{"id":"6f277f10-72fc-4d63-92e1-def9140ff167","arxiv_id":"2501.15789","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The 105 brightest protoplanetary disks in Ophiuchus have a log-normal size distribution with median about 14 au; close binaries are about 3 times smaller, and embedded disks have the same sizes as Class II disks.","lead":"Astronomers used ALMA to measure the sizes of the 105 brightest planet-forming disks in the Ophiuchus star-forming cloud, finding a typical radius of about 14 au. The results give the largest complete size catalog of such disks, and show that disks in close binary systems are much smaller and that young embedded disks are not smaller than older ones.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Two-resolution design not validated: the combined log-normal size distribution may combine differently biased HWHM measurements from the 0.15\" and 0.05\" programs.","rationale":"The paper is careful and provides multiple cross-checks, which strengthens confidence in the overall methodology. However, the central claim of a complete, log-normal size distribution depends on the measured HWHM values being unbiased across the whole sample. The most direct threat to that claim is the lack of validation for the two-resolution observing design. The reader's weakest assumption mentions this alongside binaries; I focus on the resolution issue because it affects the entire sample, including the headline numbers, rather than only the binary subsample. The proposed test is feasible with existing data: binary components observed at both resolutions provide a natural control sample, and convolution of high-resolution images to low resolution can quantify any systematic effect. If the test passes, the central claim is much stronger; if it fails, the size distribution needs re-analysis. The verdict should remain CONDITIONAL until this check is performed, consistent with the reader's assessment.","tokens_in":13111,"tokens_out":7998,"duration_ms":76391,"concrete_test":"Compare the imfit HWHM values of the binary components that were observed in both the Cycle 8 (0.15\") and Cycle 9 (0.05\") programs (Section 2.1). If these sizes differ systematically by more than ~10% or beyond the quoted uncertainties, the two-resolution design is biasing the combined size distribution. If too few such components exist, convolve the 0.05\" images of the faint sample to the 0.15\" beam and re-fit; a systematic shift in recovered HWHM would indicate a resolution-dependent bias.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.1 splits the sample by Band 6 flux: 45 objects brighter than 20 mJy observed at 0.15\" (21 au) and 55 objects with 4-20 mJy observed at 0.05\" (7 au). Since disk size correlates with flux (Fig 1, bottom-left), the resolution split is effectively a size split. The paper's internal cross-checks—image-plane vs visibility fits agree within 3%, and Frank R68% vs HWHM has a slope of 0.78—are performed across the combined dataset and do not verify that the two resolutions yield consistent sizes for the same physical source. Non-binary disks are each observed at only one resolution; the only disks observed at both resolutions are some binary components, and no comparison of their measured sizes across the two epochs is reported. If the deconvolved HWHM depends on beam size—for example, if the 0.05\" observations resolve out extended emission or the 0.15\" observations blend the disk with background—the bright and faint subsamples could carry opposite systematic biases. Combining them could create a spurious log-normal shape or shift its median and sigma. The headline median of ~14 au and sigma_log = 0.46 therefore rest on an untested homogeneity assumption.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"Dasgupta et al. present ALMA Band 8 (410 GHz) continuum observations of the 105 brightest ODISEA disks in Ophiuchus and measure their sizes using image-plane Gaussian fits (HWHM), with visibility-plane and Frank R68% cross-checks. The paper reports that the HWHM distribution is log-normal with a median of ~14 au and logarithmic standard deviation 0.46, that close binaries (projected separation < 200 au) have a median size of ~5 au, and that Class I/Flat Spectrum and Class II disks have statistically indistinguishable size distributions. The authors interpret the binary result as evidence for efficient radial drift and/or tidal truncation and the lack of SED-class evolution as evidence that pressure bumps are common at early disk stages.","tokens_in":13380,"tokens_out":8677,"duration_ms":79421,"significance":"If the size measurements are unbiased, this is a valuable demographic benchmark: the largest flux-limited, fully resolved sample of protoplanetary disk sizes in a single star-forming region, with a clean SED-class split and a quantitative binary comparison. The paper's measurement strategy has real strengths: sizes come from direct interferometric observables, image-plane and visibility-plane fits agree within 3%, the smallest source is explicitly checked against a point-source model, the Frank R68%-HWHM relation is quantified (slope 0.78), and the binary flux bias is addressed with a restricted sample. The main claims are falsifiable and should be reproducible from the data tables. The principal risk is that the two-resolution observational design and the image-plane-only binary fits are not independently validated, so the headline distribution and binary median need additional support.","major_comments":[{"comment":"The two-resolution design is confounded with source flux and size. The 45 objects brighter than 20 mJy in Band 6 were observed at 0.15\" (21 au) and the 55 objects with 4-20 mJy at 0.05\" (7 au), while the paper itself shows that disk size correlates with flux (Figure 1, bottom-left). No common-source cross-check between the two programs is reported: the image-plane versus visibility-plane comparison in Figure 3 is performed within each dataset, and the statement that some binary systems were observed at both resolutions refers to different components falling in different flux ranges, not to the same disk being measured at both resolutions. If the deconvolved HWHM depends on beam size, for example through resolved-out extended emission at 0.05\" or blending at 0.15\", the bright and faint subsamples could carry opposite systematic biases, and the combined log-normal fit in Table 2 (median 14.37 au, sigma 0.46) would not be a single unbiased distribution. Please provide an explicit validation of resolution homogeneity, for example source-injection or simulated observations of representative disk models at both resolutions, or at least split the sample by resolution and show that the fitted distributions are consistent.","section":"Section 2.1 and Figure 1"},{"comment":"All close-binary sizes rely solely on image-plane imfit fits; the paper notes that visibility fitting is challenging for binaries but does not provide any dedicated validation, such as source injection or alternative fitting, for blended close binaries. This matters directly for the second central claim: the close-binary median of 4.6 au and the K-S comparison in Table 3 (88 vs 17, p = 1e-6) depend on these fits. The small observed binary radii could in principle be an artifact of beam deconvolution or component blending, particularly for faint secondaries observed at high resolution. Please add an injection-recovery test with the actual UV coverage, or show for at least a few close binaries that image-plane and visibility-plane fits are consistent.","section":"Section 2.1.1"},{"comment":"The log-normal parameters are obtained by fitting histograms with scipy.optimize.curve_fit, but the manuscript does not specify the binning, the fitting domain, or whether the fit is to binned counts rather than to the unbinned sizes. Since the median and sigma in Table 2 are the headline results, a binned fit with arbitrary bin choice can bias both parameters. Please report an unbinned maximum-likelihood fit (or at least demonstrate that the quoted values are insensitive to binning) and state the exact fitting procedure.","section":"Section 3.1, Table 2"}],"minor_comments":[{"comment":"The text says the median HWHM is 13 au, while Table 2 reports 14.37 +/- 1.38 au and the abstract says ~14 au; please reconcile these values.","section":"Section 3.1"},{"comment":"The first K-S row compares the whole sample (N=105) with the subsample excluding close binaries (N=88), but these samples are not independent because the second is contained in the first. The p-value of 0.58 therefore does not provide the stated support for similarity; the independent comparisons (88 vs 17 and 88 vs 13) are the informative ones.","section":"Table 3"},{"comment":"The right panel reports a best-fit slope of 0.78 for R68% versus HWHM but gives no intercept, scatter, or goodness-of-fit; if this relation is intended as a conversion between size metrics, the full linear relation should be stated.","section":"Figure 3"},{"comment":"There are several typographical issues, including 'Tale 2' in the Figure 1 caption, 'the the median' in Section 3.2, and 'Analogos' in Section 4.1; a careful proofread is needed.","section":"Throughout"},{"comment":"The paper should state explicitly whether the full version of Table 1, with all 105 sources, is available in machine-readable form, since the abbreviated printed table is insufficient to reproduce the size distribution.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's scope and the data set is valuable. The main issue is not the internal consistency of the measurements but an untested observational design assumption: combining two resolutions that are correlated with flux and size without a common-source or injection-based check. This is fixable with additional analysis, so I am not recommending rejection, but the headline log-normal and binary results should not be published without that validation. I would also like the statistical fitting details to be tightened before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper before the next survey round: it delivers the largest flux-limited resolved disk sample in any star-forming region, with ALMA Band 8 sizes for 105 Ophiuchus disks. The headline result is a log-normal size distribution with median ~14 au and sigma_log 0.46, close binaries much smaller (~5 au), and no significant size difference between Class I/Flat and Class II. The data are new and the measurements are carefully done: image-plane and visibility-plane fits agree within 3%, the smallest source is explicitly shown to be resolved against a point-source model, and the Frank R68% versus HWHM correlation (slope 0.78) holds across the full size range. That is real, reproducible progress and a benchmark for disk evolution models.\n\nThe soft spots are real but not fatal. First, all binary sizes come from image-plane fits only, with no blending test or visibility-plane cross-check for close binaries. Given that the binary median size is a headline claim, that needs addressing. Second, the two-resolution design is a legitimate concern: bright disks observed at 0.15\" (21 au) and faint at 0.05\" (7 au), and since size correlates with flux, the resolution split is effectively a size split. The paper shows no direct test that the two resolutions give consistent sizes for the same physical source. The cross-checks they do show are across the combined dataset, so they do not rule out a resolution-dependent bias. The Frank correlation helps, but it too is image-based and could inherit the same resolution effects.\n\nI would not call the central result wrong on this basis, but the paper should demonstrate homogeneity explicitly—for example, by re-observing a few sources at both resolutions or by injecting model disks at the two resolutions and comparing recovered HWHM. Also, the pressure-bump conclusion in the abstract overreaches: the lack of size evolution in a flux-limited sample is suggestive, but it is not a direct detection of pressure bumps, and the sample selection should be acknowledged as a limit on that inference.\n\nBottom line: this is a solid survey paper that deserves serious peer review. It should be published after the authors tighten the binary size measurement and either validate or soften the two-resolution claim.","headline":"A valuable flux-limited disk size catalog with careful cross-checks, but the two-resolution design and image-plane-only binary sizes need scrutiny before the headline log-normal distribution is taken at face value.","tokens_in":14020,"tokens_out":1756,"would_cite":true,"duration_ms":19145,"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 paper measures the complete size distribution of the 105 brightest Ophiuchus disks and finds it log-normal with median ~14 au, close binaries at ~5 au, and no size difference between embedded and Class II sources.","keywords":["protoplanetary disks","disk size distribution","ALMA","Ophiuchus","binary disks","radial drift","SED classes","continuum emission"],"falsifier":"Compare the close-binary sizes with visibility-plane fits or forward-modeled synthetic observations of blended binaries: if the imfit half-width at half-maximum is systematically wrong for separations near or below the beam, the roughly 5 au binary median and the binary versus single comparison would not survive. A simpler check is to re-observe the 45 bright disks at 0.05 arcsec resolution and see whether their size distribution shifts.","tokens_in":12963,"feed_emoji":"🔭","tokens_out":5494,"duration_ms":46352,"temperature":0.7,"pith_summary":"This paper reports the first complete, flux-limited size census of the brightest protoplanetary disks in the Ophiuchus star-forming region, using ALMA Band 8 observations at 7-21 au resolution to resolve all 105 disks. It finds that their half-width at half-maximum sizes follow a log-normal distribution with a median of about 14 au and a logarithmic spread of 0.46, establishing a demographic baseline for disk sizes. The census shows that disks in close binaries, with projected separation under 200 au, are much smaller with a median near 5 au, consistent with efficient radial drift of dust rather than tidal truncation alone. It also finds that young embedded disks and older Class II disks have statistically indistinguishable sizes, which the authors read as evidence that pressure bumps that trap millimeter grains are already common in the first million years.","feed_headline":"Ophiuchus disk census: median size 14 au across 105 disks","feed_subtitle":"Close binaries shrink disks to about 5 au; young and old disks match in size.","key_machinery":"The central measurement object is the half-width at half-maximum (HWHM) of a two-dimensional Gaussian fitted to each disk's 410 GHz continuum image in the image plane with CASA imfit; for non-binary disks the radius enclosing 68% of the flux, $R_{68\\%}$, is computed from radial profiles made with the Frank code as a cross-check. The two methods correlate with a best-fit slope of 0.78, and image-plane and visibility-plane Gaussian fits agree within 3%, so the HWHM carries the statistical comparisons. The survey design splits the flux-limited sample, with 45 brighter disks observed at 0.15 arcsec (21 au) resolution and 55 fainter disks at 0.05 arcsec (7 au), resolving every target and allowing HWHM values down to about 0.015 arcsec.","core_discovery":"The central discovery is that the continuum sizes of the 105 brightest Ophiuchus disks follow a log-normal distribution with median HWHM $\\sim$14 au and $\\sigma_{\\log}=0.46$, spanning 1.7 to 177 au. Excluding close binaries raises the median to about 16 au and narrows the spread to $\\sigma_{\\log}=0.39$. The 17 disks in close binaries with separation under 200 au are distinctly smaller, with a median near 4.6 au, and they remain smaller at fixed millimeter flux, supporting models in which companions enhance radial drift. Embedded Class I and Flat Spectrum sources and Class II sources have indistinguishable size distributions whether or not binaries are included and whether sizes come from Gaussian fits or from $R_{68\\%}$ profiles; the authors conclude that millimeter grains must be stopped by pressure bumps from very early times.","pith_inferences":["The two-resolution design means the combined distribution could hide a resolution-dependent bias; re-observing the bright subsample at 0.05 arcsec and checking whether bright-end sizes shift would test this.","If the close-binary size deficit is driven by enhanced radial drift, gas disk sizes in the same binaries should be much larger than the dust sizes, a prediction testable with CO line observations of these binaries.","The log-normal form invites fitting the same distribution in other star-forming regions to see whether the 14 au median and $\\sigma_{\\log}=0.46$ are universal or region-specific.","Because the flux limit cuts at $M_{\\rm dust} \\gtrsim 2\\,M_\\oplus$, the full population including fainter disks likely extends to smaller sizes, and deeper observations would quantify the low-mass tail."],"forward_implications":["If the log-normal size distribution is representative, surveys that resolve only the brightest disks overestimate typical disk sizes; the median planet-forming disk is compact, near 14 au.","Close binaries produce disks more than a factor of two smaller than single stars, so compact planetary architectures rather than Uranus or Neptune analogs should be the norm around close binaries.","The size-flux relation measured in Band 8 matches the previously established $R \\propto L^{0.6}$ relation and extends it to fainter, smaller disks.","The lack of size evolution between embedded and Class II sources implies that pressure bumps or other dust-trapping substructures must be present in disks younger than about 1 Myr and at small radii.","The sample provides a benchmark for disk population synthesis models and for future comparisons with exoplanet demographics around binary systems."],"supporting_citations":[{"why":"Supplies the Band 6 flux-limited parent sample and the 4 mJy cutoff used to select the 105 disks.","marker":"Williams et al. (2019)"},{"why":"Defines the ODISEA survey whose targets and prior detections this Band 8 program extends.","marker":"Cieza et al. (2019)"},{"why":"Provides the near-IR adaptive-optics binary census down to about 6 au used to classify close binaries.","marker":"Zurlo et al. (2020)"},{"why":"Gives the Taurus binary disk sizes showing radii below 0.1 of orbital separation, the direct comparison for the Ophiuchus binaries.","marker":"Manara et al. (2019)"},{"why":"Supplies the dust evolution models predicting efficient radial drift and small continuum sizes in close binaries.","marker":"Zagaria et al. (2021b)"},{"why":"Establishes the disk size-luminosity relation and the $R_{68\\%}$ size metric that the paper compares against.","marker":"Hendler et al. (2020)"},{"why":"Introduces the radius enclosing 68% of the continuum flux as a size metric used here.","marker":"Tripathi et al. (2017)"},{"why":"Motivates pressure bumps as the mechanism preventing rapid inward drift of millimeter grains, which the SED-class result is read to support.","marker":"Pinilla et al. (2012)"}],"fun_headline_variants":["Ophiuchus disks: median 14 au, log-normal spread, binaries at 5 au","ALMA census of 105 disks: median 14 au, close binaries shrink to 5 au","Disk sizes in Ophiuchus: log-normal, median 14 au, binaries smaller","Brightest Ophiuchus disks: median 14 au, binaries ~5 au","105 Ophiuchus disks resolved: sizes log-normal, binaries halve median"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results assume that the image-plane Gaussian half-width at half-maximum measures the true dust size for every source, with no cross-check for the blended components of close binaries, and that combining 0.15-arcsec and 0.05-arcsec observations does not bias the joint distribution.","fun_headline_variants_meta":{"raw":{"variants":["Ophiuchus disks: median 14 au, log-normal spread, binaries at 5 au","ALMA census of 105 disks: median 14 au, close binaries shrink to 5 au","Disk sizes in Ophiuchus: log-normal, median 14 au, binaries smaller","Brightest Ophiuchus disks: median 14 au, binaries ~5 au","105 Ophiuchus disks resolved: sizes log-normal, binaries halve median"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000249,"raw_usage":{"total_tokens":1608,"prompt_tokens":1060,"completion_tokens":548,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":676,"completion_tokens_details":{"reasoning_tokens":431}},"tokens_in":676,"tokens_out":548,"duration_ms":5061,"temperature":1.0,"reasoning_tokens":431,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:57:11.158465+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the close-binary sizes with visibility-plane fits or forward-modeled synthetic observations of blended binaries: if the imfit half-width at half-maximum is systematically wrong for separations near or below the beam, the roughly 5 au binary median and the binary versus single comparison would not survive. A simpler check is to re-observe the 45 bright disks at 0.05 arcsec resolution and see whether their size distribution shifts.","supporting_citations":[],"review_version":1}