip filterbank spectrometers installed in the MUSCAT camera at the Large Millimeter Telescope. It describes the cryogenic and readout modifications, hot/cold-load responsivity and noise characterization, beam measurements from point sources, an on-sky NET estimate of 106 mK√s from timestreams and 115 mK√s from maps, a claimed CO(2-1) detection in NGC253, and multi-frequency maps of Orion KL, concluding that SuMAC is a pathfinder for on-chip spectroscopy on large telescopes.","tokens_in":13560,"tokens_out":5032,"duration_ms":56379,"significance":"If the calibration chain holds, this is a significant instrument demonstration: it would be the first on-sky spectroscopy with multi-pixel on-chip filterbank spectrometers on a 50-m class telescope. The measurements are grounded in several independent external references — liquid-nitrogen/room-temperature loads, Neptune point-source observations, the known CO(2-1) rest frequency, and SCUBA-2 maps — and the paper does not fit constants to its own conclusions. The beam characterization and the noise versus optical-load trends are useful system-level data for future SuperSpec-class instruments. However, the quantitative claims are conditional on an unpublished companion-paper reduction and on an unquantified beam-dilution correction, so the significance is pathfinder-level rather than fully demonstrated.","major_comments":[{"comment":"The on-sky NET and the calibrated line strengths depend on per-detector gains G_k derived from Neptune point-source observations. The only information given about the required correction is the footnote: \"Since the gains are derived from point-source observations, a beam dilution correction must be applied.\" No formula, no per-channel beam solid angle, and no validation are provided. Fig. 9 shows beams with FWHMs varying from 5–10 arcsec across channels, while Neptune is point-like and NGC253/Orion are extended. An incorrect scalar dilution factor propagates directly into R_k = G_k^{-1} in Eq. (1) and into the map calibration, biasing both the reported NET values (106/115 mK√s) and the line intensities. The consistency between timestream and map NET cannot validate the absolute scale because both use the same G_k.","section":"Sec. 4.2, Eq. (1) and footnote; Fig. 9"},{"comment":"The central spectroscopic claim — the CO(J=2→1) detection in NGC253 — is presented without error bars, without a fitted line profile, and without a signal-to-noise ratio or line-integrated intensity. The text says the line is \"highlighted\" and \"detected,\" but the figure alone does not quantitatively establish a detection. The authors should show the spectrum with 1σ per-channel uncertainties, report the fitted line amplitude, width, and significance, and state the expected line frequency/velocity used for the identification. Without these, the detection claim is not quantitatively supported.","section":"Sec. 4.2.1, Fig. 11"},{"comment":"The reduction and calibration are deferred entirely to four unpublished companion papers by the same collaboration: Kane et al. (2026) for filterbank FTS characterization, Lapuente et al. (2026) for flux calibration and data reduction, Redford et al. (2026) for noise/common-mode analysis, and Savorgnano et al. (2026) for LIM scans. The manuscript is therefore not self-contained exactly at the points where the headline numbers are set: the Neptune gain calibration, the beam-dilution correction, the atmospheric opacity correction, and the map-making parameters. At minimum, the calibration chain and the map-making procedure must be summarized in sufficient detail to be checked, or the companion papers must be made available to the referee and the reader.","section":"§1, §4.2, §4.2.1; Refs. [6]–[9]"}],"minor_comments":[{"comment":"The hot load is given as 287 K in Eq. (3.1) and in Fig. 7, but Sec. 5 says \"cold (77 K) and a hot (300 K) optical load.\" Please make the temperatures consistent.","section":"Sec. 3.1 vs Sec. 5"},{"comment":"The conversion \"10 mK/√Hz = 7.1 mK√s\" is dimensionally incorrect: 10 mK/√Hz equals 10 mK√s. The factor 1/√2 would need an explicit bandwidth or one-sided/two-sided PSD convention. This should be corrected or explained.","section":"Sec. 4.2"},{"comment":"Add explicit units to the spectral axis and color scale, and state any smoothing, binning, or filtering applied to the spectra and maps. Without this, the reader cannot assess the effective spectral resolution or the significance of features.","section":"Figs. 11 and 12"},{"comment":"Reference [8] (Redford et al.) is listed with the same title as reference [6] (Kane et al.), although the text describes [8] as the noise/common-mode analysis. Please correct the title and ensure each companion paper is uniquely identified.","section":"References"},{"comment":"The abstract lists \"preliminary line intensity mapping scans\" as a highlight, but the paper gives no LIM results or noise analysis beyond a forward reference to Savorgnano et al. (2026). Either add a brief quantitative result or remove this from the abstract.","section":"Abstract and Sec. 3.2"}],"recommendation":"major_revision","confidential_remarks":"The paper sits at the boundary of a proceedings-style instrument report and a journal article. The core hardware demonstration is credible, but the manuscript leans almost entirely on four companion papers from the same collaboration for calibration and reduction. For a journal, the editor should require that the calibration chain be made self-contained or the companion papers released, and that the detection significance be quantified. The self-citation pattern is not deceptive, but it is heavy: of the key method references, [6]–[9] are all by the SuMAC team, and some are not yet available."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know up front: this is a credible, incremental instrument paper, not a breakthrough. SuMAC is the first deployment of on-chip filterbank spectrometers on a 50-m class telescope, and the authors show a convincing CO(2-1) detection in NGC253 and multi-frequency maps of Orion. That part is real and worth having.\n\nWhat is new: three SuperSpec pixels (110 channels each for two devices, 50 for the third) operated at 95 mK in the MUSCAT cryostat at the LMT, with on-sky beam widths of 5-10 arcsec that track the diffraction limit, and a measured per-detector NET around 106-115 mK√s. The hot/cold load responsivity measurements are clearly presented. The paper cites DESHIMA and SPT-SLIM properly, so the novelty is the specific LMT deployment, not the concept.\n\nThe soft spots are exactly where the headline numbers are set. Section 4.2 estimates on-sky NET using per-detector gains derived from Neptune, with only a footnote saying 'a beam dilution correction must be applied.' No formula, no per-channel beam model, no validation. Since beam FWHMs vary from 5 to 10 arcsec across channels and Neptune is ~2 arcsec, the absolute scale of the NET and the calibrated line intensities in Fig. 11 depend on that unquantified correction. The stated consistency between timestream and map NET does not validate the scale because both use the same G_k. That is a load-bearing gap, though likely not a fatal one for the detection itself.\n\nSecond, the NGC253 spectrum in Fig. 11 has no error bars and no significance estimate. For a paper whose highlight is a line detection, that is a surprising omission. Third, there is a small unit conversion slip in Section 4.2: 10 mK/√Hz equals 10 mK√s, not 7.1. That looks like a typo but should be fixed. Fourth, most of the reduction is deferred to four unpublished companion papers (Lapuente, Redford, Kane, Savorgnano), which makes the preprint not self-contained at exactly the point where the calibration matters.\n\nOverall, the central claim—that SuMAC performed first on-sky spectroscopy with a multi-pixel SuperSpec array at a large telescope—holds. The paper is honest about the poor weather and non-optimized optics, and the internal consistency of the measurements is good. It deserves a serious referee, but the referee should ask for the beam-dilution treatment and the significance of the line detection before accepting the quantitative claims. The unit conversion should also be corrected.\n\nI would cite this if I were writing a SuperSpec follow-up, and I might bring it to a reading group focused on mm-wave instrumentation. Recommend: send to peer review with the calibration questions.\n\nBest","headline":"A credible, incremental on-sky demo of a three-pixel SuperSpec spectrometer at the LMT with a real CO detection, but the headline NET and line-flux numbers rest on an unquantified beam-dilution correction deferred to companion papers.","tokens_in":14170,"tokens_out":4602,"would_cite":true,"duration_ms":50815,"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 reports the first on-sky deployment of a multi-pixel on-chip filterbank spectrometer on a large millimeter telescope. Three SuperSpec devices were integrated into the MUSCAT camera at the Large Millimeter Telescope, and during a s","keywords":["On-chip spectroscopy","Kinetic Inductance Detectors","Millimetric instrumentation","Large Millimeter Telescope","Line Intensity Mapping","On-sky demonstration","SuperSpec","Filterbank spectrometer"],"falsifier":"Measure SuMAC's NGC253 CO(2-1) line flux against published IRAM or JCMT single-dish values, or compare the 230 GHz Orion KL map with an independently calibrated spectral map. Agreement would validate the beam-dilution correction; a source-size-dependent offset would show it is biased.","tokens_in":13203,"feed_emoji":"📡","tokens_out":6415,"duration_ms":65452,"temperature":0.7,"pith_summary":"The paper reports the first on-sky deployment of a multi-pixel on-chip filterbank spectrometer on a large millimeter telescope. Three SuperSpec devices were integrated into the MUSCAT camera at the Large Millimeter Telescope, and during a summer 2025 campaign the central pixel detected the CO(2-1) line at 230.5 GHz in NGC253, mapped Orion KL at multiple frequencies, and measured per-detector on-sky sensitivities of about 106 mK√s from timestreams and 115 mK√s from maps. The work is meant to show that integrated superconducting spectrometers can deliver moderate-resolution spectra and spectral maps on sky, laying groundwork for larger arrays and line intensity mapping.","feed_headline":"On-chip spectrometer maps Orion and detects CO in NGC253","feed_subtitle":"SuMAC's three on-chip pixels reached ~106 mK√s sensitivity, a first multi-pixel spectroscopy run on a large telescope.","key_machinery":"The central object is the SuperSpec on-chip filterbank spectrometer pixel: each pixel's broadband dual-slot antenna feeds a lithographed bank of half-wave microstrip resonators, one per spectral channel, and each resonator absorbs a narrow frequency slice and passes it to a titanium-nitride kinetic inductance detector whose resonant frequency shifts with absorbed power. This replaces the gratings and feedhorns of a conventional spectrometer with lithography, so the spectral resolution of about R≈200 is set by resonator quality factors. The argument depends on per-detector responsivity measured directly from hot/cold loads and on gains derived from Neptune point-source observations, which con","core_discovery":"SuMAC demonstrates that SuperSpec's on-chip filterbank—where dispersion and detection are integrated on a single wafer—can be operated as a multi-pixel spectroscopic instrument at a 50-meter telescope. By mounting three pixels in the MUSCAT cryostat and adjusting the readout chain to cover SuperSpec's 80–450 MHz band, the collaboration obtained 77 usable spectral channels on the central pixel across 198–296 GHz. With those channels it detected the CO(J=2→1) emission line at 230.5 GHz in the starburst galaxy NGC253 and imaged the Orion Molecular Cloud's KL region at several frequencies, with the bright structures matching SCUBA-2 850 μm continuum maps. The median on-sky NET of roughly 106–115","pith_inferences":["If the roughly fifteen-fold aperture-to-sky sensitivity gap is dominated by unoptimized optics and mirror losses, redesigning the optical chain and adding a proper beam chopper could bring on-sky NET close to the cold-load value of ~10 mK/√Hz, making small on-chip arrays competitive for targeted line work.","The NGC253 CO(2-1) line flux could be cross-checked against published IRAM or JCMT measurements; a source-size-dependent offset would isolate the beam-dilution correction as the main systematic in the calibration.","A natural next test is a frequency-resolved map of a source with known line ratios, which would verify that per-channel relative calibration is stable across the full 190–300 GHz band rather than only at the brightest lines."],"forward_implications":["On-chip filterbank spectrometers can be installed in an existing continuum camera with only filter, aperture, amplifier, and readout changes, so future large-format spectroscopic focal planes could reuse current camera infrastructure.","The consistency between timestream- and map-derived NETs (106 and 115 mK√s) indicates that the calibration pipeline yields stable per-detector sensitivity estimates.","Detecting CO(2-1) in NGC253 and mapping Orion KL at several frequencies demonstrates moderate-resolution line spectroscopy and spectral mapping of both point-like and extended sources.","The preliminary line-intensity-mapping scans open a path toward LIM experiments with on-chip spectrometers on large single-dish telescopes."],"supporting_citations":[{"why":"First-light demonstration of an integrated superconducting spectrometer (DESHIMA), the prior on-sky validation this work builds on.","marker":"[3]"},{"why":"Companion paper characterizing the SuperSpec filterbanks used in SuMAC, supplying the per-channel spectral responses.","marker":"[6]"},{"why":"Companion paper describing the flux calibration and data-reduction pipeline, providing the gains and map-making used for the NET and spectra.","marker":"[7]"},{"why":"Design, characterization, and performance of SuperSpec on-chip spectrometers, the system being deployed.","marker":"[10]"},{"why":"MUSCAT focal plane performance, describing the host camera and readout infrastructure.","marker":"[11]"},{"why":"Previous CO detections in NGC253 that identify the observed 230.5 GHz line.","marker":"[17]"},{"why":"Orion KL hot-core study that identifies the expected CO(2-1) emission and structure used to interpret the SuMAC maps.","marker":"[18]"},{"why":"Full-array noise performance of deployment-grade SuperSpec, providing the baseline for expected detector noise at higher operating temperature.","marker":"[21]"}],"fun_headline_variants":["Multi-pixel on-chip spectroscopy debuts on 50-meter telescope","First multi-pixel mm-wave spectrometer on a large dish","On-chip spectrometer sees CO in NGC253 and Orion KL","Multi-pixel filterbank spectroscopy at LMT: NGC253 CO and Orion","Three wafers bring on-chip mm-wave spectroscopy to 50-m dish"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The per-detector gains come from Neptune point-source observations, and a footnote says a beam-dilution correction must be applied but never states how it is computed or validated; with beam widths varying from 5 to 10 arcseconds across channels and sources of different sizes, an error there would directly scale the reported NET and line intensities.","fun_headline_variants_meta":{"raw":{"variants":["Multi-pixel on-chip spectroscopy debuts on 50-meter telescope","First multi-pixel mm-wave spectrometer on a large dish","On-chip spectrometer sees CO in NGC253 and Orion KL","Multi-pixel filterbank spectroscopy at LMT: NGC253 CO and Orion","Three wafers bring on-chip mm-wave spectroscopy to 50-m dish"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000595,"raw_usage":{"total_tokens":2611,"prompt_tokens":718,"completion_tokens":1893,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":462,"completion_tokens_details":{"reasoning_tokens":1802}},"tokens_in":462,"tokens_out":1893,"duration_ms":15766,"temperature":1.0,"reasoning_tokens":1802,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:27:58.508185+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure SuMAC's NGC253 CO(2-1) line flux against published IRAM or JCMT single-dish values, or compare the 230 GHz Orion KL map with an independently calibrated spectral map. Agreement would validate the beam-dilution correction; a source-size-dependent offset would show it is biased.","supporting_citations":[{"cited_title":"First light demonstration of the integrated superconducting spectrometer,","cited_arxiv_id":null,"evidence_quote":"First-light demonstration of an integrated superconducting spectrometer (DESHIMA), the prior on-sky validation this work builds on."},{"cited_title":"Data reduction pipeline for the SuMAC millimeter-wave spectrometer at the LMT,","cited_arxiv_id":null,"evidence_quote":"Companion paper describing the flux calibration and data-reduction pipeline, providing the gains and map-making used for the NET and spectra."},{"cited_title":"Superspec: On-chip spectrometer design, characterization, and performance,","cited_arxiv_id":null,"evidence_quote":"Design, characterization, and performance of SuperSpec on-chip spectrometers, the system being deployed."},{"cited_title":"Mexico-uk submillimeter camera for astronomy focal plane performance at the large millimeter telescope,","cited_arxiv_id":null,"evidence_quote":"MUSCAT focal plane performance, describing the host camera and readout infrastructure."},{"cited_title":"Carbon monoxide in the starburst of NGC 253 revisited,","cited_arxiv_id":null,"evidence_quote":"Previous CO detections in NGC253 that identify the observed 230.5 GHz line."},{"cited_title":"Orion kl: the hot core that is not a “hot core","cited_arxiv_id":null,"evidence_quote":"Orion KL hot-core study that identifies the expected CO(2-1) emission and structure used to interpret the SuMAC maps."},{"cited_title":"Full-Array Noise Performance of Deployment- Grade SuperSpec mm-Wave On-Chip Spectrometers,","cited_arxiv_id":null,"evidence_quote":"Full-array noise performance of deployment-grade SuperSpec, providing the baseline for expected detector noise at higher operating temperature."}],"review_version":1}