{"id":"4b8d9845-878e-43e3-a45a-c8c8be4bc033","arxiv_id":"2608.04803","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"A radiative transfer model of the edge-on disk d216-0939 indicates 5.4% crystalline water ice in the cold outer layers, implying outward transport of ice.","lead":"This paper builds a 3D computer model of the disk d216-0939 and finds that dust with 5.4% crystallized water ice beyond the snowline best matches JWST spectra. The result suggests crystalline ice formed near the star was carried outward, which matters for understanding how planets get water.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The outward-transport inference relies on a no-ice dust temperature field; if icy opacities warm the absorbing layers above ~130 K, in-situ crystallization cannot be excluded.","rationale":"The paper's framework is credible: it uses a 3D MCRT code, validates spectral importance sampling against wavelength-by-wavelength runs (Appendix B), and tests robustness by redoing the water-feature fit with a second continuum model (model h), obtaining a similar qualitative result. The load-bearing weak point is exactly the one identified by the reader: the temperature map used to identify cold crystalline ice is computed without the ice whose presence is being inferred. The qualitative conclusion that crystalline ice at temperatures far below the crystallization temperature requires outward transport is only established if the no-ice and ice-inclusive temperature fields agree in the absorbing layers. This is an assumption, not a demonstrated result, and no numerical check is reported in the manuscript. The specific 5.4% abundance is secondary and already weakly supported; the spatial/crystallinity claim is the more important result. Since the concern is testable by recomputing temperatures with icy opacities and by scanning the snowline/crystallization thresholds, conditional acceptance is the right level of confidence. The reader's CONDITIONAL verdict remains appropriate, so no change is needed.","tokens_in":20201,"tokens_out":10614,"duration_ms":141793,"concrete_test":"Recompute the 3D thermal equilibrium of the best-fit model with the actual icy mixture (5.4% ice, rho = 1.65 g cm^-3) placed beyond the 170 K snowline, using the MgSiO3/H2O optical constants over their valid range and the DSHARP no-ice opacities at wavelengths outside that range as a bracketing case. Compare the new temperature field with the no-ice field at the cells along the observer-to-star line of sight between the unity optical-depth surfaces at 3.1 micron (Fig. 4). If those temperatures exceed ~130 K, in-situ crystallization is not excluded and the transport claim fails; if they stay below ~100 K, the central qualitative conclusion is supported. Then re-fit the water feature using the new temperature field to see whether the 5.4% abundance and the preference for the 150 K mixture survive.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that crystalline water ice in d216-0939 requires outward transport is conditioned on the temperature field computed with the DSHARP no-ice mixture and on a fixed T_snow = 170 K (Sect. 3.4). The paper explicitly assumes that the dust temperature is not significantly influenced by the presence of water ice, but this assumption is never tested. The water-feature fit then uses icy opacities (mixtures #1-6 and the 5.4% ice variant with rho_art = 1.65 g cm^-3) while keeping that no-ice temperature field. The decisive question is the temperature of the material actually responsible for the 3 micron absorption: the paper claims it lies far below the crystallization temperature, but the temperature map that says so contains no water ice. If including ice changes the thermal balance in the upper/outer layers probed by the feature (water ice strongly affects opacity near 3 micron and the scattering albedo), those layers could be warm enough for in-situ crystallization, and the transport inference would collapse. The two-snowline test (crystalline only at 100-170 K) does not close this gap: it still uses the same no-ice temperature map and a fixed 100 K boundary rather than scanning the crystallization temperature. The 5.4% abundance is similarly tied to this model and is tuned by hand without an uncertainty estimate.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents 3D Monte Carlo radiative transfer models of the edge-on protoplanetary disk d216-0939, fitting the JWST NIRSpec/MIRI continuum SED together with archival HST H-alpha images to select a reference disk model. Water ice is then included beyond an assumed 170 K snowline using laboratory MgSiO3/H2O optical constants for amorphous and crystalline ice. The authors report that a model with 5.4% crystalline water ice by mass and an artificially lowered dust bulk density of 1.65 g/cm3 best reproduces the observed 3 micron water ice absorption feature. Because the modeled absorbing layers are colder than the crystallization temperature of water ice, they conclude that crystalline ice could not have formed in situ and must have been transported outward. The paper also decomposes the feature flux into attenuated starlight, scattered starlight, thermal dust emission, and self-scattered dust emission.","tokens_in":20512,"tokens_out":5672,"duration_ms":63604,"significance":"If established, the detection of crystalline water ice in the cold upper and outer layers of d216-0939 would provide an important observational constraint on radial transport in protoplanetary disks, with implications for planetesimal composition and for the interpretation of ice features in edge-on disks. The modeling has real strengths: the spectral importance sampling method is validated in Appendix B against POLARIS with quantified relative errors; the source-term decomposition in Fig. 6 is an illuminating diagnostic; and the robustness check with model h in Section 5.2.2 is a welcome acknowledgment of model degeneracy. However, the central astrophysical claim rests on an untested assumption about the dust temperature field and on a manually tuned ice abundance without an uncertainty estimate, so the present version does not yet establish the outward-transport conclusion.","major_comments":[{"comment":"The central inference that crystalline water ice resides 'far below the crystallization temperature' and therefore requires outward transport depends entirely on the dust temperature field shown in Fig. 4. That field is computed with the DSHARP (no ice) mixture; Section 3.4 states, but does not test, the assumption that the presence of water ice does not significantly influence the dust temperature. The icy mixtures used for the feature have substantially different absorption and scattering opacities in the 2.6-4.0 micron range (Fig. 1), so the upper and outer layers probed by the feature could plausibly be warmer in a self-consistent model. Please compute the thermal equilibrium with the best-fit icy mixture and compare temperatures along the lines of sight that contribute to the feature, or otherwise demonstrate that the no-ice temperature field remains valid. Without this, in-situ crystallization cannot be excluded and the outward-transport conclusion is not established.","section":"Sects. 3.4 and 5.2.3, Fig. 4"},{"comment":"The 5.4% ice mass fraction is presented as a well-constrained result, but it is obtained by manually setting an artificial dust density of 1.65 g/cm3 and a MgSiO3/H2O component mass fraction of 0.2 in order to match the observed feature depth. No quantitative fit statistic, confidence interval, or exploration of the degeneracy between ice fraction, artificial density, maximum grain size, and background scaling factor is provided. The statement that 'the water ice content of 5.4% is well constrained by the depth of the feature' is therefore unsupported. Please provide a quantitative best-fit search or a chi-square map over the relevant parameters and report uncertainties; alternatively, the abundance claim should be weakened to an order-of-magnitude estimate.","section":"Section 5.2.2, Fig. 6"},{"comment":"The final reference model k is selected by visual inspection of synthetic HST images, as stated in Section 5.1, and the text acknowledges that the continuum SED fit is highly degenerate. The spatial structure of the disk determines which layers lie on the line of sight and therefore directly affects the modeled water ice feature, so the visual selection criterion introduces an unquantified systematic uncertainty in the inferred ice distribution. Please add a quantitative image-comparison metric (e.g., a residual-based statistic on normalized images) and show which alternative models in Table E.1 produce significantly different ice-feature predictions. The robustness test with model h is helpful but does not remove the need for a quantitative selection criterion.","section":"Section 5.1, Table E.1"},{"comment":"The two-snowline test does not close the temperature-feedback gap described in the first major comment. It places crystalline ice only in the 100-170 K zone and amorphous ice at T<100 K using the same no-ice temperature map, with the boundary fixed at 100 K rather than scanned over the crystallization range. The conclusion that the upper and outer disk layers contain only negligible amounts of amorphous water ice is therefore conditional on that same no-ice temperature map. A meaningful test would vary the crystallization boundary (e.g., 100, 120, 140, 170 K) and, ideally, use self-consistently computed temperatures with icy opacities.","section":"Section 5.2.2"}],"minor_comments":[{"comment":"The relation is printed as α=3β−1/2; please check whether the intended expression is α=3(β−1/2), since the two differ and the latter is the standard Lin & Papaloizou form often used with the flaring exponent.","section":"Equation (3)"},{"comment":"In the flux decomposition paragraph, the second occurrence of F emi,⋆λ should presumably read F sca,⋆λ (scattered starlight), since the sentence reports the scattered contribution decreasing from 30% to 10% across the feature.","section":"Section 5.2.2"},{"comment":"The caption contains the typo 'DSHAPR (no ice)'; it should read 'DSHARP (no ice)'.","section":"Fig. D.1 caption"},{"comment":"The abstract reports '5.4% crystallized water ice' while the conclusions report '~5%'; given the lack of an uncertainty estimate, please harmonize the precision or add the uncertainty to both statements.","section":"Abstract and Section 6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well organized and the computational framework is a genuine contribution, but the headline claim about outward transport is currently conditional on an untested no-ice temperature field and on a manually tuned abundance. Both issues are fixable within the scope of the paper, so I recommend major revision rather than rejection. I saw no concerns about citation practice or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead this if you want to know where the field stands on spatially mapping ice in edge-on disks. The paper does something genuinely new: it builds a 3D MCRT model of d216-0939, fits the JWST SED and HST silhouette simultaneously, and uses spectral importance sampling to model the 3 micron water ice feature at high resolution. That combination lets them separate attenuated starlight, scattering, and dust emission, and estimate the inclination (~78 deg) and the ice mass fraction in the upper/outer layers. The validation of the importance sampling in Appendix B is careful, with relative errors below 0.5% against wavelength-by-wavelength runs. That is real technical work.\n\nThe qualitative conclusion — crystalline ice in cold upper disk layers, implying outward transport — is plausible and consistent with the feature shape. The paper also honestly reports degeneracies in the continuum fit and checks robustness with a second reference model. Good practice.\n\nNow the soft spots. The 5.4% ice abundance is not a measured number. It comes from manually lowering the bulk density to 1.65 g/cm^3 and adjusting the MgSiO3/H2O fraction until the feature depth matches. No fit metric, no error bars, no sensitivity scan. The alternative model h gives ~4%, so the true uncertainty is wider than the paper admits.\n\nThe bigger issue is the temperature field. They compute dust temperatures with ice-free DSHARP opacities and assume ice does not change the thermal balance. Then they put ice beyond the 170 K snowline and fit the feature with icy opacities. The stress-test note is right: if the icy grains in the absorbing layers are warmer than the no-ice map says, the in-situ crystallization argument weakens. The paper's two-snowline test (crystalline only between 100-170 K) still uses the same temperature map and does not scan the crystallization temperature. This is a load-bearing assumption that is asserted, not tested. It does not kill the paper, but it means the outward-transport conclusion is conditional.\n\nThe citation pattern looks fine, and the paper engages with prior 1D work (Terada & Tokunaga, Potapov et al.) rather than ignoring it.\n\nVerdict: worth sending to a competent referee. The framework and the qualitative spatial constraint are useful; the abundance and transport inference need quantitative fitting with propagated uncertainties and a temperature test with icy opacities. I would not cite the 5.4% number in my own work, but I would cite the method and the structural constraints.\n\nBring it to reading group if you are interested in disk modeling practice; otherwise a skim of the temperature caveat is enough.","headline":"A solid 3D RT modeling effort with a genuinely new spatial constraint on ice in d216-0939, but the headline abundance and outward-transport claim rest on an untested no-ice temperature field.","tokens_in":21062,"tokens_out":1325,"would_cite":false,"duration_ms":17162,"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":"This paper argues that the upper and outer layers of the protoplanetary disk d216-0939 contain about 5.4 percent crystalline water ice beyond the snowline, ice that must have been transported outward because it cannot crystallize where it…","keywords":["protoplanetary disks","water ice","crystalline ice","radiative transfer","snowline","d216-0939","JWST","edge-on disks"],"falsifier":"Recompute the disk's radiative equilibrium including water-ice opacities and ask whether the sightlines that reproduce the 3 micron absorption pass through regions with temperatures above 130 K; if they do, crystalline ice could have formed in situ and the outward-transport inference collapses. Alternatively, spatially resolved 3 micron spectroscopy across the disk could locate the crystalline signal, and if it appears inside the model's crystallization annulus rather than beyond it, in situ formation is not ruled out.","tokens_in":19964,"feed_emoji":"❄️","tokens_out":7089,"duration_ms":77157,"temperature":0.7,"pith_summary":"Water ice in a protoplanetary disk sets the raw material for rocky planets, so where the ice sits and in what form matters for how planets get their water. This paper builds a 3D radiative-transfer model of the edge-on silhouette disk d216-0939 in the Orion Nebula Cluster, fitting JWST near- and mid-infrared spectra together with an archival HST image. The best-fit model places about 5.4% crystalline water ice by mass in the upper and outer disk layers beyond the 170 K snowline. Because those layers are too cold for amorphous ice to anneal into crystals in place, the paper concludes that material has been transported outward from the warmer inner disk. If correct, this makes d216-0939 a concrete case where radial transport reshapes the icy feedstock of planet formation.","feed_headline":"A cold disk's ice is 5.4% crystalline; something moved it outward","feed_subtitle":"JWST spectra plus radiative-transfer modeling place crystallized water ice beyond the snowline of disk d216-0939.","key_machinery":"The load-bearing object is a snowline-gated effective-medium ice mixture: measured MgSiO3/H2O refractive indices are mixed into the DSHARP (no-ice) refractory dust population in proportions of 3%, 6%, or 11% water ice by mass, and the mixture is placed only where the model's dust temperature is below the assumed 170 K snowline. The radiative transfer is done with 3D Monte Carlo simulations, using spectral importance sampling so that the 2.6 to 4.0 micron feature can be computed at 141 wavelengths at manageable cost. The same machinery splits the total flux into attenuated starlight, thermal dust emission, and scattered light, which is what lets the paper attribute the feature's shape to crystalline ice and its depth to the 5.4% abundance.","core_discovery":"The central claim is that the water ice absorption feature near 3 micron in d216-0939 is produced by dust with roughly 5.4% crystalline water ice, located beyond the snowline in the disk's upper and outer layers. The paper's model replaces the ice-free DSHARP dust mixture beyond the 170 K snowline with an effective-medium mixture of MgSiO3 and water ice, using 150 K (crystalline) optical constants; a fit with 5.4% ice content and an adjusted bulk density reproduces the observed feature, while amorphous 100 K ice does not place the absorption minimum correctly. Since the model's dust temperatures in the absorbing layers are far below the roughly 130 K crystallization threshold, the crystalline ice cannot have formed in situ. The paper therefore reads the feature as indirect evidence for outward transport of material through the upper disk layers, for example by a disk wind. The same model also decomposes the feature's flux: attenuated starlight dominates at 60-70%, with scattering and thermal dust emission contributing the rest.","pith_inferences":["A testable extension would be spatially resolved 3 micron spectroscopy: the crystalline signal should appear preferentially along sightlines that cross the upper layers beyond the model's 130 K contour; a mismatch would localize where the transport story needs revision.","The assumption that ice does not change the dust temperature could be checked directly: recomputing the thermal equilibrium with ice opacities might move the snowline, and the inferred transport distance would shrink or grow accordingly.","If crystalline ice in cold outer layers is common among moderately inclined disks, then the difference between d216-0939 and strongly inclined disks with amorphous ice could be a viewing-angle effect on a vertical crystallinity gradient, linking disk transport to the ice budget of planet-forming material.","Because amorphous ice traps volatiles such as CO and CO2, outward transport of crystallized ice would also redistribute volatiles; a dedicated model of the 4.3 micron CO2 feature in the same disk could test that coupling."],"forward_implications":["If the model is right, the small grains that JWST sees in d216-0939's upper layers carry about 5.4% crystalline water ice, and amorphous ice is negligible there.","The inferred outward transport means the upper disk is not a closed chemical system: material formed near the snowline can reach tens of au in the surface layers.","The 60-70% dominance of directly attenuated starlight means the 3 micron feature is primarily an absorption diagnostic, but scattering and thermal emission contribute up to about 40% combined and must be included in any retrieval.","The same snowline-plus-spectral-importance-sampling framework can be applied to other edge-on disks with JWST ice features, as the paper states in its conclusions."],"supporting_citations":[{"why":"Supplies the JWST NIRSpec and MIRI spectra, the detection of crystalline ice, and the compositional context that motivates the MgSiO3/H2O ice mixtures.","marker":"Potapov et al. (2025)"},{"why":"First reported crystallized water ice in d216-0939 and provides the prior optical-depth estimate this work compares against.","marker":"Terada & Tokunaga (2012)"},{"why":"Defines the DSHARP (no ice) dust composition whose opacities set the disk temperature structure and the ice-free baseline.","marker":"Birnstiel et al. (2018)"},{"why":"Provides measured MgSiO3/H2O refractive indices used to build the icy effective-medium dust mixtures.","marker":"Potapov et al. (2018)"},{"why":"Supplies the 100 K and 150 K silicate-water ice optical constants that distinguish amorphous from crystalline ice in the fits.","marker":"Potapov et al. (2021)"},{"why":"ALMA-based constraints on the large-grain disk structure and the 216 au gap set the parameter ranges for the small-grain disk.","marker":"Sheehan et al. (2020)"},{"why":"The HST H-alpha silhouette image is the spatial constraint that breaks degeneracies in the SED-only fit and fixes the outer radius and inclination.","marker":"Smith et al. (2005)"},{"why":"Establishes the roughly 120-140 K crystallization onset used to argue the cold outer layers cannot make crystalline ice in situ.","marker":"Jenniskens & Blake (1994)"},{"why":"Justifies the assumed 170 K snowline temperature that determines where water ice is placed in the model.","marker":"Martin & Livio (2012)"}],"fun_headline_variants":["Crystalline ice in disk d216-0939 hints at outward transport","JWST reveals 5.4% crystalline ice beyond the snowline","Disk ice crystallinity: a sign of outward material flow","Water ice in d216-0939 is partly crystalline—why?","Crystalline ice beyond snowline: d216-0939's icy secret"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the dust temperature map computed with ice-free DSHARP opacities, together with a snowline at 170 K, correctly places the water ice; if the observed layers are actually warm enough to crystallize ice in place, the transport conclusion fails.","fun_headline_variants_meta":{"raw":{"variants":["Crystalline ice in disk d216-0939 hints at outward transport","JWST reveals 5.4% crystalline ice beyond the snowline","Disk ice crystallinity: a sign of outward material flow","Water ice in d216-0939 is partly crystalline—why?","Crystalline ice beyond snowline: d216-0939's icy secret"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000283,"raw_usage":{"total_tokens":1733,"prompt_tokens":1065,"completion_tokens":668,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":574}},"tokens_in":681,"tokens_out":668,"duration_ms":7802,"temperature":1.0,"reasoning_tokens":574,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:11:42.632122+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the disk's radiative equilibrium including water-ice opacities and ask whether the sightlines that reproduce the 3 micron absorption pass through regions with temperatures above 130 K; if they do, crystalline ice could have formed in situ and the outward-transport inference collapses. Alternatively, spatially resolved 3 micron spectroscopy across the disk could locate the crystalline signal, and if it appears inside the model's crystallization annulus rather than beyond it, in situ formation is not ruled out.","supporting_citations":[],"review_version":1}