{"id":"e7554c05-399d-44f6-9e2c-5729d6d1927f","arxiv_id":"2412.04356","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Water ice is detected at 3 microns in both scattered light and silhouette of the Orion Nebula disk 114-426 with JWST NIRCam, yielding ice-to-refractory mass ratios up to 0.2.","lead":"Using JWST NIRCam images in 12 bands, astronomers found the 3 micron water ice absorption feature in the outer parts of the edge-on planet-forming disk 114-426 in the Orion Nebula, and measured ice-to-refractory mass ratios up to 0.2. The result shows that water ice can survive in the moderately irradiated environment typical of where most stars form, which matters for understanding the availability of water for planet formation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative ice abundances and dust masses in §3.3 rest entirely on treating the silhouette as pure absorption (Eq. 1), but the forward-scattering justification is qualitative and the resulting bias is never estimated.","rationale":"The reader's CONDITIONAL verdict identifies the pure-absorption assumption of Eq. (1) as the weakest link, and I agree that this is the single most load-bearing concern. Every number in the abstract's quantitative claim—ice-to-refractory ratio, grain size range, and total dust plus ice mass—is derived from fits that set scattering opacity to zero. The paper's justification for neglecting scattering is physically plausible but qualitative, and the geometry (extended background nebula, finite pixel beam, forward-peaked scattering) makes the net effect nontrivial. The Appendix convolution test checks PSF blurring, not scattering, so it does not resolve this concern. I also considered the unresolved F335M red-edge mismatch (§4.1) and the lack of propagated uncertainties, but these affect the precision and physical interpretation of an already-detected feature rather than the core mechanism that converts the observed dip into abundance and mass. The 3 µm water-ice detection itself is well supported—it appears in both scattered-light and silhouette spectra and is accompanied by a wavelength-dependent widening of the dark lane in F300M—so the concern does not warrant rejection. The appropriate response is to keep the conditional verdict and require the scattering-inclusive modeling test as the condition for full acceptance.","tokens_in":18310,"tokens_out":8271,"duration_ms":99980,"concrete_test":"Take a representative best-fit pixel (e.g., location B in Fig. 7) with its fitted Σ_dust, ice-to-refractory ratio, and amax, and compute the transmitted F/F_bg in the fitted NIRCam bandpasses using a plane-parallel slab radiative transfer model that includes both absorption and scattering, using the same optool dust opacities, a Henyey-Greenstein phase function with the model g, and the measured nebular surface brightness as the extended background source. Compare the resulting band ratios to the pure-absorption values from Eq. (1).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—ice-to-refractory ratios up to ~0.2, maximum grain sizes of 0.25–5 µm, and 0.46 M⊕ of solids—is obtained by fitting Eq. (1), F/F_bg = exp(−κ_abs Σ_dust), which sets scattering opacity to zero. The justification in §3.3 is that scattering is preferentially forward-directed, but that argument applies most directly to a point source behind a slab. Here the background is an extended nebula filling the field, so forward-scattered photons removed from one ray can be replaced by photons scattered into that ray from adjacent background directions. Whether net attenuation survives depends on the grain albedo, the phase function, and the beam/pixel geometry; it is not guaranteed to be negligible. This is not a small correction confined to one parameter: it enters every pixel fit, and it is wavelength-dependent because albedo and asymmetry parameter vary across 1–5 µm and with ice mantle thickness. The paper offers no quantitative estimate of the resulting bias in ice-to-refractory ratio, grain size, or surface density, yet these numbers are quoted without propagated uncertainty. A scattering-inclusive radiative transfer test is therefore required before the abundance and mass values can be taken at face value. The detection of the 3 µm dip itself is on firmer ground, since it is also supported by the wavelength-dependent morphology of the dark lane (§3.1), but the quantitative translation into composition and mass is not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents JWST/NIRCam observations of the edge-on silhouette disk 114–426 in 12 bands spanning 1–5 μm. The authors report a dip at 3 μm in both the scattered-light lobes and the outer silhouette ansae, which they attribute to water ice. They fit the silhouette absorption spectra with a pure-absorption model of icy grains (Eq. 1) to derive ice-to-refractory mass ratios up to 0.18, maximum grain sizes of 0.25–5 μm, and a total solid mass of 0.46 M⊕ in the silhouette. They further report excess absorption in the Paα bands, wavelength-dependent lobe asymmetries that suggest a tilted inner disk, and a photodesorption analysis arguing that water ice can survive in the external radiation field of the Orion Nebula.","tokens_in":18583,"tokens_out":4252,"duration_ms":41716,"significance":"If the quantitative results hold, this is a valuable addition to the small sample of edge-on disks with direct ice detections, and the silhouette geometry offers a rare opportunity to measure ice abundances in absorption rather than through scattering models. The detection of the 3 μm dip is well supported by consistent F300M decrements across multiple pixels in both the scattered-light and silhouette regions, and the paper is commendably transparent about the limitations of its model, including the exclusion of Paα-contaminated bands and the imperfect fit to the red edge of the ice feature. The use of public JWST data, a documented reduction pipeline, and a standard dust-opacity code (optool) makes the analysis reproducible in principle.","major_comments":[{"comment":"The quantitative results—ice-to-refractory ratios, grain sizes, and the integrated mass—rest entirely on the pure-absorption approximation F/F_bg = exp(−κ_abs Σ_dust), with scattering opacity set to zero. The justification that scattering is preferentially forward-directed is qualitative and, as argued in the text, most directly applies to a point source behind a slab. Here the background is an extended nebula filling the field, so forward-scattered photons from neighboring background directions can replace photons removed from a given line of sight; whether net attenuation survives depends on the grain albedo, the phase function, and the pixel geometry, all of which vary with wavelength and ice mantle thickness. The paper offers no quantitative estimate of the resulting bias in the derived ice-to-refractory ratio, grain size, or surface density. A scattering-inclusive test—even a simple slab calculation with a Henyey–Greenstein phase function—is needed before the quoted abundance and mass values can be taken at face value. The detection of the 3 μm dip itself is on firmer ground, but the model-dependent numbers are not.","section":"§3.3, Eq. (1)"},{"comment":"The model systematically overpredicts the F335M band, which is included in the fit (only F182M, F187N, F212N, and F470N are excluded). The paper acknowledges in §4.1 that the red edge of the water-ice feature is not reproduced by the adopted ice optical constants and grain shape, and suggests that more sophisticated modeling (additional ice species, discrete-dipole approximation, RADMC-3D) would be needed. Given that F335M is one of the few bands sampling the long-wavelength wing of the feature, the derived ice-to-refractory ratios and maximum grain sizes are partly constrained by a band that the model fails to fit. The authors should either restrict the fit to the bands the model describes (e.g., F277W, F300M, F360M, F444W) and state the resulting parameter ranges, or explicitly present the current best-fit values as provisional pending a better red-edge treatment.","section":"§3.3 and §4.1"}],"minor_comments":[{"comment":"In Eq. (1), Σ_dust is described as the dust surface density, but the model includes the ice mantle as part of the solid mass; renaming this quantity or clarifying that it is the total solid (dust plus ice) surface density would avoid confusion, especially when the total mass is summed in §3.3.","section":"§3.3"},{"comment":"The three panels of Figure 6 lack visible color-bar labels or units in the manuscript text; if they appear in the published figure, please ensure the units (e.g., g cm^−2 for surface density) are legible.","section":"Figure 6"},{"comment":"The photodesorption balance assumes a single-angle incident FUV field, but the paper later notes that photons entering from varying angles are neglected (citing Cleeves et al. 2013). A brief sentence acknowledging that Eq. (3) shares this simplification would improve consistency.","section":"§4.4, Eq. (3)"},{"comment":"There are a few minor typographical issues, such as a missing space in '1.875µm' (p. 4) and a missing period after 'v_H2O = 216 m s^−1' (p. 12); these should be corrected in the final version.","section":"General typos"}],"recommendation":"major_revision","confidential_remarks":"The central detection of the 3 μm water-ice feature is convincing and well presented, and the paper is a good match for the journal. The main reservation is that the quantitative abundance and mass claims depend on the pure-absorption approximation of Eq. (1) without a quantified bias estimate; the overprediction of F335M further weakens the model-based numbers. These issues are addressable with additional analysis or with a more cautious presentation of the quantitative results, so I am recommending major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The 3 micron ice detection is real, and the paper deserves a proper review. The quantitative abundance and mass numbers are the soft spot: they rest on pure absorption with scattering ignored, and the paper doesn't quantify that bias.\n\nWhat's new: this is the first JWST NIRCam 12-band look at 114-426 and the first detection of the water ice feature in this disk. The evidence is good. The F300M dip appears in multiple pixels in both the scattered light and the silhouette, the dark lane morphology shows the expected wavelength-dependent opacity trend with a deviation at 3 micron, and the convolved-image sanity check keeps the signal. The Pa alpha excess is a nice added result. The authors also do careful things: aligning the images, estimating the background plane, excluding bands that might be contaminated by gas lines, and showing what changes when you homogenize the PSF.\n\nWhere it gets soft: Equation (1) sets scattering to zero because scattering is 'preferentially forward-directed.' That argument is qualitative and, for an extended background, incomplete. Photons scattered out of one ray can be replaced by photons scattered in from neighboring rays. The net attenuation depends on the albedo and phase function; it is not obviously negligible. The paper gives no estimate of the resulting bias in ice-to-refractory ratio, grain size, or surface density. So I read the maps and the 0.46 M_Earth as model-dependent estimates, not measurements. The F335M overprediction is another unresolved piece; the authors are upfront about it, but it means the red edge of the ice feature is not reproduced.\n\nNone of this kills the central detection. Water ice is there. But the quantitative reach of the paper is a step beyond what the modeling justifies as stated. The fix is either a scattering-inclusive radiative transfer test or a softer presentation of the numbers. That's a standard revision request, not a rejection reason.\n\nThe photodesorption survival analysis is a reasonable order-of-magnitude check, with honest caveats. The literature context is well handled.\n\nThis paper is for disk observers and astrochemists. I'd bring it to reading group, and I'd cite it for the ice detection, though not for precise abundances. A serious referee should engage with it.","headline":"The 3 micron water ice detection is solid and the paper deserves a real referee; the quantitative abundance and mass numbers, however, rest on an unquantified pure-absorption assumption and should be treated as model-dependent until scattering is properly handled.","tokens_in":19195,"tokens_out":3298,"would_cite":true,"duration_ms":36673,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.85.Hp"],"model":"deepseek-v4-flash","headline":"JWST images of the edge-on Orion disk 114–426 reveal a 3-micron water-ice absorption dip in both scattered light and silhouette, with ice-to-refractory ratios up to ~0.2 and 0.46 Earth masses of solids in the outer ansae.","keywords":["water ice","protoplanetary disk","silhouette disk","Orion Nebula","JWST NIRCam","3 micron absorption feature","grain growth","photodesorption"],"falsifier":"A spatially resolved spectrum at $R \\gtrsim 1000$ across 2.5–3.6 μm would measure the true profile and depth of the 3 μm feature on the silhouette; if the water-ice column required to fit that spectrum disagreed with the band-derived abundances beyond the quoted uncertainties, or if the feature shape showed no water-ice structure, the photometric detection and its quantified ratios would fail. A separate test targets the survival argument: mapping the outer-disk gas density with resolved molecular-line observations, since the adsorption–photodesorption balance predicts that below $n_H \\approx 10^8$ cm$^{-3}$ most water should be in vapor; a firmly low gas density with a persistent 3 μm dip would mean the icy grains are not steady-state residents of this disk.","tokens_in":18106,"feed_emoji":"🧊","tokens_out":16297,"duration_ms":145253,"temperature":0.7,"pith_summary":"This paper reports the detection of water ice in the outer parts of 114–426, a large edge-on protoplanetary disk in the Orion Nebula. Because the disk's outer edges are seen in silhouette against the bright nebula behind them, the authors can study planet-forming material at these radii in pure absorption, a geometry that sidesteps the scattering effects that usually complicate ice measurements in disks. JWST/NIRCam images in twelve bands from 1 to 5 μm show a dip at 3 μm, the spectral signature of water ice, in both the scattered-light lobes and the silhouette. Fitting the silhouette spectra with a model of refractory grains bearing ice mantles yields ice-to-refractory mass ratios up to ~0.2, maximum grain sizes of 0.25–5 μm, and a total of 0.46 Earth masses of ice and dust in the silhouette region. The result matters because water ice in disks is the reservoir from which planets draw their water, and this is direct evidence that ice survives in the moderately irradiated, cluster-like environment where most stars form.","feed_headline":"JWST finds water ice in an edge-on Orion disk","feed_subtitle":"A 3-micron dip in disk 114–426's silhouette puts the ice-to-rock ratio near 0.2 and totals 0.46 Earth masses of solids.","key_machinery":"The load-bearing identity is the pure-absorption law $F_\\nu/F_{\\nu,\\mathrm{bg}} = e^{-\\kappa_{\\mathrm{abs}}\\Sigma_{\\mathrm{dust}}}$ (Equation 1), which turns each pixel's measured dimming of the nebular background into a dust column density. The absorption opacity $\\kappa_{\\mathrm{abs}}$ is generated with a dust model of refractory cores bearing water-ice mantles, parametrized by the ice-to-refractory mass ratio, the maximum grain size (with a power-law size distribution of index $-3.5$ and a fixed minimum size of 0.005 μm), and the dust surface density $\\Sigma_{\\mathrm{dust}}$; a grid search minimizes $\\chi^2$ over the eight uncontaminated bands. Two sets of bands are deliberately excluded from the fit: the Paschen α bands F182M and F187N, and the narrow bands F212N and F470N centered on molecular hydrogen transitions. The second mechanism in the paper is the survivability calculation: an equilibrium between photodesorption of ice by ultraviolet photons and re-adsorption of water vapor, solved for the ice fraction as a function of the external FUV field, gas density, and water and dust abundances, which the authors use to argue that the detected ice is consistent with the disk's radiation environment.","core_discovery":"The paper's central claim is that the 3 μm dip in the disk's spectrum, present both in the scattered-light lobes and in the silhouette of the outer ansae against the Orion Nebula, is the vibrational absorption feature of water ice, and that the silhouette version can be quantified in terms of ice abundance. Each pixel's brightness relative to the nebular background is taken to obey $F_\\nu/F_{\\nu,\\mathrm{bg}} = e^{-\\kappa_{\\mathrm{abs}}\\Sigma_{\\mathrm{dust}}}$, with scattering neglected because it is preferentially forward-directed. The opacity $\\kappa_{\\mathrm{abs}}$ comes from models of ice-mantled refractory grains (87% amorphous pyroxene and 13% amorphous carbon by mass), and a grid search over three free parameters, the ice-to-refractory mass ratio, the maximum grain size, and the dust surface density, fits eight of the twelve NIRCam bands at each pixel of the silhouette. The fits return ice-to-refractory mass ratios of 0 to 0.18, maximum grain sizes of 0.25 to 5 μm, and a total solid mass of 0.46 $M_\\oplus$ in the two ansae, with the ice fraction and grain size declining with radius. The authors further report excess absorption in the two bands containing the Paschen α line, which they attribute to excited atomic hydrogen, and a wavelength-dependent brightness asymmetry between the scattered-light lobes that they interpret as evidence of a tilted inner disk. A separate equilibrium calculation between ultraviolet photodesorption and re-adsorption of water leads them to conclude that water ice can survive the external radiation field of this intermediate cluster environment.","pith_inferences":["Because the F300M filter cannot capture the full depth of the 3 μm feature, the derived ice-to-refractory ratios are arguably lower limits; a deeper or broader feature would raise them and would also explain the model's systematic tendency to overpredict the F335M band.","The same silhouette technique should transfer to other edge-on disks that project onto bright H II region backgrounds, where NIRCam's multi-band coverage could map ice abundance in absorption without the scattering complications of scattered-light measurements.","If the Paschen α absorption is confirmed spectroscopically, the disk's outer edge becomes a natural screen for mapping n = 3 hydrogen across a protoplanetary disk, a new tracer of ultraviolet-driven excitation that requires only narrowband imaging to exploit.","A radiative-transfer treatment that includes forward scattering on these same lines of sight would quantify the systematic bias in the derived grain sizes and ice fractions, and would show whether the pure-absorption assumption pushes the ice-to-refractory ratio toward its ceiling of 0.18."],"forward_implications":["If the 3 μm dip is water ice as modeled, ice makes up roughly a fifth of the solid mass relative to refractory material in the outer ansae, a direct compositional constraint for models of where planets get their water.","Grain sizes of 0.25–5 μm in the silhouette, together with the absence of millimeter grains there in ALMA data, show that outer-disk dust has grown past interstellar sizes without forming pebbles, implying stalled growth or inward drift of the largest grains.","The excess absorption in the Paschen α bands points to a reservoir of hydrogen in the n = 3 state, likely pumped by Lyβ photons, making the silhouette a probe of gas excitation as well as dust composition.","The wavelength-flipping lobe asymmetry and lateral offset of the scattered-light lobes imply a tilted inner disk, so the outer silhouette and the inner disk are not coplanar in this system.","Water ice that survives external ultraviolet fluxes near $10^2$–$10^4$ G$_0$ would establish that ice can persist in the typical clustered, intermediate-radiation environment of star formation, not only in quiescent regions."],"supporting_citations":[{"why":"Supplies the custom-reduced NIRCam mosaics of the Orion Nebula whose astrometric and photometric alignment underpins every spectrum in the paper.","marker":"McCaughrean & Pearson 2023"},{"why":"Establishes that silhouette disks like 114–426 lie in the foreground of the nebula and identifies the PSF contamination problem the paper re-checks by convolution.","marker":"McCaughrean & O'Dell 1996"},{"why":"The prior HST absorption study of the same silhouette, whose 0.2–0.7 μm grain-size result and external FUV estimate anchor the comparison and the ice-survival calculation.","marker":"Miotello et al. 2012"},{"why":"ALMA continuum and CO absorption observations that provide the 9.9 M⊕ comparison dust mass and the gas temperature used in the survival argument.","marker":"Bally et al. 2015"},{"why":"The optool code used to generate the ice-mantle grain opacity spectra that define the model fits.","marker":"Dominik et al. 2021"},{"why":"Provides the amorphous pyroxene optical constants that set the refractory-core opacity.","marker":"Dorschner et al. 1995"},{"why":"Provides the amorphous carbon optical constants in the refractory-core mixture.","marker":"Zubko et al. 1996"},{"why":"Provides the water ice optical constants that determine the shape and depth of the modeled 3 μm feature.","marker":"Warren & Brandt 2008"},{"why":"Models showing that external heating cannot remove the water snow line at this star's separation from θ1 Ori C, supporting ice survival.","marker":"Haworth 2021"},{"why":"Supplies the photodesorption yield formula at the core of the adsorption–photodesorption equilibrium used to argue the ice can persist.","marker":"Öberg et al. 2009"}],"fun_headline_variants":["JWST reveals water ice in edge-on Orion disk","3-micron dip confirms water ice in disk 114-426","Water ice survives in Orion disk, JWST shows","Ice fraction up to 0.2 in Orion disk silhouette","JWST sees water ice absorption in Orion protoplanetary disk"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the dark outer edges of the disk dim the background nebula by absorption alone, with scattering so strongly forward that it does not push background photons out of our line of sight and with no light emitted by the disk itself or by material in front of it, so that all of the derived ice fractions, grain sizes, and masses rest on that single radiative assumption.","fun_headline_variants_meta":{"raw":{"variants":["JWST reveals water ice in edge-on Orion disk","3-micron dip confirms water ice in disk 114-426","Water ice survives in Orion disk, JWST shows","Ice fraction up to 0.2 in Orion disk silhouette","JWST sees water ice absorption in Orion protoplanetary disk"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000313,"raw_usage":{"total_tokens":1879,"prompt_tokens":1143,"completion_tokens":736,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":759,"completion_tokens_details":{"reasoning_tokens":652}},"tokens_in":759,"tokens_out":736,"duration_ms":7476,"temperature":1.0,"reasoning_tokens":652,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:31:21.319646+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A spatially resolved spectrum at $R \\gtrsim 1000$ across 2.5–3.6 μm would measure the true profile and depth of the 3 μm feature on the silhouette; if the water-ice column required to fit that spectrum disagreed with the band-derived abundances beyond the quoted uncertainties, or if the feature shape showed no water-ice structure, the photometric detection and its quantified ratios would fail. A separate test targets the survival argument: mapping the outer-disk gas density with resolved molecular-line observations, since the adsorption–photodesorption balance predicts that below $n_H \\approx 10^8$ cm$^{-3}$ most water should be in vapor; a firmly low gas density with a persistent 3 μm dip would mean the icy grains are not steady-state residents of this disk.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Models showing that external heating cannot remove the water snow line at this star's separation from θ1 Ori C, supporting ice survival."}],"review_version":1}