{"id":"45a80369-9cd6-4c0a-9369-dbf3db2dbd24","arxiv_id":"2412.07862","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A JWST-based census across 19 galaxies identifies 1,816 compact 3.3 micron PAH emitters as young embedded star clusters whose PAH-emitting phase lasts less than roughly 3 million years and whose inclusion raises the count of bright young clusters by 1.8x to 8.5x.","lead":"Using JWST infrared images of 19 nearby galaxies, astronomers identified 1,816 compact sources glowing at 3.3 microns, the signature of very young star clusters still wrapped in their dusty birth clouds. If correct, ordinary optical surveys are missing roughly 2 to 8 times as many bright young clusters as previously counted, and the hidden phase lasts only about 3 million years.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The ~3 Myr PAH fade timescale relies on an optically selected HST cluster sample as an unbiased age clock; the paper's own Sec 6.1 caveats show this sample excludes dusty embedded clusters by construction.","rationale":"The reader's weakest_assumption correctly targets the most load-bearing point. The paper is careful and self-aware, and the census itself (1816 sources, CI analysis, MIRI color cuts, correlation with SFR) is credible and consistent with prior single-galaxy studies. However, the headline inference about the embedded-phase duration depends on the assumption that the optically selected HST cluster sample is an unbiased age clock for the embedded population. The paper's own text undermines this assumption in two places: the blue/UV selection effect (Sec 6.1) and the age degeneracy at ≤3 Myr (Sec 6.1). Because the same concern applies to the Hα EW comparison (the comparison baseline is also the optically selected young clusters), the ~3 Myr fade timescale is not fully supported. This warrants the CONDITIONAL verdict as given; no adjustment is needed. The suggested test would settle whether the concern actually lands.","tokens_in":38853,"tokens_out":4646,"duration_ms":65076,"concrete_test":"Recompute the fraction of 4-10 Myr PHANGS-HST clusters with F300M-F335M > 3σ using only clusters selected via the F814W band (no blue/UV detection requirement), to reduce the bias toward unobscured clusters. If the fraction remains at the ~13% level reported in Sec 6.1/Table 3, then optical selection bias is not the dominant driver of the fade timescale; if the fraction rises substantially, the ~3 Myr conclusion is not supported. A complementary check: fit SED ages for the PAH emitters that are detected in HST broad bands (category 3 of Sec 7.1); if a significant number of these PAH emitters have fitted ages >3 Myr, then PAH emission does not fade within ~3 Myr.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central timescale claim ('PAH emission fades within ~3 Myr', abstract and Sec 7.1) is derived by comparing compact PAH emitters to PHANGS-HST clusters with SED-derived ages. The comparison in Sec 6.1 shows that only ~30% of HST clusters younger than 3 Myr fall in the PAH-excess region, while only ~13% of 4-5 Myr HST clusters do. The problem is that the HST cluster catalog is optically selected: clusters are detected in NUV-U-B-V-I imaging. The paper itself concedes that 'clusters detected in blue or even UV filters are not heavily extinguished, and thus may be associated with lower than average circum- and interstellar medium' (Sec 6.1). Consequently, any cluster that remains embedded for 4-10 Myr is absent from the HST sample by construction. What the data actually show is that optically visible clusters lose their PAH excess by ~3 Myr; they do not measure the duration of the embedded phase for the general cluster population. The paper also notes that 'UV-optical colors of ≤3 Myr clusters are highly degenerate' (Sec 6.1), so the age bins themselves are uncertain. Without a clock that includes embedded objects (e.g., Pa-alpha or spectroscopy of the PAH emitters), the headline fade-timescale is not established. This is the most load-bearing weakness; the 1,816-source census and 87% cluster-like CI are not affected by this concern.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a census of compact 3.3 micron PAH emitters in 19 PHANGS-JWST galaxies, identifying 1816 sources via F300M-F335M color excess, per-galaxy 5-sigma thresholds, and mid-IR color cuts to remove evolved stars. The authors characterize the sources with concentration indices, spatial distributions, SEDs, H-alpha equivalent widths, and luminosity functions, and compare them with PHANGS-HST cluster catalogs. The headline results are that PAH emission fades within about 3 Myr and that dusty embedded clusters may increase the number of optically visible <=3 Myr clusters in PHANGS-HST by a factor of 1.8-8.5.","tokens_in":39134,"tokens_out":2750,"duration_ms":28650,"significance":"If the timescale claim holds, this is a valuable systematic measurement of the duration of the compact PAH-emitting phase and the first large, multi-galaxy census of embedded cluster candidates selected at 3.3 microns. The paper's strengths include the per-galaxy treatment of detection limits and color dispersions, contamination cuts validated against an external LMC sample, a transparent presentation of the selection criteria, and an explicit discussion of the bright-end LF caveats. The catalog itself and the characterization of 1816 sources are likely to be useful reference products for the PHANGS community and for future JWST studies of embedded star formation.","major_comments":[{"comment":"The central claim that \"PAH emission fades within ~3 Myr\" is not established for the general embedded population, because the age clock is the optically selected PHANGS-HST cluster sample. As the paper itself concedes in Sec. 6.1, clusters detected in blue or UV filters are not heavily extinguished, so the HST sample is biased by construction toward clusters that have already cleared their dust. The data show that optically visible clusters lose their F300M-F335M excess by ~3 Myr, but they do not constrain how long clusters that remain embedded continue to emit PAHs. This is load-bearing for the abstract and Section 7.1; the conclusion should be reframed as a constraint on the PAH-emitting phase of optically detected clusters, or supported by an independent age or embedded-phase indicator such as Pa-alpha or mid-IR spectroscopy.","section":"Sec. 6.1 and Sec. 7.1 (Figs. 12-14, Table 3)"},{"comment":"The age comparison relies on the reliability of SED-derived ages in exactly the range where the paper says ages are most uncertain: the text notes that UV-optical colors of <=3 Myr clusters are highly degenerate and that 3-5 Myr clusters with A_V >~ 1 can be confused with 5-10 Myr clusters. Because the evolutionary statement depends on separating <=3 Myr from 4-10 Myr, the paper should quantify the effect of these age uncertainties on the reported fractions, for example by propagating SED-fitting age posteriors or showing the sensitivity of the Table 3 fractions to alternative age cuts.","section":"Sec. 6.1 (age bins used in Table 3 and Figs. 12-14)"},{"comment":"The factor of 1.8-8.5 increase in the number of <=3 Myr clusters is presented as a headline result, but it depends on treating bright PAH emitters as clusters and on the relative completeness of the PAH and HST cluster samples. The paper correctly notes that only the bright end is used and that fainter sources are incomplete, but Fig. 15 shows the PAH LF turning over near the completeness limit and having a steeper bright-end slope than the HST LF. A quantitative completeness and contamination treatment for the overlap region, or an explicit upper/lower bound derived from the CI distribution and the LMC contamination benchmark, is needed before this factor can be taken as a robust measurement rather than an illustrative estimate.","section":"Sec. 7.2 (LF-based increase factor)"}],"minor_comments":[{"comment":"The phrase \"identify of 1816 sources\" should read \"identify 1816 sources.\"","section":"Abstract"},{"comment":"The word \"Neverthelsss\" is misspelled; it should be \"Nevertheless.\"","section":"Sec. 6.1"},{"comment":"The text \"age < 107 Myr\" appears to have a typesetting error; it should presumably be \"age < 10^7 yr\" or \"age < 10 Myr,\" and the caption for that figure should state the intended meaning.","section":"Sec. 6.3"},{"comment":"The gray region denotes the absence of age estimates for PAH emitters, but the emitters are placed to the left of the HST clusters to suggest an evolutionary sequence. The caption should state explicitly that this horizontal placement is an assumption, not an age measurement.","section":"Fig. 17 and Fig. 18"},{"comment":"The two references listed as Whitmore et al. 2023a and 2023b appear to be identical; please check and correct the bibliographic entries.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The census and source characterization are solid and likely publishable after revision. The main issue is that the headline timescale is overinterpreted given the optically selected HST comparison sample, and the LF-based boost factor needs a more explicit incompleteness treatment. I would suggest the editor ask the authors to reframe the timescale as applying to optically visible clusters or to obtain an independent embedded-phase clock before the abstract claims a general ~3 Myr PAH fade time."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a careful extension of a single-galaxy method to 19 galaxies, and the census itself is worth having. The paper does not fully establish the ~3 Myr PAH fade timescale, though; that claim leans on optically selected HST cluster ages, and the paper's own Section 6.1 admits the selection bias. The census and the 1.8–8.5x boost factor are the real contribution, not the timescale.\n\nWhat's new: the 1,816-source catalog with per-galaxy 5-sigma color thresholds and detection limits, the contamination cuts validated against the LMC sample, the concentration index analysis (87% cluster-like), and the H-alpha EW comparison that independently suggests the PAH emitters are younger than HST clusters. The luminosity function analysis and the per-galaxy boost factor quantification also go beyond prior single-galaxy work. These are solid, reproducible steps: the selection recipe is transferable, and the external LMC benchmark is a good check.\n\nThe soft spots are concentrated in the timescale argument. The comparison in Section 6.1 uses HST clusters with SED ages to say PAH emission fades within ~3 Myr. But those clusters are, by construction, visible in NUV-U-B-V-I; heavily embedded clusters that stay embedded longer never enter the sample. The paper notes the UV-optical age degeneracy for <=3 Myr clusters and admits that blue/UV-detected clusters are not heavily extincted. So the data demonstrate that optically visible clusters lose PAH excess fast; they do not measure the embedded phase for the general population. The H-alpha EW comparison is suggestive, but it is not calibrated to an independent age scale for the embedded objects. The boost factor also has no quoted uncertainty, and the final catalog is not shipped with the arXiv paper—that limits immediate use.\n\nOverall, this is a useful census paper for anyone working on embedded cluster populations and the PHANGS datasets. The central census and the boost factor are likely correct in direction, but the timescale claim should be softened or backed by a clock that includes embedded objects, like Paschen-alpha or spectroscopy. I would send it to a good referee with a request for revision, not desk reject. A serious referee can push on the timescale and the missing catalog.","headline":"Solid multi-galaxy census of embedded cluster candidates, but the headline ~3 Myr fade timescale rests on optically selected ages and should be softened.","tokens_in":39935,"tokens_out":1690,"would_cite":true,"duration_ms":17962,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Compact 3.3 µm PAH sources in 19 nearby galaxies are dust-embedded young star clusters, and their PAH-bright phase fades within about 3 Myr, so the optically visible cluster census is missing a substantial embedded population.","keywords":["star formation","star clusters","dust-embedded clusters","3.3 micron PAH emission","JWST NIRCam","nearby galaxies","cluster formation timescales","HST cluster catalogs"],"falsifier":"Measure independent, extinction-robust ages for a sample of the 3.3 µm PAH emitters, for example from resolved stellar populations in the nearest galaxies or from radio free-free emission probing embedded ionizing stars; if a substantial fraction turn out older than about 5 Myr, the central claim that PAH emission marks a ≤3 Myr embedded phase would be refuted.","tokens_in":38589,"feed_emoji":"🔭","tokens_out":9572,"duration_ms":80411,"temperature":0.7,"pith_summary":"This paper argues that compact sources glowing at 3.3 µm, the signal of polycyclic aromatic hydrocarbons heated by very young massive stars, are star clusters still wrapped in their natal dust and gas, too obscured for HST's optical bands. Using NIRCam photometry in 19 nearby galaxies, it selects 1,816 such sources by an F300M–F335M color excess above 5σ. When these sources are placed against the ages of optically selected clusters from the same galaxies, PAH emission is common only among clusters younger than about 3 Myr, and the PAH emitters have Hα equivalent widths 1–2.8 times higher than the youngest optical clusters. The bright end of the luminosity function therefore implies that including these dusty clusters would raise the number of known ≤3 Myr clusters by a factor of roughly 1.8–8.5 per galaxy.","feed_headline":"JWST finds 1,816 hidden young star clusters","feed_subtitle":"Their 3.3-micron PAH glow in 19 galaxies is invisible to Hubble and fades within about 3 million years.","key_machinery":"The central object is the compact 3.3 µm PAH emitter, selected as a 5σ excess of F335M over the F300M continuum in NIRCam color-magnitude diagrams. The 3.3 µm feature is a C–H stretching mode of small PAH molecules, excited by single UV photons, so its presence marks photodissociation regions and the hardest, most intense radiation fields around very young massive stars; the F300M–F335M excess therefore acts as a short-lived clock for the embedded phase. Two auxiliary diagnostics carry the argument: the F200W concentration index, an aperture-difference measure used to separate extended cluster-like objects from point-like stars, and the Hα equivalent width, which is higher for younger populations and does not rely on UV-optical SED ages.","core_discovery":"The central claim is that the compact 3.3 µm PAH emitter is a distinct, earlier phase of cluster evolution: a dusty, optically invisible or faint object that becomes a normal young HST cluster after it clears its surroundings. The paper establishes a selection recipe with a median 5σ color threshold of F300M−F335M = 0.67 mag at F335M = 20, removes likely evolved stars with a F2100W/F1000W > 3 cut, and keeps 1,816 sources. About 87% of these have F200W concentration indices in the cluster regime, and they sit preferentially in dust lanes, spiral arms, rings, bar ends, and galaxy centers. On the HST cluster age scale, roughly 30% of clusters younger than 3 Myr show the PAH color excess, while essentially none of the 4–10 Myr clusters do; this leads the authors to conclude that PAH emission at cluster scale fades within about 3 Myr. The Hα equivalent widths and the bright-end luminosity function comparison support the idea that the PAH emitters are younger than the youngest optically detected clusters, increasing the census of ≤3 Myr clusters by factors of 1.8–8.5 (median 3.3) when restricted to bright, complete samples.","pith_inferences":["Inference: if the ~3 Myr fading timescale is correct, optical cluster catalogs underestimate the current cluster formation rate by the same factor, and the correction may be largest in high-star-formation, gas-rich galaxies such as NGC 1385 and NGC 1365.","Inference: the steeper bright-end luminosity function of PAH emitters compared with HST clusters hints that the embedded population is weighted toward lower-mass clusters; a completeness-corrected mass function would test whether the embedded-to-visible ratio depends on mass.","Inference: cross-matching PAH emitters with ALMA molecular gas maps would test the paper's implied emergence sequence, predicting that PAH emitters sit near CO intensity peaks while HST clusters lie toward the cloud edges.","Inference: follow-up mid-infrared spectroscopy could decide whether the apparent 10 µm dip in the median SED is silicate absorption or simply continuum between the 7.7 and 11.3 µm PAH features, which would refine the selection diagnostics."],"forward_implications":["The dust-embedded phase of cluster formation, as traced by compact 3.3 µm PAH emission, lasts no more than about 3 Myr.","At the bright, complete end of the luminosity function, PAH-selected embedded clusters increase the count of ≤3 Myr clusters by a factor of 1.8–8.5 per galaxy, with a median of 3.3.","Around 87% of the 1,816 PAH emitters have concentration indices consistent with star clusters rather than individual stars.","PAH emitters have Hα equivalent widths 1–2.8 times higher than the youngest optically visible clusters, placing them in an earlier evolutionary phase.","The number of compact PAH emitters tracks galaxy star formation rate and CO luminosity, tying the hidden embedded population to the gas supply of the galaxy."],"supporting_citations":[{"why":"Establishes the F300M–F335M color-excess method on one galaxy, NGC 7496, that this paper extends to 19 galaxies.","marker":"Rodríguez et al. 2023"},{"why":"Provides the JWST Cycle 1 Treasury imaging, filter set, exposure times, and galaxy distances for the sample.","marker":"Lee et al. 2023"},{"why":"Supplies the HST optical cluster catalogs and morphological classifications used for the comparison and concentration-index reference.","marker":"Maschmann et al. 2024"},{"why":"Supplies SED-fitting ages for HST clusters, the clock used to infer the ~3 Myr PAH fading timescale.","marker":"Thilker et al. 2024"},{"why":"Provides the Hα narrowband imaging and the Hα+[NII] continuum-subtraction method used for equivalent widths.","marker":"Chandar et al. 2024"},{"why":"Defines the F200W concentration index used to separate cluster-like from point-like sources.","marker":"Whitmore et al. 2023a"},{"why":"Gives the cluster aperture strategy and aperture corrections adopted for HST and NIRCam photometry.","marker":"Deger et al. 2022"},{"why":"SExtractor with deblending is used to recover sources in crowded regions missed by the peak-finding algorithm.","marker":"Bertin & Arnouts 1996"},{"why":"Supplies the LMC infrared point-source comparison used to set the F2100W/F1000W cut that removes evolved stars.","marker":"Jones et al. 2017"}],"fun_headline_variants":["JWST reveals 1,816 hidden newborn star clusters","JWST's infrared eye spots 1,816 dust-enshrouded clusters","1,816 hidden clusters: JWST sees star birth in 19 galaxies","PAH glow uncovers 1,816 young clusters hidden from Hubble","JWST finds 1,816 young clusters, gone in 3 million years"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument that the PAH-bright phase lasts only about 3 Myr rests on trusting the HST-derived SED ages to separate clusters younger than 3 Myr from 4–10 Myr clusters, even though the paper notes that UV-optical colors of very young clusters are highly degenerate and that optically selected clusters are biased toward objects that have already cleared their dust.","fun_headline_variants_meta":{"raw":{"variants":["JWST reveals 1,816 hidden newborn star clusters","JWST's infrared eye spots 1,816 dust-enshrouded clusters","1,816 hidden clusters: JWST sees star birth in 19 galaxies","PAH glow uncovers 1,816 young clusters hidden from Hubble","JWST finds 1,816 young clusters, gone in 3 million years"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001725,"raw_usage":{"total_tokens":6939,"prompt_tokens":1183,"completion_tokens":5756,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":799,"completion_tokens_details":{"reasoning_tokens":5657}},"tokens_in":799,"tokens_out":5756,"duration_ms":34281,"temperature":1.0,"reasoning_tokens":5657,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:30:22.289622+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure independent, extinction-robust ages for a sample of the 3.3 µm PAH emitters, for example from resolved stellar populations in the nearest galaxies or from radio free-free emission probing embedded ionizing stars; if a substantial fraction turn out older than about 5 Myr, the central claim that PAH emission marks a ≤3 Myr embedded phase would be refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies SED-fitting ages for HST clusters, the clock used to infer the ~3 Myr PAH fading timescale."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Hα narrowband imaging and the Hα+[NII] continuum-subtraction method used for equivalent widths."}],"review_version":1}