{"id":"07ac0e4b-31a2-43f5-92e9-5d864abd9c68","arxiv_id":"2411.16474","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"FESTIM v1.3's kinetic surface model is verified with the method of manufactured solutions and validated against hydrogen retention experiments and MHIMS/TESSIM-X simulations.","lead":"The paper adds a kinetic surface model to FESTIM, an open-source code for hydrogen transport in materials, and tests it with manufactured solutions, four experiments, and two other simulation codes. It matters because it extends an open tool used in fusion research to regimes where surface and bulk hydrogen are not in equilibrium.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Table 4 gives ΓD=5.8×10^18 m^-2 s^-1 for EUROFER Case 4 while the text and Fig. 8 state ~9×10^19; the validation outcome depends on this flux, so the paper's stated inputs are internally inconsistent.","rationale":"The paper's central claim is that FESTIM v1.3's kinetic surface model is correctly implemented and demonstrates reliability by reproducing four experimental hydrogen retention cases. In good faith, the MMS test and the cross-code comparisons do provide useful evidence that the implementation is internally consistent and matches other codes. However, a verification and validation paper must report inputs that are complete and internally consistent, since the experimental agreement is the main external support for the model. The Case 4 discrepancy between Table 4 and the text/Fig. 8 is directly load-bearing: ΓD controls adsorption, abstraction, and implantation, so a factor-of-15 ambiguity changes the computed TDS spectra and makes the claimed agreement with experiment unreproducible from the paper as written. The reader's weakest assumption concerned parameters fitted to the same experiments; this is a related concern about parameter integrity, but the specific numerical inconsistency is more concrete and more easily resolved, hence 'partial' agreement. If the repository script resolves the discrepancy, the paper still needs an erratum or explicit clarification; if not, the validation claim is significantly weakened. The MMS section would also benefit from an explicit convergence-rate table, but the flux inconsistency is the strongest single concern. The appropriate final verdict remains conditional: the paper should be accepted only after the Case 4 flux value is reconciled and the validation inputs are made reproducible.","tokens_in":14994,"tokens_out":7930,"duration_ms":72505,"concrete_test":"Download the archived repository (Zenodo doi:10.5281/zenodo.14217067), inspect the Case 4 input script (e.g., the D flux and temperature time profiles), and compare its ΓD value and time profile with Table 4 and Fig. 8. Then run Case 4 twice, once with ΓD=5.8×10^18 and once with ΓD=9×10^19 (keeping all other inputs fixed), and identify which output matches the FESTIM curves in Fig. 9. Also check whether the listed gas pressure of 1 Pa and molecular flux formula (eq. 10) are used consistently. This single check determines whether the inconsistency is a typographical error or a substantive error in the stated simulation setup.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The most load-bearing weakness is an internal inconsistency in the key input for validation Case 4 (Section 4.4). The text says samples were irradiated with a low-energy D flux of ≈9×10^19 m^-2 s^-1 and Fig. 8's y-axis is labeled in units of 10^19 m^-2 s^-1 with values reaching roughly 9. However, Table 4 lists ΓD = 5.8×10^18 m^-2 s^-1, a factor ~15 lower. The surface flux terms (Jads ∝ ΓD2, Jloss ∝ ΓD(1-r)θ, and the implantation source S) all scale with this flux, so the difference changes the predicted retention and TDS spectra. Because the abstract's central claim rests on reproducing the four experimental cases, this parameter must be correct and reproducible from the paper. If Table 4 is the actual value used, then the text/Fig. 8 are wrong and the experiment is misrepresented; if the text is right, Table 4 cannot reproduce the shown FESTIM curves. This is not a question of physics consensus but of internal consistency: the reader cannot tell from the paper which flux was simulated. The paper's own note that some parameters 'differ between the published paper and the input files of TESSIM-X' and that others came from private communication makes the omission more consequential. A secondary concern is the MMS section (Section 3) claims convergence only verbally with no convergence study; however, the flux inconsistency is the primary risk to the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes the implementation of a kinetic surface model in the open-source hydrogen transport code FESTIM v1.3, in which the surface concentration of adsorbed hydrogen is treated as an independent field evolved by a flux balance equation coupled to bulk diffusion-trapping transport. The implementation is checked with a method-of-manufactured-solutions (MMS) test, and then applied to reproduce four experimental datasets on hydrogen isotope retention in Ti, oxidized W, self-damaged W, and EUROFER, including comparisons with the MHIMS and TESSIM-X codes. The authors claim that the MMS test proves correctness of the implementation and that the four validation cases demonstrate reliability of the model.","tokens_in":15384,"tokens_out":2069,"duration_ms":21170,"significance":"If the implementation and validation are sound, the paper is valuable: FESTIM is a widely used open-source package, and a documented, flexible kinetic surface model with a public validation script repository [27] would be a useful community resource. The MMS test is a standard and appropriate verification tool, and the cross-code comparisons with MHIMS and TESSIM-X provide a meaningful consistency check. However, the validation strength is limited by the reuse of parameters fitted to the same or closely related experiments in earlier work, by the absence of quantitative error measures and experimental uncertainty bars, and by an internal inconsistency in the key flux input of Case 4 that must be resolved before the central claim can be accepted.","major_comments":[{"comment":"The text states that EUROFER samples were irradiated with a low-energy D flux of ≈9×10^19 m^-2 s^-1, and Fig. 8 shows flux values around 9 on an axis labeled in units of 10^19 m^-2 s^-1, but Table 4 lists ΓD = 5.8×10^18 m^-2 s^-1, a factor of about 15 lower. Since the surface fluxes Jads (∝ ΓD), Jloss (∝ ΓD(1-r)θ), and the implantation source S all scale with this flux, the two values lead to quantitatively different predicted retention and TDS spectra. The reader cannot determine from the paper which value was actually used in the FESTIM simulations. This is a load-bearing inconsistency for the fourth validation case and for the paper's overall claim of reproducing the experimental cases; it must be corrected and clarified.","section":"Section 4.4, Table 4 and Fig. 8"},{"comment":"The paper claims that 'these errors decrease with decreasing stepsizes' but reports only single L2 error values (Ebulk = 2.33×10^-5, Esurf = 4.29×10^-5) for one mesh size and one time step, with no convergence table or convergence-rate analysis. A verification claim of 'correctness' in a V&V paper should include a systematic refinement study showing the expected order of accuracy. Please add a grid/time-step convergence table or at least a log-log plot of error versus step size.","section":"Section 3, MMS verification"},{"comment":"The validation cases inherit many parameters that were fitted to the same or closely related experimental data in prior studies: the coverage-dependent desorption parameters in Table 2 are from Hodille et al. [5], the trap concentrations and detrapping energies in Table 3 are from [45], and Case 4 relies on parameters from [46] plus additional values from private communication. Consequently, the agreement with experiment is largely a consistency check that FESTIM solves the same equations as the earlier codes, not an independent confirmation of the underlying physics. The paper should state this limitation explicitly and temper the claim that the model's 'reliability' is demonstrated by these cases.","section":"Section 4.2, Section 4.3, Section 4.4"}],"minor_comments":[{"comment":"The sentence 'The front surface (at x = 0) is assumed...' and the following paragraph contain 'ad hocloss' which should be 'ad hoc loss channel'.","section":"Section 4.4, text after Eq. (24)"},{"comment":"The symbol for the sample cross-section area appears to be missing in the text; it is denoted only by '= 1.3 × 10^-4 m^2' after Eq. (12). Please define the symbol explicitly.","section":"Section 4.1, Eq. (12)"},{"comment":"In the sentence before Fig. 9, '(see Fig. 9.' is missing a closing parenthesis; it should read '(see Fig. 9).'","section":"Figure 9"},{"comment":"The table and text state that the D flux is 5.8×10^18 m^-2 s^-1, which is consistent with the cited experiment, but no experimental uncertainty is quoted for the NRA data or the flux; adding error bars or a stated uncertainty would strengthen the comparison.","section":"Section 4.3, Table 3"},{"comment":"The paper would benefit from a reproducibility statement explicitly listing the version of FESTIM, the FESTIM V&V book reference, and the exact contents of the Zenodo repository [27], since the repository link is central to the verification and validation claims.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The flux inconsistency in Case 4 is the main technical blocker; it is a straightforward factual issue that the authors should resolve unambiguously. The MMS section also needs a convergence study to support the verification claim. The validation is weaker than the abstract suggests because of inherited fitted parameters and private-communication inputs; the authors should be encouraged to frame the experimental comparisons as consistency checks rather than independent validation. The paper fits the scope of the journal and, after these revisions, could be acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper delivers what it says: a flexible kinetic surface model in FESTIM v1.3, verified by MMS and tested against four experimental cases plus cross-code comparisons. The MMS test is clean and gives real evidence the implementation is correct, though the claim that errors decrease with step size would be stronger with a convergence table. The arbitrary Jvs interface is a genuinely useful engineering feature, and the open-source scripts make the work reproducible in principle. The validation cases cover useful ground—Ti absorption, oxidised W, self-damaged W, and EUROFER—and the agreement with MHIMS and TESSIM-X is meaningful because it shows the model reproduces the same equations as established codes.\n\nThe load-bearing problem is in Section 4.4. The text and Fig. 8 say the EUROFER samples were exposed to a D flux of ≈9×10^19 m^-2 s^-1, but Table 4 lists ΓD = 5.8×10^18 m^-2 s^-1—a factor of ~15 lower. This is not a minor typo: the adsorption, loss, and implantation terms all scale with this flux, so the predicted TDS spectra depend on which value was actually used. The reader cannot tell from the paper which simulation was run. That directly undermines the reproducibility of the central validation claim. The paper's own note that some parameters came from private communication or differ from the TESSIM-X input files makes this omission worse.\n\nOther soft spots are minor by comparison. The validation reuses fitted parameters from prior work, so the experiments are consistency checks rather than independent confirmations of the physics—but the paper does not overclaim otherwise. The MMS section is adequate but would benefit from a formal convergence study.\n\nOverall, the implementation and verification are solid, and the cross-code agreement is encouraging. The flux inconsistency is fixable with a corrected table or text, but it must be resolved before this can be trusted as a validation paper. A serious referee should see it, with the expectation of a revision.\n\nRecommendation: send to peer review, but flag the Table 4 flux discrepancy as a major required fix.","headline":"A useful open-source V&V contribution with a genuine internal inconsistency in the key flux for one validation case; fixable but needs to be corrected before publication.","tokens_in":15882,"tokens_out":1335,"would_cite":true,"duration_ms":13780,"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":"FESTIM's new kinetic surface model is verified by manufactured solutions and validated against four hydrogen-retention experiments in Ti, W, and EUROFER.","keywords":["kinetic surface model","FESTIM","hydrogen transport","method of manufactured solutions","validation","hydrogen isotope retention","thermal desorption spectroscopy","finite element method"],"falsifier":"Run the same four cases with a manufactured solution whose L2 errors do not decrease as the mesh and time step are refined, or reproduce one of the experiments using independently measured sticking and desorption parameters not fitted to that experiment; either failure would show that the model or its validation is not conclusive.","tokens_in":14819,"feed_emoji":"🧪","tokens_out":5142,"duration_ms":46959,"temperature":0.7,"pith_summary":"This paper reports that FESTIM v1.3's newly added kinetic surface model is correctly implemented and usable for hydrogen transport simulations. It verifies the numerical implementation with the method of manufactured solutions on a coupled bulk-surface diffusion problem, obtaining small L2 errors that decrease with refinement. It then claims reliability by reproducing four published experiments on hydrogen isotope uptake and release in titanium, tungsten, and EUROFER, using surface flux expressions taken from earlier models. Cross-code comparisons with MHIMS and TESSIM-X match essentially curve-for-curve, which the authors present as evidence that the new model solves the same physics as established packages.","feed_headline":"FESTIM kinetic surface model matches four H-retention experiments","feed_subtitle":"MMS verification plus Ti, W, and EUROFER experiments back the v1.3 surface-kinetics feature.","key_machinery":"The load-bearing object is the kinetic surface model itself: a surface concentration $c_s$ (particles per area) evolved by balancing the subsurface-to-surface flux $J_{bs}$, the surface-to-subsurface flux $J_{sb}$, and an arbitrary user-supplied net vacuum-to-surface flux $J_{vs}$, with the bulk connected through the Robin condition $\\mathbf{J}\\cdot\\mathbf{n} = \\lambda_{IS}\\,\\partial c_m/\\partial t + J_{bs} - J_{sb}$. The model generalizes the earlier concentration-only surface treatment, which assumed adsorbed and absorbed populations are in equilibrium; the new variable allows surface kinetics to be rate-limiting, as in low-energy atom exposure or low-temperature sorption. The machinery is completed by Arrhenius frequency factors and site-blocking factors $(1 - c_s/n_{surf})$.","core_discovery":"The central claim is that the kinetic surface model — which tracks an explicit adsorbed surface concentration $c_s$ governed by $dc_s/dt = J_{bs} - J_{sb} + J_{vs}$, coupled to the bulk McNabb–Foster diffusion-trapping equations through a Robin condition — is a correct and reliable extension of FESTIM. Correctness is established by manufactured-solution convergence; reliability is established by agreement with four experiments spanning different materials (Ti absorption at 473–923 K, D desorption from oxygen-covered W(110), NRA-measured D retention in self-damaged W, and TDS from damaged EUROFER) and with two independent codes.","pith_inferences":["A consequence the paper leaves implicit: the experiments constrain only a few surface parameters; many come from fits to the same or closely related measurements, so the validation primarily shows that FESTIM reproduces the equations of prior codes rather than independently confirming the surface physics.","A stricter test would use the kinetic surface model with parameters measured from independent surface-science experiments (for example, desorption energies from density functional theory and sticking coefficients from beam experiments) and predict a retention experiment without refitting.","The steady-state reduction the authors mention could be formalized by taking $dc_s/dt \\approx 0$ and eliminating $c_s$, yielding an effective surface boundary condition; a useful extension would be to quantify when this approximation breaks down, such as during fast transients or large $J_{vs}$.","Extending the model beyond 1D would let surface diffusion couple adjacent faces; the current 1D restriction means predictions for geometrically complex components still rely on the equilibrium surface assumption."],"forward_implications":["FESTIM users can model surface-limited uptake and release processes, such as low-energy atom exposure, sorption experiments, and transient plasma events, with an explicit adsorbed hydrogen concentration.","The manufactured-solution test provides a benchmark problem for the kinetic surface model that other hydrogen transport codes can reproduce.","Because FESTIM matches MHIMS and TESSIM-X on the same inputs, the four validation cases can serve as a shared reference set for cross-code comparisons.","The near-instant surface equilibration seen in the self-damaged tungsten case justifies a steady-state approximation of surface kinetics for long-time simulations.","The validation cases will be added to FESTIM's verification and validation book, giving users documented confidence in the new feature."],"supporting_citations":[{"why":"Supplies the method of manufactured solutions used to verify the numerical implementation.","marker":"[32]"},{"why":"Provides the Pick-Sonnenberg surface flux balance that the kinetic model generalizes.","marker":"[29]"},{"why":"Provides the surface flux expressions and parameters used for the titanium absorption validation case.","marker":"[36]"},{"why":"Supplies the coverage-dependent desorption model and MHIMS results used for the oxygen-covered tungsten validation case.","marker":"[5]"},{"why":"Supplies the surface process model for atomic deuterium exposure in self-damaged tungsten.","marker":"[31]"},{"why":"Provides trap parameters and the observation of rapid surface equilibration used in the self-damaged tungsten case.","marker":"[45]"},{"why":"Supplies the EUROFER model setup, parameters, and TESSIM-X comparison for the final validation case.","marker":"[46]"},{"why":"Establishes the FESTIM code framework that the kinetic surface model extends.","marker":"[12]"}],"fun_headline_variants":["FESTIM kinetic surface model verified and validated","FESTIM v1.3 surface kinetics: four experiments agree","Kinetic surface model in FESTIM matches four experiments","FESTIM surface model: MMS verification plus 4 experiments","FESTIM surface kinetics: verified by MMS, matched by 4 experiments"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The validation assumes that the surface flux expressions and parameter values taken from earlier models are the correct physics for these experiments, even though several of those parameters were fitted to the same or closely related data.","fun_headline_variants_meta":{"raw":{"variants":["FESTIM kinetic surface model verified and validated","FESTIM v1.3 surface kinetics: four experiments agree","Kinetic surface model in FESTIM matches four experiments","FESTIM surface model: MMS verification plus 4 experiments","FESTIM surface kinetics: verified by MMS, matched by 4 experiments"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000592,"raw_usage":{"total_tokens":2710,"prompt_tokens":816,"completion_tokens":1894,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":432,"completion_tokens_details":{"reasoning_tokens":1806}},"tokens_in":432,"tokens_out":1894,"duration_ms":14123,"temperature":1.0,"reasoning_tokens":1806,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:07:25.965812+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same four cases with a manufactured solution whose L2 errors do not decrease as the mesh and time step are refined, or reproduce one of the experiments using independently measured sticking and desorption parameters not fitted to that experiment; either failure would show that the model or its validation is not conclusive.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Pick-Sonnenberg surface flux balance that the kinetic model generalizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the surface process model for atomic deuterium exposure in self-damaged tungsten."},{"cited_title":"Delaporte-Mathurin, J","cited_arxiv_id":null,"evidence_quote":"Establishes the FESTIM code framework that the kinetic surface model extends."}],"review_version":1}