{"id":"b449d02e-7796-4715-934d-a3d99dee70d1","arxiv_id":"2608.00918","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Star formation rate surface density in ~100 pc HII regions scales as roughly the 1.85th power of molecular gas surface density, and this steep relation extends to 500 pc scales and to high-redshift starbursts.","lead":"Astronomers combined HST, JWST, and ALMA images to measure star formation rates in 353 bright HII regions in three nearby galaxies and found that star formation rises steeply with molecular gas, a log-log slope of about 1.85. The result suggests local star-forming regions, distant starbursts, and high-redshift clumps follow one continuous star formation law, a claim that could affect how galaxies are modeled across cosmic time.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The steep slope is contingent on the annulus-mode diffuse subtraction: without it the slope is 1.35, and no fully independent subtraction has been tested on the same data.","rationale":"The reader identified the same load-bearing concern: the steep slope depends on the diffuse-subtraction method, and the no-subtraction comparison (n=1.35) shows how much of the effect comes from that assumption. I agree that this is the most fragile premise. However, the paper provides several internal checks that mitigate the concern: three fitting algorithms agree, the 500 pc re-analysis reproduces the slope, expanding the annulus has small effects, and the steep residuals do not show obvious systematics. The comparison with earlier studies using different diffuse-subtraction approaches (HIIPhot, unsharp masking, Nebulosity Filter) adds external support. Therefore the CONDITIONAL verdict remains appropriate. The concrete test I propose would directly validate the annulus-mode method on the same data and would settle whether the steepening is physical or a subtraction artifact. If that test came out negative, the verdict would need to move toward REJECT; if positive, the central claim would be substantially strengthened. Since the test has not yet been run, keeping the paper CONDITIONAL is the right call.","tokens_in":57477,"tokens_out":5273,"duration_ms":53534,"concrete_test":"On the same 353 apertures, replace the annulus-mode subtraction with an independent diffuse map: fit a smooth 2D galaxy model (e.g., polynomial or spline) to the 21µm and Hα maps after masking the selected HII regions, subtract that model from the apertures, and recompute the LINMIX fit. If the slope is 1.85±0.15, the annulus method is not the cause; if it shifts toward 1.35, the headline steepening is a subtraction artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5.2 shows that not removing the diffuse emission yields a slope of n=1.35±0.07; removing it produces n=1.85±0.12. Thus the headline steepening is entirely attributed to the decomposition of each aperture's flux into a compact HII region component and a diffuse galaxy component. The decomposition uses the mode of pixel values in a 60–100 pc annulus after sigma-clipping (Section 3.1). The paper's robustness test only increases the outer annulus radius by a factor of two, changing SFRs by ~0.03 dex; this tests the annulus size, not the validity of attributing all annulus light to the underlying galaxy. The physical argument that the diffuse 21µm correlates with stellar mass rather than current SFR (Figure 3) is model-dependent: the F300M stellar-mass map is a smooth galaxy tracer, and any smooth component—including light from young populations outside the central peak—will correlate with it. If the annulus mode removes 21µm (or Hα) from young stars in the extended star-forming complex, faint regions lose a larger fraction of their flux, which moves them downward in Σ_SFR more than bright regions, steepening the fitted slope. Because the 500 pc re-analysis uses the same subtraction procedure, it does not independently break this degeneracy. The consistency with earlier studies (Liu+11, Momose+13, Kumari+20) is reassuring but those use different galaxies/resolutions and also rely on diffuse-subtraction assumptions.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a new measurement of the resolved molecular star formation law at ~100 pc scales in three nearby galaxies (NGC 628, NGC 5194, NGC 5236), combining HST Hα, JWST MIRI 21 µm and NIRCam imaging, and CO maps. After subtracting a local diffuse background via an annulus-mode estimator, the authors find a logarithmic slope n = 1.85 ± 0.12 between Σ_SFR and Σ_mol for 353 HII regions, consistent across three regression algorithms, and a similar slope (n = 1.88 ± 0.20) at 500 pc scales. They argue that the diffuse 21 µm emission correlates with stellar mass rather than current star formation, and that the steep slope connects local HII regions to starbursts and high-redshift star-forming clumps in a single sequence.","tokens_in":57747,"tokens_out":6093,"duration_ms":57562,"significance":"If the measurement is robust, this paper provides a resolved molecular star formation law at ~100 pc scales that is significantly steeper than the canonical kpc-scale law and closer to Milky Way cloud values, with implications for star formation efficiencies and for the interpretation of high-redshift observations. The authors have been careful in several respects: three independent fitting algorithms agree, the 500 pc re-analysis reproduces the slope, and alternative α_CO prescriptions are tested in Appendix G, showing that the steep slope is not an artifact of the adopted CO-to-H2 conversion. The main caveat is the sensitivity of the headline slope to the diffuse-emission subtraction, which is the central point I raise below.","major_comments":[{"comment":"The headline slope n = 1.85 is conditional on the diffuse-emission subtraction: Section 5.2 reports that without subtracting the diffuse component, the LINMIX fit yields n = 1.35 ± 0.07, so the steepening is entirely attributed to the subtraction. The robustness test in Section 3.1 only increases the outer annulus radius by a factor of two, which changes SFRs by ~0.03 dex; this tests the annulus size but does not test whether the annulus mode correctly measures only the underlying galaxy's diffuse emission, as opposed to extended emission from the same star-forming complex. The physical argument in Section 3.2 and Figure 3, that diffuse L(21) correlates with stellar mass, is plausible but model-dependent; any smooth component tracing the disk could correlate with the stellar mass map, including light from young populations outside the central peak. If the annulus subtraction removes 21 µm or Hα from young stars in the region's periphery, faint regions lose a larger fraction of their flux, which would artificially steepen the fitted slope. The 500 pc re-analysis in Appendix F uses the same subtraction procedure, so it does not independently break this degeneracy. I recommend that the authors either (a) add an independent diffuse-subtraction test, such as using a different background estimator or a stellar-population-based model for the diffuse component, or (b) explicitly frame the steep slope as conditional on the subtraction and provide a quantitative estimate of the systematic uncertainty in n from plausible variations in the subtraction approach.","section":"5.2, 3.1"}],"minor_comments":[{"comment":"The title on the first page reads 'The HII Regions' Molecular Law of Star F ormation' with an erroneous space in 'F ormation'.","section":"Title"},{"comment":"In the acknowledgments, 'The dara were obtained' should read 'The data were obtained'.","section":"Acknowledgments"},{"comment":"The right panel of Figure 3 would benefit from an axis label including units for the diffuse 21 µm surface density; currently only the stellar mass axis has explicit units in the caption.","section":"Figure 3"},{"comment":"The reported scatter differs substantially among the three algorithms (0.18 for LINMIX versus 0.30 and 0.27 for the others); a brief explanation of this difference would help readers interpret the reliability of the scatter estimates.","section":"Table 3"},{"comment":"In the sentence 'The solid lines in the left panel of Figure 3 show the possible range of ratios permitted by a range of star formation histories using the models in Calzetti et al. (2025, briefly described in Appendix B)', the phrasing could be clarified to indicate that these lines are model predictions rather than fits to the data.","section":"3.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well executed and the result is potentially important, but the central slope is heavily dependent on the diffuse-subtraction assumption, and the current robustness test is insufficient to establish that the subtraction is correct. An independent subtraction test would substantially strengthen the paper; at a minimum, the conditional nature of the headline slope should be stated prominently and the systematic uncertainty quantified. I also note that the paper relies heavily on the authors' own previous calibration work (Calzetti et al. 2025); this is acceptable given the continuity of the method, but the referee should be alert to any circularity concerns in the SFR calibration."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should know two things about arXiv:2608.00918. First, it is a careful, well-documented measurement of the molecular star formation law at ~100 pc scales in three nearby galaxies, using JWST 21um plus HST Halpha plus ALMA/PdBI CO. The central result is a steep slope n≈1.85 after subtracting the galaxy's diffuse emission. Second, that slope is contingent on the subtraction method: without removing the diffuse component, the same data give n=1.35±0.07. The paper's own Section 5.2 shows this, and the robustness test only changes the annulus outer radius, not the assumption that all annulus light belongs to the galaxy rather than to the regions themselves.\n\nWhat is actually new: the resolution and homogeneous sample. Steep slopes after diffuse subtraction have been reported before at 250-500 pc by Liu+11, Momose+13, Kumari+20, and others. This paper adds ~100 pc resolution, three well-observed galaxies, and an explicit claim that local HII regions, high-redshift clumps, and starbursts lie on one sequence. The measurement is executed honestly: three independent fitting algorithms agree, a 500 pc re-analysis reproduces the slope, and Appendix G explores alternative alpha_CO prescriptions, showing the steep slope is not a simple artifact of the CO conversion. The epsilon_ff covariance analysis in Appendix E is also thoughtful.\n\nThe soft spot is exactly where you would expect. The annulus-mode subtraction removes the mode of pixel values in a 60-100 pc annulus after sigma-clipping. If some of that light comes from young stars in the extended star-forming complex—rather than from an old stellar population—faint regions lose a larger fraction of their flux, which moves them down in Sigma_SFR more than bright regions, and artificially steepens the fitted slope. The paper's argument that diffuse 21um correlates with stellar mass is plausible, but any smooth component would correlate with stellar mass, so it does not fully distinguish between the two interpretations. The 500 pc re-analysis uses the same subtraction, so it does not independently validate the decomposition.\n\nThe high-redshift \"single sequence\" claim also needs a hedge. The high-z points are sparse, heterogeneous, and sometimes lower limits, and the paper itself calls the comparison \"brief and incomplete.\" The abstract states the sequence more firmly than Section 6.3 supports.\n\nWho is this for? Anyone working on star formation laws, resolved SFR tracers, or CO-to-H2 conversion. It deserves a serious referee: the measurement is important, the internal tests are decent, and the weaknesses are clear enough that a referee can push for a quantitative hedge on the high-z claim and a genuinely independent test of the diffuse subtraction—for example, applying a median-filter or model-based background to the same data. I would send it to review and would probably cite it for the data and method, though I would not treat n=1.85 as established until the subtraction is independently verified.\n\nRegards,","headline":"A careful 100 pc resolved measurement of the molecular star formation law, but the steep n≈1.85 slope rests on a diffuse-subtraction assumption that the paper does not independently validate.","tokens_in":58501,"tokens_out":4729,"would_cite":true,"duration_ms":37813,"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":"The star formation law steepens to slope 1.85 at HII-region scales.","keywords":["star formation law","HII regions","molecular gas","JWST","infrared dust emission","diffuse background subtraction","starburst galaxies","high-redshift star formation"],"falsifier":"A decisive check would be to measure SFR in the same 353 regions with a tracer that is insensitive to dust heated by old stars—for example, radio free-free emission or a recombination line ratio with a much deeper extinction correction—and to refit $\\Sigma_{\\rm SFR}$ versus $\\Sigma_{\\rm mol}$; if the slope returned to near $n \\simeq 1$ instead of $1.85$, the diffuse-light subtraction would be implicated. A second, complementary test would be to run the paper's exact annular-mode background recipe on simulated galaxy images with a known input star formation law and see whether the recovered slope is biased.","tokens_in":2160,"feed_emoji":"🌠","tokens_out":2657,"duration_ms":96502,"temperature":0.7,"pith_summary":"The paper analyzes 353 HII regions of roughly 100 pc size in three nearby galaxies, NGC 628, NGC 5194, and NGC 5236, using HST Hα, JWST 21 µm, and CO maps. It claims that the molecular star formation law at these scales is steep, $\\Sigma_{\\rm SFR} \\propto \\Sigma_{\\rm mol}^{1.85 \\pm 0.12}$, much steeper than the near-linear relation found for kiloparsec-scale galaxy regions and close to the slopes of Milky Way molecular clouds. The steepening comes from subtracting the galaxy's diffuse background light from each region; without that subtraction the slope falls to $1.35 \\pm 0.07$. The paper further claims that local HII regions, high-redshift star-forming clumps, and low- and high-redshift starburst galaxies fall on one continuous sequence of star formation across three orders of magnitude in gas surface density. A sympathetic reader would care because a single steep law from 100 pc HII regions to distant starbursts would make HII regions true scaled-down versions of extreme star-forming systems, and would mean that unresolved galaxy measurements that do not remove diffuse light underestimate the steepness of the real star formation law.","feed_headline":"Star formation steepens to slope 1.85 at HII region scales","feed_subtitle":"Removing diffuse galaxy light reveals one molecular star-formation law from 100-pc HII regions to redshift-5 starbursts.","key_machinery":"The central objects are the surface-density quantities $\\Sigma_{\\rm SFR}$, from dust-corrected Hα combined with JWST 21 µm emission, and $\\Sigma_{\\rm mol}$, from CO total-intensity maps converted with a disk-weight-dependent $\\alpha_{\\rm CO}$. The mechanism that carries the argument is annular background subtraction: for each 60 pc aperture, the diffuse galaxy emission is estimated as the mode of pixel values in a surrounding 60–100 pc annulus after $\\sigma$-clipping, and this diffuse component is removed from both the SFR and molecular gas tracers. Removing that component changes the fitted slope from $n \\simeq 1.35$ to $n \\simeq 1.85$, and the paper argues the diffuse 21 µm light is heat from old stars, not current star formation, because it tracks stellar mass rather than region luminosity. Supporting machinery includes a fixed 60 pc free-fall radius used to derive efficiencies per free-fall time, and a disk-weight-dependent CO-to-H$_2$ conversion that removes galaxy-to-galaxy offsets without steepening the within-galaxy relations.","core_discovery":"On the paper's own terms, the central discovery is a resolved molecular star formation law for HII regions: for the 353 regions above the censoring limits, $\\log \\Sigma_{\\rm SFR} = (1.85 \\pm 0.12)\\,\\log \\Sigma_{\\rm mol} - (4.27 \\pm 0.26)$, with a scatter of 0.18 dex in the LINMIX fit. This is significantly steeper than the $n \\sim 0.9$–$1.3$ slopes typical of kiloparsec-scale galaxy regions and is close to the $n \\sim 1.6$–$2$ slopes of Milky Way molecular clouds. The relation remains steep at 500 pc scales, with slope $1.88 \\pm 0.20$, and the paper interprets the steepness as the signature of current star-forming regions isolated from the diffuse emission of the underlying galaxy. It also finds that the diffuse 21 µm emission correlates with stellar mass surface density rather than with current star formation, and argues that leaving that diffuse component in the photometry flattens the measured slope to $1.35 \\pm 0.07$. The paper extends the relation to lower and higher gas surface densities by adding local starbursts and high-redshift clumps, concluding that all these systems form a single star formation sequence over three orders of magnitude in $\\Sigma_{\\rm mol}$.","pith_inferences":["A natural extension the authors do not pursue: applying the same annular diffuse-light subtraction to existing kiloparsec-resolution surveys would probably steepen their measured slopes toward $n \\sim 1.5$–$1.8$, changing published molecular depletion times.","If the steep law is universal, then unresolved high-redshift galaxy stacks that mix diffuse and compact emission may be fitting a flatter effective slope; resolved ALMA-scale observations of individual clumps at $z \\sim 2$–$4$ would provide a direct test.","The slope could be tracer-dependent: using a high-density gas tracer such as HCN instead of CO might yield a shallower relation, since the paper's $\\Sigma_{\\rm mol}$ includes lower-density gas that is not directly forming stars.","A testable consequence for simulations: the same 60–100 pc annular background recipe applied to mock galaxy images should recover the input star formation law; if it artificially steepens it, the observational slope is partly a measurement effect."],"forward_implications":["If the slope is truly $n \\simeq 1.85$, the molecular gas depletion time decreases with activity, $\\tau_{\\rm dep} \\propto \\Sigma_{\\rm SFR}^{-0.5}$ or steeper, meaning that more intensely star-forming regions consume their gas faster.","The efficiency per free-fall time stays low on average, about 1%, but increases by roughly a factor of 3.5 from faint to bright regions, so star formation feedback does not need to be finely tuned to hold efficiencies down.","Kiloparsec-scale laws with $n \\sim 1$ appear to be partly a mixing artifact: if diffuse galaxy light is not removed, the resolved law flattens toward the old galaxy-wide value, so the true small-scale law may be steep everywhere.","Local HII regions, high-redshift star-forming clumps, and starbursts join into one sequence over three orders of magnitude in $\\Sigma_{\\rm mol}$, giving a common calibration for interpreting unresolved high-redshift measurements.","Physical star formation models must reproduce both a steep slope and a large scatter at fixed region size; models adding a power-law tail to the gas density probability distribution are the ones the paper identifies as capable of bracketing the data."],"supporting_citations":[{"why":"Supplies the Hα+21 µm SFR calibration and the star-formation-history-dependent scaling used for all region SFRs.","marker":"Calzetti et al. (2025)"},{"why":"Provides the source selection and photometry approach for NGC 628 that the paper extends to the other two galaxies.","marker":"Calzetti et al. (2024)"},{"why":"Provides the LINMIX Bayesian regression algorithm used for the headline slope fit.","marker":"Kelly (2007)"},{"why":"Supplies the disk-weight-dependent CO-to-H2 conversion factor whose choice the paper tests in an appendix.","marker":"Bolatto et al. (2013)"},{"why":"Supplies the dust model used to identify 21 µm emission from old stellar populations as diffuse contamination.","marker":"Draine & Li (2007)"},{"why":"Provides the ALMA CO maps and $R_{21/10}=0.65$ conversion for NGC 628 and NGC 5236.","marker":"Leroy et al. (2021a)"},{"why":"Provides the PdBI/PAWS CO map of NGC 5194 used to derive molecular gas surface densities.","marker":"Schinnerer et al. (2013)"},{"why":"Provides the local starburst sample that anchors the high-$\\Sigma_{\\rm mol}$ end of the unified sequence.","marker":"Kennicutt & De Los Reyes (2021)"},{"why":"Provides 400–500 pc starburst-region data used to place local U/LIRGs on the same relation.","marker":"Wilson et al. (2019)"},{"why":"Provides the redshift-5 galaxy data point used in the high-redshift comparison.","marker":"Accard et al. (2025)"}],"fun_headline_variants":["HII regions steepen molecular star formation law to slope 1.85","Removing diffuse light reveals steep HII region star formation law","One steep star formation law from 100-pc HII regions to starbursts","100-pc HII regions expose steeper molecular law of star formation","Steep molecular star formation law holds for HII regions and starbursts"],"cache_read_input_tokens":60416,"weakest_assumption_plain":"The load-bearing premise is that the diffuse 21 µm and Hα light inside each 60 pc aperture belongs to the galaxy's old stellar population, not to the region's current star formation, and that this diffuse component can be measured as the mode of pixel values in the surrounding 60–100 pc annulus; the paper's own no-subtraction fit gives a shallower slope of $1.35 \\pm 0.07$, so if this decomposition is wrong the headline steepening weakens.","fun_headline_variants_meta":{"raw":{"variants":["HII regions steepen molecular star formation law to slope 1.85","Removing diffuse light reveals steep HII region star formation law","One steep star formation law from 100-pc HII regions to starbursts","100-pc HII regions expose steeper molecular law of star formation","Steep molecular star formation law holds for HII regions and starbursts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001526,"raw_usage":{"total_tokens":6193,"prompt_tokens":1111,"completion_tokens":5082,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":727,"completion_tokens_details":{"reasoning_tokens":4981}},"tokens_in":727,"tokens_out":5082,"duration_ms":31419,"temperature":1.0,"reasoning_tokens":4981,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:15:12.882834+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be to measure SFR in the same 353 regions with a tracer that is insensitive to dust heated by old stars—for example, radio free-free emission or a recombination line ratio with a much deeper extinction correction—and to refit $\\Sigma_{\\rm SFR}$ versus $\\Sigma_{\\rm mol}$; if the slope returned to near $n \\simeq 1$ instead of $1.85$, the diffuse-light subtraction would be implicated. A second, complementary test would be to run the paper's exact annular-mode background recipe on simulated galaxy images with a known input star formation law and see whether the recovered slope is biased.","supporting_citations":[],"review_version":1}