{"id":"09071483-042d-4976-93e6-cb44987ba3c4","arxiv_id":"2504.19528","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Molecular dynamics and thermodynamic modeling of BCC iron show that coexisting H and He form a core-shell structure in cavities, with He in the core and H on the surface, and that H trapping can persist above 600 K at high H concentration.","lead":"Simulations show that when hydrogen and helium both enter tiny cavities in iron, helium settles in the core while hydrogen forms a shell on the cavity surface, at temperatures from 300 K to 973 K. The result matters for fusion reactor steels, where hydrogen-helium synergy drives cavity swelling and hydrogen retention in structural materials.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The assembled Fe-H-He potential is not validated in the paper; the MD core-shell picture and H trapping thresholds depend on it.","rationale":"The reader's weakest-assumption analysis identifies the unvalidated assembled Fe-H-He potential as the key risk, and my review reaches the same conclusion. The paper's strongest claim is a kinetic MD statement about where H and He sit inside cavities; every MD figure supporting that statement is generated by a potential that is neither fitted transparently nor benchmarked against DFT or experiment in this manuscript. This is not a mere consensus disagreement: the literature example of the Ramasubramaniam Fe-H potential, cited by the authors themselves, shows that an empirical potential can produce qualitatively wrong H clustering behavior. Because the Wen EAM potential was developed for bulk Fe-H interactions and not specifically for cavity surfaces or H2 formation inside cavities, its use for the shell/H2 conclusions is a genuine correctness risk. I also note that the thermodynamic analysis in Section 4.2 uses DFT-derived binding energies from the authors' previous work, and therefore is not fully independent of the potential issue; if the MD potential is wrong, the claimed kinetic confirmation of the thermodynamic thresholds is unsupported, even if the thresholds themselves might still be correct. The concern is concrete and testable by direct comparison with a DFT-trained deep potential. Because the reader already assigned a conditional verdict based on this same issue, my analysis does not change the recommended outcome; it strengthens the reason for keeping the verdict conditional rather than accepting the paper as-is.","tokens_in":21102,"tokens_out":4142,"duration_ms":48256,"concrete_test":"Use the Wu deep potential (Ref. 30, trained on DFT) to evaluate, for the same 2.86 nm cavity with He/V = 0.5 at 723 K: (i) the binding energy of individual H atoms at the inner cavity surface, (ii) the H-H pair interaction as a function of separation inside the cavity, and (iii) a short cyclic-insertion trajectory, at least 2 ns, sufficient to measure the steady-state H radial fraction. Compare the H radial fraction and the number of H2 molecules with Figs. 4(e) and 5(d). If the radial profile changes qualitatively, or if any binding energy differs by more than 0.1 eV from the assembled empirical potential, re-fit the empirical potential and re-run the full 20 ns protocol. If the core-shell profile and H2 counts persist under the DFT-trained potential, the concern would be resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The weakest load-bearing point is the interatomic potential. Section 2.1 assembles the Gao Fe-He s-band potential, the Wen Fe-H EAM potential, the Aziz He-He potential, and the Hayward H-He potential, with unspecified fitting of the functional forms and parameters. No benchmark against DFT, the Wu deep potential, or experimental H/He trapping energies is reported in the manuscript. The Wen EAM potential was designed primarily for H in bulk Fe; nothing in the paper demonstrates that it transfers reliably to cavity surfaces, where H-H interactions and H2 formation are central to the claimed shell structure. Because the central kinetic statement—\"the cavity surface acts as the energetically favorable trapping site for H atoms, resulting in a stable shell-like structure\"—and the H2 molecule counts in Figs. 4-8 come directly from MD simulations using this potential, an overbinding of H to Fe surfaces or an incorrect H-H repulsion would directly generate the observed shell and H2 signals. The paper itself notes that the earlier Ramasubramaniam Fe-H potential produced artificial H clustering (Ref. 26), so this risk is concrete rather than hypothetical. The thermodynamic thresholds in Section 4.2 are computed from the authors' prior DFT binding energies rather than from this potential, so those thresholds are partially independent; however, the kinetic evidence for the shell and the claim that kinetics support the thermodynamics are not independent. Quantitative capture efficiencies and critical concentrations would shift if the potential were re-fitted or re-validated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper combines molecular dynamics (MD) simulations with a statistical thermodynamic model to study the evolution of H and He atoms trapped in pre-existing cavities in BCC Fe. The authors assemble a hybrid Fe-H-He interatomic potential from previously published Fe-He, Fe-H, He-He, and H-He potentials. MD simulations with cyclic insertion of H and He at a 4:1 ratio show that He forms a core-like distribution at the cavity center while H forms a shell near the cavity surface, with H2 molecules forming inside the cavity at higher temperatures. A thermodynamic model based on DFT binding energies from prior work predicts that H trapping by cavities is favorable at temperatures above 600 K provided the matrix H concentration exceeds a critical value (~10 appm at 723 K and ~100 appm at 973 K). The authors argue that these results explain experimental observations of core-shell H-He-V cavities in ion-irradiated RAFM steels.","tokens_in":21352,"tokens_out":7496,"duration_ms":76063,"significance":"If the assembled potential is quantitatively reliable, the study provides a plausible kinetic mechanism for the experimentally observed core-shell structure and yields a falsifiable thermodynamic condition for H retention at elevated temperatures. The thermodynamic analysis is based on independent DFT data rather than the MD potential, which is a strength. However, the central MD results are not validated against reference calculations or experiments, and all quantitative outcomes derive from single trajectories, so the significance is currently conditional on these caveats.","major_comments":[{"comment":"The assembled Fe-H-He hybrid potential is not validated anywhere in the manuscript. The text states that 'the final interatomic potential was derived by fitting the functional forms of these potentials along with their associated parameters' but provides no fitting procedure, no parameter values, and no benchmark against DFT (e.g., the Wu deep potential [30]), DFT-based binding energies, or experiments. The central kinetic claims—that H forms a stable shell on the cavity surface and that H2 molecules form in the cavity center (Secs. 3.1.2 and 3.2, Figs. 4-8)—depend directly on the accuracy of H-surface and H-H interactions in this potential. The authors themselves note that the earlier Ramasubramaniam Fe-H potential overestimated H-H attraction and produced artificial clustering (Ref. 26), demonstrating the sensitivity of such MD results to the potential. Without validation, the core-shell picture is not established.","section":"2.1"},{"comment":"All reported concentrations, capture efficiencies, and H2 counts (e.g., capture efficiencies in Figs. 2 and 4, H2 counts in Fig. 5(d), and CH values in Fig. 7) are obtained from single MD trajectories. No error bars, multiple independent runs, or statistical measures are reported. Given the stochastic insertion scheme and the finite system size (2.5x10^5 Fe atoms), quantitative trends such as the claim that 723 K maximizes H2 formation (Fig. 5(d)) may be sensitive to initial conditions and thermal noise. This undermines the quantitative nature of the kinetic evidence.","section":"3.1-3.2"},{"comment":"The thermodynamic model is based on binding energies from previous DFT studies [25,62], not from the MD potential used in the kinetic simulations. While this independence is a strength, the paper claims that the thermodynamic calculations 'provide substantial support' for the MD results, yet the two frameworks use different energy surfaces, so only qualitative agreement can be expected. More importantly, the thermodynamic model assumes a stable cavity with a constant He/V ratio (Eq. 8), whereas in the MD simulations the He/V ratio increases as He is trapped. The critical concentration thresholds (~10 appm at 723 K, ~100 appm at 973 K, Figs. 10-11) are therefore not directly comparable to the MD conditions, and their sensitivity to the assumed binding energies is not assessed.","section":"4.2, Eq. (8)"},{"comment":"The cyclic insertion scheme introduces H and He at extremely high rates (up to 16,000 appm H and 4,000 appm He over 20 ns). The authors acknowledge that the rates exceed experimental values, but the consequences for the observed H2 formation and shell structure are not discussed. It is plausible that the high insertion rate drives the system into an oversaturated regime where H2 formation is artificially enhanced. A discussion of this limitation, or tests at lower insertion rates, is needed.","section":"2.3"}],"minor_comments":[{"comment":"The phrase 'strong consistency with experimental results' is overly strong; the comparison with experiment is qualitative (Sec. 4), based mainly on the general core-shell morphology rather than quantitative distributions or concentrations.","section":"Abstract/Conclusion"},{"comment":"The manuscript should clarify whether each component potential is used with its original published parameters or whether any refitting was performed; as written, the description of the 'fitting' is opaque and prevents reproduction.","section":"2.1"},{"comment":"In the sentence beginning 'Clowers et al. [10] provided clear evidence...', the phrase 'at irradiation temperature' appears twice; this editorial error should be fixed.","section":"4.2"},{"comment":"The panel labels in Fig. 5 are not defined in the caption; the reader must infer that panels (a-c) correspond to 300 K, 723 K, and 973 K from the text. Please add explicit labels.","section":"Fig. 5"}],"recommendation":"major_revision","confidential_remarks":"The manuscript addresses an important and timely problem, and the thermodynamic model offers a testable prediction. However, the central MD claims rest on an unvalidated empirical potential, and the statistics come from single trajectories. I recommend major revision. The authors should validate the potential against DFT or the Wu deep potential they cite, report error bars from multiple runs, and clarify the mapping between the fixed-He/V thermodynamic model and the dynamically changing MD conditions. The claimed 'strong consistency' with experiment should be tempered unless a more quantitative comparison is provided."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one thing to know: the core-shell H/He picture is almost certainly right, but the quantitative trapping thresholds rest on a hybrid potential that the paper never validates. That split—robust qualitative result, shaky quantitative edifice—should guide how you read it.\n\nWhat's new: this is the first systematic MD survey of H/He co-evolution around pre-existing cavities in BCC-Fe across temperatures (300–973 K), cavity radii, and He/V ratios, and it turns the old 600 K de-trapping limit into a concentration-dependent threshold. That is a genuinely useful reframing. The paper also does something right that many MD papers skip: it uses vacuum surfaces so excess H/He can escape, and it checks for H2 via the radial PDF at 0.74 Å. The comparison to the Clowers halo-core EELS data is direct and appropriate.\n\nCredit where due: the cyclic-insertion protocol is described well enough to reproduce; the distinction between H on the surface and He in the core is consistent with earlier ground-state work (Hayward et al.), and the new piece—that kinetics at finite temperature produce the same pattern—is a legitimate addition. The thermodynamic argument in Sec. 4.2 is partly independent of the MD because it uses DFT binding energies from their previous papers, so the circularity concern is minor, not fatal.\n\nThe soft spots, in rough order of seriousness. First, the assembled Fe-H-He potential is the load-bearing element: it stitches together Gao Fe-He, Wen Fe-H, Aziz He-He, and Hayward H-He with 'fitting' that is never specified, and no benchmark against DFT or the Wu deep potential appears anywhere. Given that the paper itself notes the earlier Ramasubramaniam Fe-H potential produced artificial H clustering, the risk is concrete. If the Wen EAM overbinds H at cavity surfaces or gets H-H repulsion wrong, both the shell structure and the H2 counts could be artifacts. Second, every concentration and capture efficiency comes from single trajectories with no error bars; at 20 ns simulation time with 100 insertion cycles, sampling is thin. Third, no code or data are released, which matters more here than usual because the potential parameters are needed to assess the results.\n\nMy own verdict: conditional. I would not reject, and I would not accept as-is. Send it to review—the qualitative mechanism deserves an archival home—but the referee list should ask for a potential validation section, uncertainty estimates, and the potential file. If those come back solid, the concentration-dependent threshold claim becomes much more interesting.","headline":"The core-shell mechanism is likely robust, but the paper's quantitative claims rest on an unvalidated hybrid potential and single trajectories, so treat thresholds as provisional until the potential is benchmarked.","tokens_in":21921,"tokens_out":2432,"would_cite":false,"duration_ms":25667,"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":"When hydrogen and helium coexist in BCC iron, cavities evolve into a stable core-shell configuration—He at the center, H on the surface—and H trapping remains possible above 600 K once the matrix H concentration passes a critical threshold.","keywords":["H-He synergies","trapping effects","core-shell structure","molecular dynamics","thermodynamic theory","BCC iron","cavity evolution","fusion structural materials"],"falsifier":"One decisive check would be a direct comparison at 723 K and 973 K between this potential's predictions and a DFT-trained machine-learning potential for Fe-H-He: if a cavity with He/V = 0.5 does not retain an H surface shell above the critical concentration, or if the predicted critical H concentration shifts by more than an order of magnitude, the core-shell conclusion fails. Experimentally, one could irradiate a ferritic/martensitic steel with He and H so that the matrix H concentration at 723 K exceeds roughly 10 appm and look for an H halo around He-filled cavities by electron energy-loss spectroscopy; its absence would contradict the thermodynamic threshold.","tokens_in":20911,"feed_emoji":"🫧","tokens_out":9657,"duration_ms":82361,"temperature":0.7,"pith_summary":"The paper tries to establish that when hydrogen and helium are both present in body-centered cubic (BCC) iron—the situation in fusion-reactor steels—cavities do not just absorb gas randomly; they settle into a predictable core-shell configuration. Helium atoms sit stably at the cavity center, while hydrogen atoms prefer the cavity surface, and this geometry is produced by both molecular-dynamics kinetics and a Gibbs free-energy stability condition. A second claim is that the widely cited 600 K hydrogen de-trapping temperature is not an absolute limit: above 600 K, cavities still trap hydrogen as long as the matrix hydrogen concentration exceeds a critical value that grows with temperature. If these claims hold, fusion-relevant steels at high hydrogen production rates can retain hydrogen in cavities even at elevated irradiation temperatures, and the halo-core contrast seen in electron microscopy of irradiated steels has a concrete atomic mechanism.","feed_headline":"Cavity surfaces trap hydrogen above 600 K when H is plentiful","feed_subtitle":"MD plus thermodynamics shows H-He cavities form a stable core-shell structure above the old de-trapping temperature.","key_machinery":"The argument is carried by two coupled instruments. The first is a newly assembled three-species Fe-H-He empirical potential: an s-band many-body Fe-He description, an embedded-atom-method Fe-H potential that reproduces DFT-level H-H and H$_2$ behavior, a short-range repulsive H-He pair potential, and a compatible He-He pair form. It lets the authors run nanosecond-scale kinetic simulations with cyclic H/He insertion and a vacuum layer that allows atoms to escape, so trapping and de-trapping are both observable in one run. The second instrument is a statistical-thermodynamic stability condition, $\\Delta G(x,y,z) = -\\sum_l E_b^H(l,y,z) - x k_B T \\ln[C_H/(1-C_H)]$, built from H binding energies and the bulk H concentration: when the reversible work of adding one more H atom to the cavity is negative, trapping is spontaneous. Taken together, these convert a static ground-state picture of bubble energetics into a temperature- and concentration-dependent rule for cavity evolution.","core_discovery":"On the paper's own terms, the central discovery is that an H-He-vacancy cavity in BCC-Fe evolves into a stable core-shell structure: He is trapped at the cavity center because He-vacancy binding (about 2 eV) dominates, while H binds more weakly (about 0.75 eV) and therefore decorates the inner and outer cavity surfaces, which are low-electron-density regions. The molecular-dynamics simulations show this for temperatures 300–973 K, cavity radii 0.57–2.86 nm, and He/V ratios of 0, 0.5, and 0.8, with H$_2$ molecules forming in the cavity interior at intermediate temperatures and low He/V ratios. Statistical thermodynamics then supplies the quantitative condition: H trapping by a cavity is spontaneous, with $\\Delta G < 0$, when the matrix H concentration exceeds a critical threshold—about 10 atomic parts per million (appm) at 723 K and 100 appm at 973 K—even though both higher He/V ratio and higher temperature make trapping harder. This reconciles the old rule that H detraps above 600 K with experiments that observe H halos at 723 K, and it explains why the experimentally seen core-shell contrast is the expected steady state.","pith_inferences":["A testable extension: the same Gibbs criterion predicts that the critical He/V ratio for H trapping shifts upward as matrix H concentration increases, so a concentration-temperature phase diagram for the H shell could be drawn from this model.","The concentration-threshold logic implies that post-irradiation cooling history matters: H retained on cavity surfaces during irradiation can redistribute or de-trap as the matrix H concentration drops, so ex-situ measurements may undercount the H that was present in service.","If the core-shell structure is the long-time attractor, then cavity swelling models for ferritic/martensitic steels should treat the H shell as part of the cavity pressure balance, since H$_2$ in the interior and H adsorbed on surfaces contribute differently to stress.","The same hybrid potential and free-energy criterion could be applied to H/He behavior at grain boundaries, where the competition between He core trapping and H surface trapping may set incubation times for boundary bubble arrays."],"forward_implications":["Above 600 K, hydrogen retention by cavities becomes concentration-limited rather than absolutely forbidden; around 723 K a matrix H concentration of roughly 10 appm is enough to keep trapping thermodynamically spontaneous.","Triple-beam irradiation experiments at high H/He production rates should expect H-He-V cavities to display a He-rich core and an H-rich surface shell, with the H shell detectable at the cavity periphery.","Increasing the He/V ratio suppresses H$_2$ formation in the cavity interior, but the pressure-driven expansion of the cavity surface and the surrounding strain field add H-trapping sites, so hydrogen retention moves outward to the shell.","Interstitial H in the matrix promotes the formation of hydrogen-tagged helium-vacancy clusters outside cavities, creating competing traps and reducing H/He accumulation in pre-existing cavities.","An intermediate temperature near 723 K maximizes hydrogen storage in cavities because it balances thermal activation of H$_2$ formation against escape of interstitial H."],"supporting_citations":[{"why":"supplies the experimental EELS observation of an H halo around He-filled cavities at 723 K that the simulations are designed to explain.","marker":"[10]"},{"why":"gives ab initio evidence that vacancies screen the H-He repulsion, allowing H and He to co-segregate in a cavity.","marker":"[22]"},{"why":"first predicted a He-rich core/H-shell bubble configuration and supplies the H-He pair potential used in the model.","marker":"[23]"},{"why":"provides DFT H-He-vacancy binding energetics feeding the thermodynamic free-energy analysis.","marker":"[25]"},{"why":"supplies the Fe-H embedded-atom-method potential used here, chosen for its DFT-accurate H-H and H2 interactions.","marker":"[28]"},{"why":"establishes the ternary potential construction methodology that the present potential follows.","marker":"[32]"},{"why":"supplies the s-band Fe-He many-body potential used for helium interactions in the assembled ternary potential.","marker":"[34]"},{"why":"provides the nano-cavity hydrogen behavior regime model, including the critical H-H separation for H2 formation.","marker":"[57]"},{"why":"provides quantitative He-nanocavity energetics supporting the claim that He core trapping is more favorable than H shell trapping.","marker":"[62]"}],"fun_headline_variants":["H-He cavities form stable core-shell: He center, H surfaces","Critical H level flips H trapping in Fe cavities at 723 K","Core-shell He-H cavities: He locks in, H decorates surfaces","Fusion cavities: He cores, H shells stable above 600 K","Trapping H in Fe cavities needs >10 appm at 723 K"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the newly assembled Fe-H-He potential describes H/He/Fe interactions accurately at the high temperatures, high concentrations, and cavity surfaces studied here, even though the paper reports no direct benchmark of this combined potential against density-functional theory or experiments.","fun_headline_variants_meta":{"raw":{"variants":["H-He cavities form stable core-shell: He center, H surfaces","Critical H level flips H trapping in Fe cavities at 723 K","Core-shell He-H cavities: He locks in, H decorates surfaces","Fusion cavities: He cores, H shells stable above 600 K","Trapping H in Fe cavities needs >10 appm at 723 K"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000213,"raw_usage":{"total_tokens":1482,"prompt_tokens":1064,"completion_tokens":418,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":680,"completion_tokens_details":{"reasoning_tokens":322}},"tokens_in":680,"tokens_out":418,"duration_ms":4630,"temperature":1.0,"reasoning_tokens":322,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:50:16.420240+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"One decisive check would be a direct comparison at 723 K and 973 K between this potential's predictions and a DFT-trained machine-learning potential for Fe-H-He: if a cavity with He/V = 0.5 does not retain an H surface shell above the critical concentration, or if the predicted critical H concentration shifts by more than an order of magnitude, the core-shell conclusion fails. Experimentally, one could irradiate a ferritic/martensitic steel with He and H so that the matrix H concentration at 723 K exceeds roughly 10 appm and look for an H halo around He-filled cavities by electron energy-loss spectroscopy; its absence would contradict the thermodynamic threshold.","supporting_citations":[{"cited_title":"Clowers, Z","cited_arxiv_id":null,"evidence_quote":"supplies the experimental EELS observation of an H halo around He-filled cavities at 723 K that the simulations are designed to explain."},{"cited_title":"Ab initio study of helium and hydrogen interactions in $\\alpha$-Fe","cited_arxiv_id":"1205.6374","evidence_quote":"gives ab initio evidence that vacancies screen the H-He repulsion, allowing H and He to co-segregate in a cavity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides DFT H-He-vacancy binding energetics feeding the thermodynamic free-energy analysis."},{"cited_title":"Wen, A new interatomic potential describing Fe -H and H-H interactions in bcc iron, Comput","cited_arxiv_id":null,"evidence_quote":"supplies the Fe-H embedded-atom-method potential used here, chosen for its DFT-accurate H-H and H2 interactions."},{"cited_title":"Huang, Y.H","cited_arxiv_id":null,"evidence_quote":"establishes the ternary potential construction methodology that the present potential follows."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the s-band Fe-He many-body potential used for helium interactions in the assembled ternary potential."},{"cited_title":"Hayward, R","cited_arxiv_id":null,"evidence_quote":"provides the nano-cavity hydrogen behavior regime model, including the critical H-H separation for H2 formation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides quantitative He-nanocavity energetics supporting the claim that He core trapping is more favorable than H shell trapping."}],"review_version":1}