{"id":"4836d7c3-c6e1-4ed9-8e2a-8eac7d6450b5","arxiv_id":"2607.24934","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Coreshine SPHEREx spectra map 3 µm H2O ice across four prestellar cores; the two densest show an unexplained central drop in ice absorption that standard Bonnor-Ebert scattering models cannot reproduce.","lead":"SPHEREx spectra of scattered near-infrared light (coreshine) yield the first uniform ice-absorption maps of four nearby prestellar cores. Two dense cores show a central drop in water-ice absorption that simple scattering models cannot explain, opening a new way to map ices without background stars.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The high-density RADMC-3D null tests do not resolve the 0.002 pc flat core that dominates extinction, so the claimed failure to reproduce the central τice dip is not yet numerically secure.","rationale":"The reader correctly identified adequacy of the simplified Bonnor–Ebert/radiative-transfer null model as the weak point, but the sharper load-bearing issue is numerical resolution of the central region in the very calculation used to reject multiple-scattering explanations. The paper deserves credit for testing foregrounds, continuum choices, anisotropic illumination, grain growth, ice thresholds, saturation, and CO–H2O mixing, and for validating SPHEREx τice against background-star measurements. Those checks support the mapping demonstration and the reality of the observed central decrement. They do not yet establish that a properly resolved high-density scattering model cannot produce it. I would therefore retain the reader’s CONDITIONAL verdict: accept the proof-of-concept ice mapping, while explicitly requiring converged high-resolution radiative-transfer calculations before treating the central dip as unexplained by null scattering/dust models.","tokens_in":25265,"tokens_out":4875,"duration_ms":177559,"concrete_test":"Repeat the fiducial nH2 = 10^7 cm−3 RADMC-3D model with unchanged dust, ISRF, wavelengths, and photon count, but on a nested/AMR grid resolving r0 by at least four cells (Δx ≤ 5×10−4 pc), ideally alongside a 40/80/160-cell convergence sequence. Convolve the output to the SPHEREx grid and run the identical quadratic-continuum τice pipeline. If central τice and the central-versus-annulus ordering change by ≤10–15% and remain centrally peaked, the concern does not land; a larger change or an emerging dip would require revising the model-exclusion claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The observational τice maps, continuum tests, foreground checks, and L1544 background-star cross-calibration are persuasive; the concern is the inference that scattering/null models cannot explain the central dip. §3.1 adopts r0 = 0.002 pc and R = 0.2 pc, while Appendix C states that the opacity is dominated by the central flat region and that single scattering fails at nH2 = 10^7 cm−3 (τ ≈ 10). The high-density interpretation therefore rests on RADMC-3D. However, §3.3 uses a 40×40×40 Cartesian grid. If uniform over the 0.4 pc diameter, the cell size is ≈0.01 pc—five times r0—and an even-numbered grid need not sample the core center. The ρ0 plateau and steep inner r−2 rise are thus badly underresolved, potentially misrepresenting the multiple-scattering source function and wavelength-dependent escape from precisely the region producing the putative dip. The analytical model cannot rescue this regime because the paper itself says its single-scattering assumption fails there. This does not challenge the existence of the observed central drop, but it weakens the stronger statement that the tested models robustly exclude scattering or dust/ice explanations.","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The authors present the first spatially resolved maps of the 3 µm H2O ice absorption feature measured against coreshine (Galactic radiation scattered by micron-sized grains in dense cores), using SPHEREx spectrophotometry of four nearby prestellar cores (L1544, CrA 151, L260, L1512). The two lower-density cores show ice absorption peaking toward the center, as expected, while the two densest cores show a central drop in tau_ice. The observational pipeline is carefully validated: continuum-fit variants and S/N cuts change tau_ice by ≲10%, foreground subtraction cannot reverse the central-dip trend except at unphysical foreground levels (App. B), and SPHEREx tau_ice for L1544 background stars agrees with Goto et al. (2021) within ~10% (App. A). To interpret the maps, the authors build a single-scattering analytical model and RADMC-3D Monte Carlo models of a Bonnor-Ebert-like sphere; the models reproduce the continuum central dip but not the central tau_ice drop, and variants testing ISRF geometry, grain growth to 100 µm, ice-formation thresholds, and CO–H2O mixing also fail. Only an ad hoc central ice-saturation model reproduces the dip. The authors conclude that an unexplained physical or chemical effect operates in the densest regions, and advertise coreshine spectroscopy as a new probe of ice density and composition.","tokens_in":25572,"tokens_out":4168,"duration_ms":140545,"significance":"If it holds, this is a genuinely useful methodological advance: spatially continuous ice-abundance maps for any coreshine-bright core, rather than maps limited to the sparse set of background-star sightlines — enabling statistically large samples with an all-sky survey. The pipeline is reproducible (public SPHEREx/WISE/Herschel data, public RADMC-3D and OpTool codes), carries an external ~10% cross-calibration against IRTF background-star measurements (App. A), and is tested against foreground and continuum-fitting systematics. The reported central ice-absorption drop in the two densest cores, if confirmed as non-scattering in origin, would be a new and interesting constraint on ice chemistry at n_H2 ~ 10^6-10^7 cm^-3. The proof-of-concept value of the technique stands independently of how the interpretive question resolves.","major_comments":[{"comment":"§3.3 vs §3.1/App. C: the high-density null tests appear to under-resolve the region that dominates the physics. The BE profile uses r0 = 0.002 pc inside a 0.2 pc truncation radius, so a uniform 40^3 Cartesian grid spanning the cloud gives ~0.01 pc cells — 5x r0 — and an even-numbered grid need not place a cell at the center. App. C itself states that 'the opacity is dominated by the central flat region' (tau ~ kappa rho0 r0) and that the single-scattering assumption fails at n_H2 = 10^7 cm^-3 (tau ~ 10), so the analytical model cannot validate the Monte Carlo result precisely where the claimed discrepancy lives. The statement that the 40^3 grid was 'found sufficient' is asserted without a convergence test. Since the abstract's claim that 'none of these [models] can explain the reduced central absorption' rests entirely on these high-density RADMC-3D runs, a convergence demonstration is n","section":"§3.3, App. C (Eq. C1)"},{"comment":"Conclusion 3 and the abstract state the null result more strongly than the explored parameter space supports, given the magnitude of the mismatch. In the high-density fiducial model the central tau_ice reaches ~1.5 (Fig. 4, top right), while the observed central values are ~0.6-0.8 with a ring peaking near ~1 (Fig. 2) — a factor ~2 offset, well above the 10-15% measurement errors. The §4 variants test only coarse departures (a sharp 5-to-100 um a_max jump, a single 30 deg cap ISRF orientation, one threshold density). The body text is appropriately hedged ('Other physical explanations... cannot be ruled out', §4.4.1), but the abstract and Conclusions should be brought in line — e.g., 'within the model family tested' — or the variant suite should be broadened (continuous rgrowth, swept ISRF directions combined with the Major-Comment-1 convergence test). As written, a reader could take the","section":"Abstract; §5, item 3"}],"minor_comments":[{"comment":"§2.2: the 7.2-arcsec grid spacing is justified against the 6.2-arcsec native pixel, but adjacent grid points are not independent given the SPHEREx PSF (~2 pixels); please state the effective spatial resolution of the tau_ice maps and confirm the central dip spans more than one independent beam.","section":"§2.2"},{"comment":"§3.3: please justify the 10^6 photons per wavelength and the 200 MJy/sr flux floor. Photon noise and the floor could both affect the derived central tau_ice in the faint central depression; a brief noise estimate for the simulated images would help.","section":"§3.3"},{"comment":"Fig. 3 legends: the density labels (e.g., '0 = 0.1x10^6') do not state whether the quantity is n_H2 or rho0, nor units; the axis label 'ice' should be tau_ice. Same notation issue in Fig. 5.","section":"Fig. 3, Fig. 5"},{"comment":"Table 1's 'Central Density' column contains 'Low'/'High' rather than densities; either give approximate central densities (with the references already cited in §2.1) or rename the column.","section":"Table 1"},{"comment":"Fig. 2 radial profiles: describe the annulus binning and how the +/-1 sigma band is constructed (the text mentions only adding 10% to the upper error for the bandpass correction).","section":"§2.3, Fig. 2"},{"comment":"Typos/style: 'gasesous' (§4.4.1), 'upto' (§4.2), 'focussing' (§2.1), 'roughly corresponding' (§2.2); 'megajansky per steradian' in §3.3 without distance scaling makes the absolute flux comparison to data unclear.","section":"Various"}],"recommendation":"major_revision","confidential_remarks":"The methodology closely follows Hora et al. (2026), with the genuinely new element being the coreshine (rather than Galactic background) light source on core scales; the authors disclose this clearly. If the editor wants a second opinion on the radiative-transfer numerics, a referee with RADMC-3D Monte Carlo experience would be valuable, since the load-bearing question is grid convergence in the optically thick central core rather than the observations themselves."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The thing worth knowing is that this is a genuine methods first: uniform 3 µm H2O ice maps built from SPHEREx coreshine spectra on four nearby cores, not background stars. The observational side is careful. Continuum-fit variants move τ_ice by ≲10%, S/N cuts do not gut the trough, App. A recovers Goto et al. 2021 on the L1544 stars to ~10%, and foreground subtraction does not erase the central dip except at unphysical levels. Two lower-density cores peak at center; L1544 and CrA 151 drop. That pattern is in the data.\n\nWhat they do well is show that coreshine spectra carry ice information at ~1000 AU scales and that, for fixed geometry and density structure, RADMC τ_ice tracks η_ice. That alone makes the technique useful for anyone who already has dust/gas density maps and wants ice without hunting background stars. Citations to Pagani/Steinacker coreshine and to Hora et al. on large-scale SPHEREx ice are in the right places; they are not overselling prior art.\n\nThe soft spot is the stronger interpretive sentence—that Bonnor-Ebert scattering models with linear ice scaling, geometry tweaks, grain growth, thresholds, or simple CO mix cannot produce the dip, so something physical/chemical is required. The analytical single-scattering model is already outside its validity at n~10^7 (they say τ~10). The load-bearing nulls are the 40³ RADMC runs. With r0 = 0.002 pc and a ~0.4 pc box, cell size is ~0.01 pc—several times the flat core that dominates extinction—so the inner source function and wavelength-dependent escape are under-resolved. That does not invent the observed dip; it weakens the claim that the tested nulls robustly rule out radiative-transfer or microphysics explanations. Sample is four coreshine-selected cores; absolute columns still need external density structure. Those are real limits, not fatal ones.\n\nThis is for people doing ice chemistry and dense-core dust. Bring it to reading group for the method and the open puzzle. I would cite the mapping demonstration. It deserves a serious referee; ask for better-resolved RT (or a clear statement that the nulls are qualitative) before locking the chemistry claim. Accept with that revision path.","headline":"First real coreshine ice maps from SPHEREx; the maps and method hold up, but the claim that simple RT excludes a scattering origin for the central dip is not yet numerically secure.","tokens_in":26601,"tokens_out":601,"would_cite":true,"duration_ms":20455,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Scattered near-infrared light maps water ice across prestellar cores, and the densest cores show a central absorption drop that simple scattering models cannot explain.","keywords":["coreshine","water ice","prestellar cores","SPHEREx","ice absorption","near-infrared scattering","Bonnor-Ebert sphere","molecular clouds"],"falsifier":"Deep multi-band ice maps or background-star sightlines through the true centres of the densest cores that either confirm lower central water-ice column than at intermediate radii, or erase the dip once full multi-scattering and anisotropic illumination are included.","tokens_in":26329,"feed_emoji":"❄️","tokens_out":895,"duration_ms":38676,"temperature":0.7,"pith_summary":"This paper shows that light scattered inside dense star-forming cores (coreshine) can be turned into spatially resolved maps of the 3 µm water-ice absorption band, using all-sky SPHEREx spectra instead of sparse background stars. In four nearby prestellar cores, ice absorption rises toward the centre in the two lower-density objects, but drops in the innermost regions of the two densest cores. Analytical and Monte Carlo models of a Bonnor-Ebert sphere lit by scattering confirm that coreshine spectra track ice mass fraction when the density structure is known, yet none of the tested geometries, grain-growth cases, ice-formation thresholds, or simple CO–water mixes reproduce the central dip. The result both establishes a uniform-resolution ice probe for many cores and flags an unexplained physical or chemical effect in the densest prestellar gas.","feed_headline":"Scattered light maps ice in cores—densest ones dip at center","feed_subtitle":"Models cannot explain the central drop, flagging new physics in dense prestellar gas","key_machinery":"Coreshine ice-depth maps: pixel-by-pixel continuum-subtracted optical depth of the 3 µm H2O band measured against diffuse scattered Galactic light inside the core, interpreted with single-scattering analytics and 3D Monte Carlo radiative transfer of a Bonnor-Ebert sphere.","core_discovery":"Coreshine-derived SPHEREx spectra yield the first uniform, spatially resolved maps of 3 µm H2O ice absorption in four nearby prestellar cores. Two cores show the expected centre-peaked ice absorption; the two densest show a surprising central drop. Models of scattered light in a Bonnor-Ebert sphere demonstrate that such spectra robustly trace spatially varying ice density and composition, but standard geometric, grain-growth, and ice-composition variants cannot produce the reduced central absorption, implying an unexplained effect in the densest regions.","pith_inferences":["A confirmed central water-ice deficit would tighten links between dense-core freeze-out, oxygen budget, and the inventory of solids that later feed planet-forming disks.","A statistical sample of coreshine cores sorted by central density and evolutionary stage could separate local chemistry from illumination geometry as the driver of the dip.","Joint coreshine ice maps with gas-phase water and CO freeze-out tracers would test whether oxygen is reallocated among ices rather than simply depleted."],"forward_implications":["Ice absorption can be mapped at roughly thousand-AU resolution across the large share of cores that show coreshine, without needing background stars.","When dust or gas tracers fix the density structure, coreshine ice depth can constrain ice mass fraction and column.","The unexplained central dip marks densest, evolved prestellar regions as places where ice may not scale linearly with gas.","The same scattered-light method can be extended with SPHEREx to other ice bands and much larger core samples."],"fun_headline_variants":["SPHEREx coreshine maps show ice dip at centers of densest cores","First scattered-light ice maps reveal central drop in densest cores","Coreshine ice maps flag unexplained absorption dip in dense cores","SPHEREx traces H2O ice via coreshine—densest cores defy peak trend","Scattered-light maps find central ice drop models cannot explain"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"That a simple spherical density profile with constant or simply thresholded ice fraction, ordinary grain sizes, and a basic radiation field is a fair null model—so failure to match the central ice dip means real physics or chemistry rather than missing complexity in dust or light.","fun_headline_variants_meta":{"raw":{"variants":["SPHEREx coreshine maps show ice dip at centers of densest cores","First scattered-light ice maps reveal central drop in densest cores","Coreshine ice maps flag unexplained absorption dip in dense cores","SPHEREx traces H2O ice via coreshine—densest cores defy peak trend","Scattered-light maps find central ice drop models cannot explain"]},"model":"grok-4.5","effort":"low","cost_usd":0.006652,"raw_usage":{"total_tokens":1735,"prompt_tokens":885,"num_sources_used":0,"completion_tokens":79,"cost_in_usd_ticks":66524000,"prompt_tokens_details":{"text_tokens":885,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":771,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":885,"tokens_out":79,"duration_ms":13707,"temperature":1.0,"reasoning_tokens":771,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T05:29:02.290579+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Deep multi-band ice maps or background-star sightlines through the true centres of the densest cores that either confirm lower central water-ice column than at intermediate radii, or erase the dip once full multi-scattering and anisotropic illumination are included.","supporting_citations":[],"review_version":1}