{"id":"094f7d07-19ac-41d0-9ed2-20a8aec89fa4","arxiv_id":"2504.20742","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A numerical framework couples non-equilibrium condensation and freezing kinetics with exhaust plume flow, predicting delayed and spatially different contrail formation than equilibrium models.","lead":"This paper builds a computer model of the first seconds of contrail formation, tracking how water vapor condenses into droplets and freezes into ice inside an aircraft exhaust plume. It shows that accounting for the slow, energy-limited pace of nucleation changes where and how much ice forms compared with older equilibrium models.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. 10's heterogeneous condensation rate omits soot number/surface area, so the non-equilibrium reference solution may not depend on soot loading as intended.","rationale":"The paper's strongest claim is that resolving nucleation and growth kinetics changes contrail onset, location, and morphology relative to equilibrium models. For that claim to hold, the non-equilibrium model must be a faithful representation of the microphysics. The reader's RANS/Sct concern is legitimate and is acknowledged by the authors as a limitation, but it affects both equilibrium and non-equilibrium solutions as a modeling-uncertainty issue. The heterogeneous nucleation closure in Eq. 10 is a more targeted internal inconsistency that directly controls a dominant nucleation pathway: the bimodal droplet distribution and the kerosene-versus-hydrogen contrast are presented as key quantitative results, and both depend on the relative strength of heterogeneous and homogeneous nucleation. If the soot number concentration and surface area do not enter Eq. 10, the model cannot quantitatively represent soot-activated condensation, and the claimed non-equilibrium delay and spatial confinement are not a well-posed function of fuel and soot loading. I am not claiming the qualitative conclusion is false; the direction of equilibrium overprediction is physically plausible, and the paper is honest about scope. However, the equations as printed do not support the quantitative claim without a correction or clarification. The conditional verdict is therefore appropriate; no change in verdict is needed, but the manuscript should be revised to fix or clarify Eq. 10 and demonstrate soot-concentration sensitivity before the central quantitative comparison is relied upon.","tokens_in":18785,"tokens_out":9951,"duration_ms":107025,"concrete_test":"Run a zero-dimensional sensitivity test with fixed thermodynamic state and supersaturation: evaluate the heterogeneous condensation source in Eqs. 10-11 at soot number concentrations of 10^6, 10^8, and 10^10 cm^-3, keeping the log-normal size distribution shape unchanged. If the source term is unchanged, Eq. 10 lacks the N_p 4 pi r_p^2 prefactor. Then compare the implemented J_het with the Fletcher expression J_het = N_p 4 pi r_p^2 K_het exp(-f(theta) Delta G / kT) at the baseline state and report the ratio. If the ratio differs by more than an order of magnitude, re-run the unmixed nozzle baseline and the hydrogen no-soot case with the corrected heterogeneous nucleation closure and re-check whether the equilibrium versus non-equilibrium differences in Sec. IV.B survive.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section II.B presents heterogeneous condensation as the soot-dependent pathway that anchors the kerosene baseline and the hydrogen comparison, but Eq. 10 defines J_het with the same homogeneous prefactor and only a reduced barrier exp(-f(theta) Delta G / kT). There is no factor proportional to soot number concentration or surface area. The text reinforces this: soot influence is said to be determined solely by the assumed contact angle, and soot is passively transported in Eq. 26. In standard heterogeneous nucleation theory (Fletcher 1958), the volume nucleation rate is J_het proportional to N_p 4 pi r_p^2 K exp(-f(theta) Delta G / kT), so the prefactor must include the substrate surface-area concentration. The absence matters because the central comparison between equilibrium and non-equilibrium predictions, the bimodal droplet spectrum in Fig. 5, and the no-soot hydrogen case in Sec. IV.C all depend on the relative strength of heterogeneous and homogeneous pathways. If Eq. 10 is implemented as written, varying the stated soot concentration (10^8 cm^-3) and its log-normal size distribution has no effect on condensation; the model is only toggling a mechanism on or off, not testing soot-mediated nucleation. The heterogeneous freezing source in Eq. 14 does include n_p,i 4 pi r_p,i^2, making the omission in Eq. 10 an internal inconsistency rather than a stylistic choice. This is a correctness risk in the non-equilibrium reference solution, not merely the acknowledged turbulence-closure caveat.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a numerical framework for near-field contrail formation that couples compressible, multi-component RANS flow with non-equilibrium phase-change modeling: homogeneous and heterogeneous condensation, homogeneous and immersion freezing, and polydispersed droplet/ice/soot populations via a method-of-moments approach. The framework is applied to two simplified axisymmetric high-bypass turbofan nozzle configurations under cruise conditions. The central claim is that resolving finite-rate nucleation and growth materially changes the onset, location, and morphology of condensed phases relative to an equilibrium Schmidt–Appleman-style treatment, which is said to overpredict condensed mass and spatial extent. A parametric hydrogen-like case and a mixed-nozzle geometry are also examined.","tokens_in":19035,"tokens_out":8050,"duration_ms":84467,"significance":"If the central claim holds, the paper would provide a useful physics-based complement to equilibrium contrail criteria and a modular platform for studying fuel and nozzle effects on early plume microphysics. It has clear strengths: the thermophysical properties are taken from established sources (IAPWS-95, REFPROP), the microphysics models are adapted from published frameworks, a grid-convergence check is reported, and the solver is described in enough detail to be reproduced by a specialist group. However, the heterogeneous condensation model as written does not depend on soot loading, which directly affects the soot-mediated pathway, the bimodal droplet spectra, and the kerosene/hydrogen comparison. The quantitative predictive claims are therefore conditional on a correction to Equation (10) and on sensitivity of the results to the mixing and microphysical parameters.","major_comments":[{"comment":"The heterogeneous condensation rate J_het in Eq. (10) is written with the same kinetic prefactor as the homogeneous rate and depends on soot only through the reduced barrier exp(-f(theta) Delta G* / kT). It contains no factor proportional to the soot number concentration N_p or to the soot surface-area concentration. The text immediately below Eq. (11) confirms this: \"Their influence on nucleation is determined solely by the assumed contact angle.\" Consequently, the soot loading specified in Table 2 (10^8 cm^-3, log-normal distribution with mean radius 50 nm) has no effect on the heterogeneous condensation rate as written, and the kerosene versus no-soot hydrogen comparison in Section IV.C reduces to toggling the mechanism on or off rather than testing soot-mediated condensation. This also undermines the attribution of the bimodal droplet spectrum in Fig. 5 to a soot-activated heterogeneous mode. Standard heterogeneous nucleation theory (Fletcher, Ref. [29]) requires a prefactor proportional to the aerosol surface-area concentration (e.g., N_p 4 pi r_p^2 times a suitable kinetic coefficient). This is an internal inconsistency with Eq. (14), where the heterogeneous freezing rate explicitly contains n_{p,i} 4 pi r_{p,i}^2. The authors should correct Eq. (10) and re-evaluate the affected results, or justify in detail why a volume-based rate without soot surface area is appropriate.","section":"II.B, Eq. (10)"},{"comment":"The quantitative claims about onset, location, and morphology depend on the steady RANS representation of turbulent mixing, specifically the Spalart–Allmaras closure with a constant turbulent Schmidt number Sct = 0.7. Since nucleation rates are exponentially sensitive to the local supersaturation, errors in entrainment and scalar mixing can shift the supersaturated region and thus the nucleation timing and location that are central to the equilibrium versus non-equilibrium comparison. The manuscript acknowledges in Section V that turbulence effects are not treated in detail and will be addressed with turbulence-resolving simulations, but the central claim would be substantially strengthened by at least a sensitivity study over Sct (for example, 0.5 to 1.0) or by a comparison with a turbulence-resolving simulation for one configuration. Without such a check, the reader cannot distinguish model-specific mixing errors from the physical non-equilibrium effects that the paper emphasizes.","section":"III.A, III.B, IV.B"},{"comment":"The microphysical parameters that control the phase-change kinetics are fixed without uncertainty quantification: the contact angle theta = 90 degrees, the ice-active surface site density n_s = 10^8 m^-2, and the soot concentration and size distribution. Given the exponential sensitivity of nucleation rates to the barrier and the direct proportionality of heterogeneous rates to n_s and to soot surface area, the reported quantitative differences (for example, the liquid mass-fraction peaks in Section IV.B) should be accompanied by a sensitivity range, at least for theta and n_s. This is especially important for the hydrogen-like case in Section IV.C, where the claim that homogeneous nucleation alone can initiate contrail formation depends on the absence of soot and on the kinetic competition between pathways.","section":"II.B, Table 2, IV.B, IV.C"}],"minor_comments":[{"comment":"The coefficients alpha and beta in Eq. (16) are not defined in the text or nomenclature beyond being called \"parameters according to Young\"; the reader is told that alpha = 11 and beta = 0 but not their physical meaning or units.","section":"II.B, Growth Kinetics, Eq. (16)"},{"comment":"In the right panel of Fig. 5, the ordinate is described as \"normalized\" but the normalization reference is not stated, and the axis labels are missing; please specify the normalization used for both nucleation-rate profiles.","section":"Fig. 5"},{"comment":"The statement that \"the simulated ice crystal concentrations and sizes remained below typical optical visibility thresholds for contrails\" is given without a quantitative threshold or citation; please state the visibility criterion used and its source.","section":"IV.B"},{"comment":"In the mixed-nozzle case, the statement that freezing is \"delayed\" and that \"no ice is detected within the region shown\" should be tied explicitly to the downstream extent of the plotted domain, since ice may form farther downstream.","section":"IV.C"},{"comment":"There are several typographical inconsistencies, including occasional use of \"homogenous\" where \"homogeneous\" is intended (e.g., Fig. 5 caption and Section II.B); a careful proofreading pass would improve readability.","section":"General"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nWhat is actually new: the authors take steam-turbine non-equilibrium condensation machinery (CNT with Kantrowitz correction, Young growth law, method of moments) and couple it to cloud-ice microphysics (Koop homogeneous freezing, Vali-style immersion freezing) inside a compressible RANS solver, then apply the combination to a simplified high-bypass nozzle. I do not know of another paper doing that exact integration, and the equations are laid out carefully enough to reproduce. The paper is honest about scope: no experimental validation, simplified geometry, steady RANS, no turbulence-resolving treatment. The equilibrium-versus-non-equilibrium comparison is a sensible target, and the bimodal droplet spectrum is the most interesting result. Grid convergence is reported, external benchmark models are used, and the citations look appropriate. No fitted-to-output circularity; the self-citations are to the solver, which is tooling.\n\nThe stress-test note is correct, and the reader's report was too lenient. Eq. 10 gives the heterogeneous condensation rate the same homogeneous prefactor and only a reduced barrier exp(-f(theta) dG/kT). There is no factor proportional to soot number concentration or surface area. Standard Fletcher-type theory has a volumetric rate proportional to N_p 4 pi r_p^2 K exp(...). The text in Section II.B even says soot influence is determined solely by contact angle. Compare Eq. 14, where immersion freezing includes n_p,i 4 pi r_p,i^2. So the two soot-dependent pathways are internally inconsistent as written. If Eq. 10 is implemented literally, changing soot concentration from 1e8 cm^-3 to zero does not affect condensation; the model can only switch a mechanism on or off, and the hydrogen-like no-soot case in Section IV.C is not testing soot loading. That matters because the bimodal spectrum, the heterogeneous-versus-homogeneous partitioning in Fig. 5, and the fuel comparison all depend on relative pathway strength. This is a formulation error in the non-equilibrium reference solution, not just the acknowledged turbulence caveat.\n\nThe other soft spots are the expected ones: steady RANS with fixed Sct=0.7 controls the supersaturation field, no validation, no code or data release. Those are revision items, not fatal. The broad qualitative conclusion—finite kinetics delay and reduce condensed mass relative to equilibrium—is physically plausible and would probably survive a corrected Eq. 10, but the quantitative pathway partitioning needs to be redone.\n\nWho it is for: people working on near-field contrail models or planning hydrogen-contrail experiments. It deserves a serious referee. I would send it out and ask the referees to pin down the heterogeneous condensation prefactor, its numerical implementation, and exactly how the no-soot case was imposed.","headline":"Useful framework paper with an honest scope, but the heterogeneous condensation source term as written makes soot concentration irrelevant to condensation—a load-bearing bug that has to be fixed before the hydrogen comparison can be trusted.","tokens_in":19609,"tokens_out":5037,"would_cite":false,"duration_ms":51094,"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":"Non-equilibrium phase-change kinetics, resolved through a moment-based multiphase CFD model, shift and shrink predicted early contrails compared with equilibrium criteria.","keywords":["contrail formation","non-equilibrium condensation","homogeneous nucleation","heterogeneous nucleation","ice nucleation","method of moments","aircraft exhaust plume","RANS multiphase flow"],"falsifier":"Compare the predicted non-equilibrium condensation onset—its axial and radial position and peak droplet number—against test-rig or in-flight measurements in a well-characterized high-bypass plume: if droplets appear as early and as broadly as the equilibrium model predicts, the kinetic delay is an artifact; if they appear when and where the non-equilibrium model predicts, the equilibrium-based overprediction is confirmed.","tokens_in":18523,"feed_emoji":"✈️","tokens_out":5662,"duration_ms":56339,"temperature":0.7,"pith_summary":"The paper builds a numerical model of the first seconds of contrail formation that does not assume water instantly condenses and freezes once saturation is reached. Instead it resolves the finite rates at which vapor clusters nucleate into droplets, droplets freeze into ice, and both grow, using kinetic models adapted from steam-turbine and cloud-microphysics work. Run on a high-bypass turbofan nozzle at cruise conditions, the model predicts that equilibrium-style assumptions overstate how soon, how much, and how widely condensed and frozen water appears. The authors argue that capturing these kinetic delays changes the predicted shape of the young contrail and matters for assessing fuels and nozzle designs.","feed_headline":"Kinetics trim contrail predictions: equilibrium models overshoot","feed_subtitle":"A CFD study resolves nucleation rates and droplet growth to show where contrails really form.","key_machinery":"The carrying object is a coupled, moment-based population balance for each dispersed phase, embedded in a steady compressible RANS flow solver. For liquid droplets, ice crystals, and soot, the radius distribution is tracked through its first moments, with nucleation sources from classical nucleation theory (homogeneous and soot-assisted) and freezing from water-activity and active-site parameterizations, and with growth rates taken from a droplet-growth law including Knudsen and Prandtl corrections. A surface-area-weighted radius closes the growth term, and the moments feed back into the flow through interphase mass, momentum, and energy source terms. That coupling is what lets kinetics alter the flow and the flow alter kinetics, and it is what the equilibrium treatment omits.","core_discovery":"On its own terms, the paper's central discovery is that non-equilibrium phase-change kinetics are first-order for early contrail structure, not a small correction. In the unmixed nozzle case, the equilibrium treatment condenses immediately at saturation and produces a broad, continuous ice shell; the non-equilibrium treatment delays condensation, narrows and displaces the liquid region, and concentrates ice along the plume edges, with peak condensed mass substantially lower. The model attributes these differences to the nucleation energy barrier, finite growth rates, and separate momentum transport of vapor, liquid, and ice phases. It also reports a bimodal droplet spectrum—many tiny homogeneously nucleated droplets plus fewer larger soot-activated droplets—and shows that a hydrogen-like case with no soot can still produce ice through homogeneous nucleation. A mixed-flow nozzle delays and weakens condensation by homogenizing the exhaust.","pith_inferences":["Beyond the paper, the same kinetic core could be used as a screening tool: before full engine data exist, the model's homogeneous-nucleation result suggests soot-free contrails may still form and freeze early, shifting the question from whether they form to how many ice crystals they seed.","Beyond the paper, the bimodal droplet spectrum implies a freezing cascade in which larger heterogeneous droplets freeze first while small homogeneous droplets remain supercooled, so contrail ice crystal number may depend sensitively on the relative rates of the two nucleation pathways.","Beyond the paper, the model's sensitivity to the mixing closure suggests a direct computational test: replacing the steady turbulence closure with a turbulence-resolving simulation would show whether the computed nucleation timing and location survive a more faithful representation of entrainment.","Beyond the paper, the mixed-nozzle result reframes nozzle design as a possible contrail mitigation lever, since internal mixing can delay and weaken the early condensed phase before any atmospheric processes take over."],"forward_implications":["Equilibrium contrail criteria should be read as upper bounds on condensed mass and contrail extent, not as precise onset predictors.","Near-field contrail morphology is sensitive to nozzle mixing: unmixed exhaust produces earlier and stronger condensation than mixed exhaust.","Fuel-type effects enter through water vapor content and soot availability; a hydrogen-like high-water, soot-free case can still nucleate ice homogeneously.","Droplet size distributions at the contrail start are bimodal, so liquid-phase microphysics should be part of downstream contrail-evolution models.","The numerical framework is modular enough to swap fuels, nozzle geometries, and microphysical parameterizations without changing core assumptions."],"supporting_citations":[{"why":"Supplies the modern form of the equilibrium contrail onset criterion that the paper argues against.","marker":"[7]"},{"why":"Representative large-scale equilibrium model that assumes instantaneous condensation at saturation.","marker":"[2]"},{"why":"Provides the microphysical pathway context for soot and homogeneous nucleation in contrail formation.","marker":"[9]"},{"why":"Supplies the water vapor fraction and soot-free emission assumptions used for the hydrogen-like case.","marker":"[10]"},{"why":"Gives the droplet growth law used for condensation and ice growth.","marker":"[18]"},{"why":"Source for the steam-turbine nucleation and growth modeling adapted to the exhaust plume.","marker":"[19]"},{"why":"Provides the water-activity-based homogeneous freezing rate for supercooled droplets.","marker":"[34]"},{"why":"Supplies the active surface site density approach for immersion freezing.","marker":"[36]"},{"why":"Introduces the method of moments used to track polydispersed particle size distributions.","marker":"[38]"},{"why":"Provides the surface-area-weighted radius closure for the moment equations.","marker":"[40]"}],"fun_headline_variants":["Kinetics delay ice: equilibrium overpredicts contrail mass","Non-equilibrium kinetics reshape early contrail structure","Equilibrium models overshoot contrail ice: kinetics wins","Kinetic effects trim contrail ice predictions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the steady RANS turbulence model with a constant turbulent Schmidt number faithfully reproduces the entrainment and mixing that set local supersaturation, so that the computed nucleation timing and location are not artifacts of the mixing closure.","fun_headline_variants_meta":{"raw":{"variants":["Kinetics delay ice: equilibrium overpredicts contrail mass","Non-equilibrium kinetics reshape early contrail structure","Equilibrium models overshoot contrail ice: kinetics wins","Kinetic effects trim contrail ice predictions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000687,"raw_usage":{"total_tokens":3076,"prompt_tokens":865,"completion_tokens":2211,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":481,"completion_tokens_details":{"reasoning_tokens":2148}},"tokens_in":481,"tokens_out":2211,"duration_ms":15796,"temperature":1.0,"reasoning_tokens":2148,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:21:05.221030+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the predicted non-equilibrium condensation onset—its axial and radial position and peak droplet number—against test-rig or in-flight measurements in a well-characterized high-bypass plume: if droplets appear as early and as broadly as the equilibrium model predicts, the kinetic delay is an artifact; if they appear when and where the non-equilibrium model predicts, the equilibrium-based overprediction is confirmed.","supporting_citations":[{"cited_title":"On Conditions for Contrail Formation from Aircraft Exhausts,","cited_arxiv_id":null,"evidence_quote":"Supplies the modern form of the equilibrium contrail onset criterion that the paper argues against."},{"cited_title":"The microphysical pathway to contrail formation,","cited_arxiv_id":null,"evidence_quote":"Provides the microphysical pathway context for soot and homogeneous nucleation in contrail formation."},{"cited_title":"Contrail formation on ambient aerosol particles for aircraft with hydrogen combustion: a box model trajectory study,","cited_arxiv_id":null,"evidence_quote":"Supplies the water vapor fraction and soot-free emission assumptions used for the hydrogen-like case."},{"cited_title":"SpontaneousCondensationofSteaminSupersonicNozzles,","cited_arxiv_id":null,"evidence_quote":"Gives the droplet growth law used for condensation and ice growth."},{"cited_title":"Classical Nucleation Theory and Its Application to Condensing Steam Flow Calculations,","cited_arxiv_id":null,"evidence_quote":"Source for the steam-turbine nucleation and growth modeling adapted to the exhaust plume."},{"cited_title":"Water Activity as the determinant for homogeneous ice nucleation in aqueous solutions,","cited_arxiv_id":null,"evidence_quote":"Provides the water-activity-based homogeneous freezing rate for supercooled droplets."},{"cited_title":"A comparison of modelling methods for polydispersed wet-steam flow,","cited_arxiv_id":null,"evidence_quote":"Provides the surface-area-weighted radius closure for the moment equations."}],"review_version":1}