{"id":"4dfff3bd-881c-4c62-a358-418087bea22e","arxiv_id":"2411.18738","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A 22-galaxy SOFIA infrared atlas with new 19.7-214 μm images and fluxes shows that the average nuclear SED peak is ~40 μm and hints that more luminous AGN peak at shorter wavelengths.","lead":"This paper releases a 20-214 μm imaging atlas of 22 nearby active galaxies observed with the SOFIA telescope, including 41 new images and measured nuclear fluxes. A generalist might read it because it provides the best-sampled mid-to-far-infrared spectral coverage of nearby AGN to date, a legacy dataset for studying the dusty structures around supermassive black holes.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Nuclear SED peak position is set by sparse photometry plus a Spitzer spectrum normalized to it; an unmodeled host contribution at 30–40 μm would shift many peaks below 40 μm before the '~40 μm average' claim is tested.","rationale":"I agree with the reader's weakest-assumption identification: the single-Gaussian host model in §4.1 is the structural soft spot, and the paper itself acknowledges its failure mode for NGC 4388. My stress-test adds that the concern propagates not only to the four extended sources but to the entire peak-wavelength result, because the 'nuclear SED' for the point-like majority is a few-arcsecond aperture measurement in which the same host contamination is unresolved rather than subtracted. The Figure 8 correlation is a secondary claim and is less load-bearing than the headline 40 μm peak, which drives the abstract's physical interpretation. My proposed test is deliberately cheap and model-based: it asks whether the headline survives plausible host-subtraction and calibration perturbations. I would keep the verdict at CONDITIONAL (not REJECT), because the atlas itself, the photometry, and the morphology descriptions are well documented and independently useful; only the secondary claims need reframing or caveating after the test.","tokens_in":30459,"tokens_out":1786,"duration_ms":15589,"concrete_test":"Reconstruct each nuclear SED from the released fluxes, then repeat the peak-wavelength determination twice: (i) with the 31.5/37.1 μm fluxes perturbed by ±10% (calibration plus host-subtraction uncertainty) and (ii) after replacing the single-Gaussian host model with a Moffat or multiple-Gaussian host fitted to the same images. If the median peak in νFν shifts below 35 μm, or if the Spearman correlation in Figure 8 loses significance (p > 0.05) in either variant, the '~40 μm average peak' and the luminosity–peak correlation should be explicitly downgraded.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central science claim (abstract and §5) is that the average nuclear SED peak is ~40 μm in νFν, attributed to AGN-heated extended dust. However, the peak wavelength is determined from only 2–4 SOFIA photometry points (19.7–53 μm), a Spitzer spectrum, and generally 1–2 Herschel points; the SED bins are coarse, and for the four extended sources the 30–40 μm nuclear flux relies on the §4.1 PSF-plus-single-Gaussian decomposition. The reader already flags this; the paper itself limits NGC 4388 in this range. The deeper issue is that for the majority of the 22 objects, the 'nuclear SED' at 31.5–37.1 μm is simply an unresolved 3–4 arcsec aperture flux, so host-galaxy star-forming or NLR dust inside the beam is not actually separated. For the four extended sources, the 2D Gaussian host model is a single elongated Gaussian; bars, rings, and dust lanes at 10–100 pc scales are known in many of these galaxies (NGC 2273 star-forming ring, NGC 7469 circumnuclear ring, NGC 3081 resonance rings), so the extracted nuclear flux is a model-dependent quantity that can be biased low or high depending on whether the host is oversubtracted. A direct test is to re-derive the SEDs with a 10–20% host-subtraction error applied to the 31.5 and 37.1 μm points; the claimed ~40 μm average peak is a one-point shift away from ~30 μm in several objects. The correlation with L_bol (§5.1, |R|~0.63, p=0.0015) is presented without errors on λ_peak and without a selection-function check despite the sample being X-ray flux-limited and incomplete at the high-luminosity end.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a SOFIA/FORCAST and HAWC+ imaging atlas of 22 nearby Seyfert AGN at 19.7–214 μm, supplemented by archival Herschel 70–500 μm data. The authors provide photometry and images for 69 observations, 41 of them previously unpublished, and classify most sources as point-like while identifying four extended systems (Centaurus A, Circinus, NGC 1068, NGC 4388). Resolved nuclear fluxes are obtained either by aperture photometry or by a PSF-plus-single-2D-Gaussian decomposition for extended emission. The paper constructs the resulting mid- to far-IR SEDs, reports that the average peak wavelength in νFν is about 40 μm, associates this with unresolved AGN-heated dusty emission, and claims a correlation between bolometric luminosity and peak wavelength. It also presents a mid- to far-IR color-color diagram and discusses AGN versus star-formation dominated sources.","tokens_in":30818,"tokens_out":6017,"duration_ms":58985,"significance":"If the central claims hold, the atlas is a valuable legacy dataset for studying the mid- to far-IR properties of local AGN, because it provides homogeneous SOFIA imaging across a wide wavelength range and the best-sampled 30–500 μm SEDs available for this sample. The paper's strengths are its careful documentation of the reductions, the explicit treatment of calibration and background uncertainties, and the public release of 41 new images. The conclusion that the typical νFν peak is near 40 μm is physically relevant for torus and dusty-outflow models, as it points to dust at roughly 70–100 K, and the color analysis provides a useful comparison with larger samples. However, the quantitative peak-wavelength claim is not yet supported to the precision implied by the abstract, because the peak positions are selected from a sparse wavelength grid and are sensitive to the host-subtraction assumptions. The atlas value alone supports publication after the central SED claims are either strengthened with an uncertainty analysis or softened.","major_comments":[{"comment":"The extraction of unresolved nuclear fluxes for the four extended sources relies on an 11-parameter model in which the host is a single elongated 2D Gaussian plus a constant background, with the unresolved nucleus represented by the standard-star PSF. For Centaurus A, Circinus, NGC 1068, and NGC 4388, known host structures such as dust lanes, bars, and circumnuclear rings are not representable by one Gaussian, so the fitted PSF amplitude can be biased in either direction depending on whether the host is over- or under-subtracted. Because the SED peaks and the L_bol–λ_peak correlation in Sections 5 and 5.1 depend on these fluxes, please add a robustness test that varies the host model (for example, a two-Gaussian host or a PSF-amplitude perturbation on the 31.5–53 μm images) and quantify the resulting shift in the nuclear fluxes and in λ_peak.","section":"§4.1 and Figure 14"},{"comment":"The peak wavelength is determined from a coarse and non-uniform grid consisting of the Spitzer/IRS spectrum plus photometry at 31.5, 37.1, 53, 70, 89, and longer wavelengths. For most objects the maximum in νFν falls either on the last covered Spitzer wavelength or on one of the SOFIA/Herschel band centers, so the abstract's statement that the 'average peak wavelength' is about 40 μm is substantially a statement about the chosen wavelength grid rather than about a measured continuum peak. No per-object uncertainties on λ_peak are reported. Please report λ_peak with uncertainties, for example by bootstrap-perturbing the measured fluxes within the quoted errors and recomputing the peak, or by fitting a smooth interpolation to the νFν points; also state how the sample average changes if the 31.5 μm or the 37.1 μm point is dropped for objects where the peak falls on that point.","section":"§5 and Figure 7"},{"comment":"For NGC 4388, the SED points at 31.5 and 37.1 μm are total aperture fluxes, not PSF-subtracted nuclear fluxes, because the text states that almost all of the extended emission lies within the FWHM and that total fluxes are therefore used for the SED in the 30–40 μm range. These points include NLR dust and possibly host-galaxy emission and are not directly comparable with the PSF-extracted nuclear fluxes at longer wavelengths. Since this object contributes a short-wavelength peak to the sample average, the peak and the average should be re-evaluated using either aperture-matched measurements at all wavelengths or an explicit correction for the known NLR/host contribution.","section":"§5 and §4.2, NGC 4388"},{"comment":"The claimed correlation between L_bol and λ_peak (|R|≈0.63, p=0.0015) is presented without error bars on λ_peak, without stating whether R is a Pearson or Spearman coefficient, and without testing how the discrete sampling of λ_peak affects the result. With 22 objects and λ_peak taking only a few discrete values (for example 18–20, 31.5, 37.1, 53, or 70 μm), the significance may be inflated by the grid rather than by a physical trend. Please add a Monte Carlo test that perturbs fluxes within their uncertainties, recomputes λ_peak and the correlation each time, and report the resulting distribution of R and p; also provide a version of Figure 8 with horizontal error bars on λ_peak.","section":"§5.1 and Figure 8"}],"minor_comments":[{"comment":"The word 'organzed' in the paragraph beginning 'The manuscript is organzed as follows' is a typo and should read 'organized'.","section":"§1"},{"comment":"The row for NGC 4388 at 31.5 μm appears to list 'HA F545' as the mission/program identifier even though the observation was made with FORCAST; this is likely a typo and should be corrected to the appropriate FORCAST mission identifier.","section":"Table 3"},{"comment":"The text states 'Half (11) of the objects in the sample show ratios Fν(31)/Fν(70) < 1 while Fν(70)/Fν(160) > 1' and then lists 12 object names (Circinus, Mrk 231, Mrk 573, NGC 1275, NGC 2110, NGC 3081, NGC 3227, NGC 3281, NGC 4151, NGC 4941, NGC 5506, NGC 7469); the count or the list should be corrected.","section":"§5.2"},{"comment":"The text mentions that the Spitzer spectrum for NGC 1068 does not align with the SOFIA photometry because of PSF subtraction, but the same issue may affect other objects where Spitzer spectra are simply overplotted on photometry obtained with different apertures; a brief statement about the matching or non-matching of apertures for all objects would help readers interpret the SED peaks.","section":"§5, Figure 7"}],"recommendation":"major_revision","confidential_remarks":"The atlas itself is a useful contribution and the data products appear reliable, but the abstract's central quantitative claim about the average peak wavelength needs either a proper uncertainty treatment or a softer formulation. The requested robustness tests are feasible with the existing data and should not require new observations. I do not see any issue of circularity or model-fitting bias beyond the host-subtraction sensitivity already identified. The manuscript is likely to be acceptable after these revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a data-release paper, and the data release is the product. The 41 new SOFIA images plus reprocessed Herschel data, assembled into 30-500 um SEDs for 22 AGN, are genuinely useful. Anyone working on AGN torus and dust SEDs will want this atlas. The reductions are standard, the uncertainties are carried through, and the PSF-plus-Gaussian decomposition for extended sources is described well enough to be reproduced. Credit where due: they explicitly restrict the NGC 4388 SED to >40 um because the 30-40 um PSF subtraction is unreliable. That is the right call.\n\nThe soft spots are in the secondary science claims. The \"average peak ~40 um\" is more fragile than the abstract suggests. For most objects the peak is set by 2-4 SOFIA photometry points, a Spitzer spectrum normalized to them, and 1-2 Herschel points. The 31.5 and 37.1 um fluxes are aperture photometry on 3-4 arcsec scales, so host-galaxy or NLR dust inside the beam is not actually separated. For the four extended sources, the nuclear flux rests on a single elongated 2D Gaussian host model. A 10-20% shift in the 31.5 or 37.1 um points would move several peaks below 40 um. The paper acknowledges this for NGC 4388, but the same concern applies in milder form to the point-source SEDs. The abstract's phrasing \"unresolved extended dusty region heated by the AGN\" is an interpretation; these are arcsecond-scale SEDs, and the AGN-heating attribution is not demonstrated against host contributions.\n\nThe Lbol-peak correlation (|R|~0.63, p=0.0015) is presented without errors on peak wavelength and without any selection-function discussion. The sample is X-ray flux limited and incomplete at the high-luminosity end, so the correlation could be driven partly by sample construction. That does not invalidate the atlas, but it means the correlation should be framed as a tentative hint, not a measured result.\n\nWho is this for? AGN observers who need mid-to-far-IR photometry of nearby Seyferts. They will cite the tables and images. The paper deserves a serious referee; the data are valuable and the analysis is mostly careful. My recommendation is conditional acceptance: keep the atlas and photometry as the centerpiece, move the peak and correlation claims to a more explicitly exploratory register, and add the caveats about sparse wavelength sampling and host contamination to the abstract. A short revision fixes this.","headline":"Solid SOFIA/Herschel atlas of 22 nearby AGN; treat the ~40 um peak and luminosity correlation as suggestive, not measured to that precision.","tokens_in":31573,"tokens_out":1605,"would_cite":true,"duration_ms":17579,"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":"Nearby active galaxies emit most of their nuclear dust light near 40 microns, and the more luminous the nucleus, the shorter its spectral peak.","keywords":["active galactic nuclei","infrared imaging atlas","SOFIA observations","dust torus","spectral energy distribution","Seyfert galaxies","host galaxy subtraction","40 micron peak"],"falsifier":"Re-analyze the same images with a host model that includes a stellar bar and a circumnuclear ring; if the average $\\nu F_\\nu$ peak moves away from ~40 $\\mu$m or the luminosity–peak correlation disappears, the central result is an artifact of the Gaussian decomposition. A direct observational test is JWST/MIRI imaging at 20–28 $\\mu$m of NGC 1068 and NGC 4388, which would resolve out the extended NLR dust and reveal whether the compact nucleus truly peaks near 40 $\\mu$m.","tokens_in":30217,"feed_emoji":"🌌","tokens_out":10535,"duration_ms":84223,"temperature":0.7,"pith_summary":"This paper assembles the best-sampled 20–214 $\\mu$m imaging dataset of nearby active galactic nuclei to date: 69 SOFIA images of 22 Seyferts, 41 of them previously unpublished. By separating each nucleus from its host galaxy, it builds arcsecond-scale spectral energy distributions from 30 to 500 $\\mu$m and asks where AGN-heated dust actually radiates. The central result is that the average nuclear SED peaks near 40 $\\mu$m in $\\nu F_\\nu$, the energy-per-logarithmic-frequency measure, implying dust at roughly 70–100 K in an unresolved extended region rather than only in the classical hot torus. The paper also reports that peak wavelength shortens as bolometric luminosity rises, and that the edge-on Seyfert NGC 4388 changes from narrow-line-region-aligned dust at 30–40 $\\mu$m to host-galaxy-aligned dust at longer wavelengths.","feed_headline":"Active galaxies' dust shines brightest near 40 microns","feed_subtitle":"A 20–214 μm atlas of 22 Seyferts pinpoints where AGN-heated dust radiates and links peak wavelength to luminosity.","key_machinery":"The analysis is carried by the atlas itself: 69 SOFIA/FORCAST (19.7–40 $\\mu$m) and HAWC+ (53–214 $\\mu$m) images of 22 AGN, complemented by Herschel 70–500 $\\mu$m images and Spitzer/IRS spectra. The load-bearing tool is the flux extraction: point-like sources are measured with aperture photometry, while extended sources are fit with a two-component model in which the unresolved nucleus is a scaled standard-star point-spread function and the host galaxy is a single elongated 2D Gaussian plus a constant background, with 11 free parameters. This same model is applied to the Herschel images, producing the nuclear flux tables from which the SEDs, peak wavelengths, luminosities, and colors are all derived. PSF-subtracted residual images are what reveal extended AGN-heated structures such as the NLR dust in NGC 4388.","core_discovery":"The paper's central discovery claim is that the several-arcsecond nuclear emission of nearby Seyferts, measured homogeneously from 20 to 500 $\\mu$m, peaks on average at ~40 $\\mu$m in $\\nu F_\\nu$. This peak lies longward of the 10–30 $\\mu$m region where torus emission is usually characterized, and the authors attribute it to an unresolved extended dusty region heated by the AGN, with characteristic dust temperatures of 70–100 K. A secondary claim is that the peak wavelength anti-correlates with bolometric luminosity ($|R| \\sim 0.63$, $p = 0.0015$), so more luminous AGN show hotter dust dominating their infrared output. The atlas also demonstrates a clean separation in NGC 4388 between narrow-line-region dust at 30–40 $\\mu$m and host-galaxy dust at longer wavelengths.","pith_inferences":["If the 40 $\\mu$m peak is generic across the AGN population, surveys that use single 20–30 $\\mu$m bands (such as WISE 22 $\\mu$m) may undercount the bolometric dust output of Seyferts; comparing WISE W4 fluxes against the 40 $\\mu$m peaks in this atlas would test that directly.","The unresolved extended dusty region may be the pc-scale dusty wind or NLR dust seen in interferometry; a test would be to compare the ~40 $\\mu$m peak with subarcsecond interferometric sizes from instruments such as MATISSE to see whether the emitting region matches the dust sublimation radius or the NLR base.","The 2D Gaussian decomposition could be stress-tested on simulated images containing bars and rings; a re-analysis with a multi-component host model would either confirm or shift the reported nuclear fluxes and peak wavelengths."],"forward_implications":["The average ~40 $\\mu$m peak implies that models of AGN dust must reproduce a substantial far-infrared bump longward of the classical 10–30 $\\mu$m torus bands, meaning a significant fraction of the dust-reprocessed energy escapes in the 30–60 $\\mu$m range.","The luminosity–peak anti-correlation makes infrared color a rough luminosity diagnostic: brighter AGN will be identified by shorter-wavelength peaks and hotter dust.","The NGC 4388 morphology switch shows that wavelength-dependent imaging can separate AGN-heated narrow-line-region dust from star-forming host dust without spectroscopy, which can sharpen color-based AGN selection.","The elevated average $F_\\nu(70)/F_\\nu(160)$ ratio of 1.4 ± 0.7, compared with ~0.8 for larger star-formation-dominated samples, suggests far-infrared colors can pick out AGN-heated dust."],"supporting_citations":[{"why":"Established the 31.5 $\\mu$m photometry and the SED turnover beyond 31.5 $\\mu$m; this atlas extends that sample and method.","marker":"Fuller et al. 2016"},{"why":"Found PSF-subtracted extended emission at 37.1 $\\mu$m in Mrk 3, NGC 4151, and NGC 4388, directly motivating the NGC 4388 NLR interpretation and providing prior fluxes.","marker":"Fuller et al. 2019"},{"why":"Modeled the NGC 1068 torus with 20–53 $\\mu$m SOFIA and ALMA data, showing the torus peaks at 30–40 $\\mu$m with 70–100 K dust, the interpretive basis for the 40 $\\mu$m average peak.","marker":"Lopez-Rodriguez et al. 2018"},{"why":"Clumpy torus models whose predictions (shallow silicate features, Type 1/Type 2 SED similarity) motivate the need for longer-wavelength SED coverage and the sample selection.","marker":"Nenkova et al. 2008a,b"},{"why":"The original spectropolarimetric evidence for the obscuring torus, the physical picture the survey's dust studies target.","marker":"Antonucci & Miller 1985"},{"why":"GATOS survey paper defining the parent sample of molecular torus/outflow targets drawn from the Swift/BAT catalog.","marker":"García-Burillo et al. 2021"},{"why":"The 70-month Swift/BAT ultralhard X-ray catalog from which the flux-limited GATOS parent sample is selected.","marker":"Baumgartner et al. 2013"},{"why":"Reports the hard-X-ray vs mid-IR correlation for Type 1 AGN against which the ~18–20 $\\mu$m SED peaks are compared.","marker":"García-Bernete et al. 2017"},{"why":"Provides the $F_\\nu(70)/F_\\nu(160)$ dust-temperature proxy and the ~0.8 average ratio for a larger sample used as the comparison for the atlas's 1.4 ratio.","marker":"Meléndez et al. 2014"},{"why":"Documents the FORCAST instrument and reduction pipeline used for the 20–40 $\\mu$m images and flux calibration.","marker":"Herter et al. 2012"}],"fun_headline_variants":["AGN dust peaks at 40 microns in new SOFIA-based atlas","Seyfert galaxy dust shines brightest at 40 μm","Bright AGN host hotter dust, shifting peak to shorter wavelengths","40-micron peak ties AGN dust to luminosity in 22 Seyferts"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extracted nuclear fluxes and the ~40 $\\mu$m peak rest on the assumption that every unresolved nucleus is a scaled point-spread function sitting on a single smooth elliptical host component, so any real host structure (bars, rings, dust lanes) that the Gaussian cannot represent will bias the measured SED peak and the luminosity correlation.","fun_headline_variants_meta":{"raw":{"variants":["AGN dust peaks at 40 microns in new SOFIA-based atlas","Seyfert galaxy dust shines brightest at 40 μm","Bright AGN host hotter dust, shifting peak to shorter wavelengths","40-micron peak ties AGN dust to luminosity in 22 Seyferts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000283,"raw_usage":{"total_tokens":1755,"prompt_tokens":1115,"completion_tokens":640,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":731,"completion_tokens_details":{"reasoning_tokens":562}},"tokens_in":731,"tokens_out":640,"duration_ms":6185,"temperature":1.0,"reasoning_tokens":562,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:55:35.683059+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-analyze the same images with a host model that includes a stellar bar and a circumnuclear ring; if the average $\\nu F_\\nu$ peak moves away from ~40 $\\mu$m or the luminosity–peak correlation disappears, the central result is an artifact of the Gaussian decomposition. A direct observational test is JWST/MIRI imaging at 20–28 $\\mu$m of NGC 1068 and NGC 4388, which would resolve out the extended NLR dust and reveal whether the compact nucleus truly peaks near 40 $\\mu$m.","supporting_citations":[{"cited_title":"2016, , 462, 2618, 10.1093/mnras/stw1780","cited_arxiv_id":null,"evidence_quote":"Established the 31.5 $\\mu$m photometry and the SED turnover beyond 31.5 $\\mu$m; this atlas extends that sample and method."},{"cited_title":"2017, , 469, 110, 10.1093/mnras/stx795","cited_arxiv_id":null,"evidence_quote":"Reports the hard-X-ray vs mid-IR correlation for Type 1 AGN against which the ~18–20 $\\mu$m SED peaks are compared."},{"cited_title":"L., Adams , J","cited_arxiv_id":null,"evidence_quote":"Documents the FORCAST instrument and reduction pipeline used for the 20–40 $\\mu$m images and flux calibration."}],"review_version":1}