{"id":"838047f2-f147-4e94-bbc8-cb83a810fea8","arxiv_id":"2505.14171","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Nanoconfined water is shown to be a molecular superionic that conducts protons via chain-like Grotthuss hopping, enabled by short oxygen contacts and dangling hydrogen bonds.","lead":"Using machine-learned molecular dynamics and electronic structure analysis, the authors show that water confined in a nanoscale slit can conduct protons while remaining made of intact water molecules. This molecular superionic state works through a Grotthuss proton-hopping mechanism enabled by low transfer barriers and a flexible hydrogen-bond network, which suggests design rules for fast ionic conductors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper invokes the 0.1 S/cm superionic threshold but reports no conductivity for nanoconfined water; the classification and transferable design rules rest on an unevidenced quantitative claim.","rationale":"I read the paper as attempting two linked things: to classify nanoconfined water as a molecular superionic and to identify the mechanistic conditions enabling that classification. The molecular part is well supported: the narrow H-O-H angle distribution, sharp O-H RDF first peak, two-peak ICOBI distribution, and Voronoi coordination analysis all consistently indicate intact water molecules. The chain-like correlated diffusion is also supported by the geometric form of the chain-length distributions and by the sensitivity analysis in the SI. The weakest spot is the unquantified conductivity threshold. The word 'superionic' is defined by a numerical conductivity criterion in the paper itself, but no conductivity, diffusion coefficient converted to conductivity, or explicit reliance on a prior computed conductivity is provided. Defect mobility and correlated hops are necessary for superionic conduction but are not sufficient to establish that the system crosses 0.1 S/cm. The ML-potential concern raised by the reader is real but less decisive for the central classification: the nanoconfined potential was validated by Kapil et al., and the qualitative structural conclusions are robust to the level of theory, as the SI comparison for bulk superionic water suggests. A moderate shift in proton-transfer barriers would not erase the molecularity evidence, whereas even a perfectly accurate potential would not supply the missing conductivity number. Thus the most load-bearing failure is logical rather than numerical: the paper asserts a threshold-based classification without presenting the threshold-relevant observable. This does not overturn the qualitative mechanistic picture, so the reader's CONDITIONAL verdict remains appropriate. I would make the conductivity estimate or an explicit deferral to prior work a required condition for acceptance.","tokens_in":18622,"tokens_out":5776,"duration_ms":61906,"concrete_test":"Compute the ionic conductivity from the existing 500 K, 12 GPa nanoconfined trajectory, either via the Nernst-Einstein relation using proton/defect self-diffusion coefficients and the known carrier concentrations (pKw ≈ 1.8–3) or via Green-Kubo integration of the charge-current autocorrelation, and compare with 0.1 S/cm. If a conductivity value is not reported, the authors should instead explicitly cite and quote the value from the prior prediction (Kapil et al., Nature 609, 512) and state that the present conclusions are conditional on it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing gap is quantitative. The paper's own criterion for superionicity (Results: 'Materials have conventionally been considered superionic if they have a conductivity that exceeds 0.1 S/cm') is never evaluated for nanoconfined water. Defect MSDs show that hydronium and hydroxide species are mobile, and chain-length statistics show correlated hops, but neither establishes a conductance above the threshold. The Discussion asserts that 'all three' definitions of superionicity—high conductivity, chain-like diffusion, and highly diffusive defects—apply to nanoconfined superionic water, yet only the latter two are evidenced in the manuscript. If the phase's conductivity is below 0.1 S/cm, the central classification and the proposed design rules ('low PT barrier plus flexible H-bond network as key characteristics of fast ionic conduction in molecular superionics') are not supported as stated. This concern is independent of ML-potential accuracy: even granting revPBE0-D3 and the Morse-wall confinement model, the superionic label requires a number that the paper does not provide. Prior work (Kapil et al.) may supply that number, but the present paper neither quotes nor derives it, so the central claim overreaches the evidence presented here.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper combines machine-learned molecular dynamics, DFT/ICOBI bonding analysis, and trajectory-based correlation functions to argue that nanoconfined water at ~12 GPa and 400–550 K is both molecular and superionic. It contrasts this phase with bulk superionic water, ice VII, pressurized liquid water, and superionic AgI. The authors find that nanoconfined water retains intact molecules (narrow H-O-H angle distribution, molecular ICOBI signature), hosts diffusive hydroxide and hydronium defects, and exhibits geometric (chain-like) rather than Poisson hop statistics. They propose a two-step Grotthuss mechanism—proton transfer followed by hydrogen-bond network rearrangement—and identify low proton-transfer barriers and dangling-hydrogen-bond flexibility as the two key characteristics of molecular superionicity. The paper concludes that nanoconfined superionic water is a molecular superionic and that these criteria can guide discovery of other molecular superionic materials.","tokens_in":18817,"tokens_out":4238,"duration_ms":44892,"significance":"If fully established, the claim would be significant: it would identify a phase of pure water composed of intact molecules that conducts protons via a Grotthuss-like chain mechanism at conditions far milder than bulk superionic ice, and it would propose transferable design rules for molecular superionics. The paper has clear strengths: the nanoconfined simulations use a revPBE0-D3 neural network potential from prior work; the chain-length analysis is tested against several spatial and temporal cutoffs in the SI; a hybrid-functional cross-check for bulk superionic water is provided; and the comparison with AgI grounds the mechanism in the broader superionic literature. The structural, bonding, and dynamical analyses are internally consistent. The principal weakness is quantitative: the paper invokes the conventional 0.1 S/cm conductivity threshold but never computes or quotes a conductivity for nanoconfined water, leaving the central classification and the generality of the proposed design rules under-supported.","major_comments":[{"comment":"The paper sets the conventional superionic threshold as conductivity exceeding 0.1 S/cm, but nowhere in the main text or SI is a conductivity reported for nanoconfined water. The defect MSDs (Fig. 2b) and chain-length distributions (Fig. 2c) show mobile defects and correlated hops, but neither metric establishes that the conductivity exceeds 0.1 S/cm. This matters because the title, abstract, and Discussion depend on the classification 'superionic'. The later statement that 'all three' definitions of superionicity apply is therefore not supported for the conductivity definition. Please add an explicit conductivity estimate for the nanoconfined phase (e.g., a Nernst-Einstein estimate from defect mobility with a stated Haven-ratio assumption, or a Green-Kubo calculation), or alternatively cite and clearly use a previously computed value from Kapil et al. and state that present evidence concerns the mechanism rather than the conductivity number.","section":"Results and Discussion, paragraph beginning 'Materials have conventionally been considered superionic...'"},{"comment":"The paper proposes that 'low charge-carrier barrier' and 'a hydrogen bonding network that actively rearranges' are the two key characteristics of molecular superionicity, and then states that high conductivity, chain-like diffusion, and highly diffusive defects are 'transferable criteria of superionicity'. The evidence presented establishes these features for one nanoconfined water system and draws a qualitative analogy with ammonia. It does not test whether the two proposed characteristics are sufficient or necessary across a family of materials, nor does it show that the mechanism-specific criteria are equivalent to the conductivity threshold. The design-principle claims should be framed as hypotheses for future testing, backed by the ammonia comparison, rather than as established transferable criteria. This is a load-bearing framing issue because the paper's broader significance rests on the generality of these rules.","section":"Discussion and Conclusions, final paragraph"},{"comment":"The central mechanism depends on the accuracy of the revPBE0-D3 neural network potential and the implicit Morse-wall confinement model for nanoconfined water. The SI provides a hybrid-functional cross-check for bulk superionic water, but there is no equivalent validation of proton-transfer free-energy barriers or hydrogen-bond rearrangement dynamics for the nanoconfined system at 400–550 K and 12 GPa. Since the two-step Grotthuss mechanism rests precisely on these quantities, I ask the authors to either provide such a validation (e.g., selected DFT or alternative-potential checks of the free-energy surfaces in Fig. 3e) or explicitly discuss the expected sensitivity of the conclusions to these potential choices. This is a correctness-risk concern rather than a demonstrated error, but it is central to the mechanistic claim.","section":"Methods: Potentials and dynamics; SI Section S7"}],"minor_comments":[{"comment":"Equation (S2) appears to contain a typo: the second term inside 'min' should be ||r_i(t+Δt) − r_j(t)||, not ||r_j(t+Δt) − r_j(t)||.","section":"SI Section S8, Eq. (S2)"},{"comment":"The table lists 'Ice VII (50000 GPa)' and 'Pressurised water (500 GPa)', which are inconsistent with the main text values of 50 GPa and 0.5 GPa; the units or values should be corrected.","section":"SI Table S1"},{"comment":"The definition of δ uses square-root notation without explicit vector norms; please rewrite it with |r_H − r_O1| and |r_H − r_O2| to avoid ambiguity.","section":"Methods, Eq. (4)"},{"comment":"The caption uses the symbol ∂ for the proton-transfer coordinate in the free-energy panels, while Methods defines this coordinate as δ; please unify the notation.","section":"Figure 3 caption"},{"comment":"The reference list contains duplicate and misnumbered entries (e.g., two entries labelled '9', two labelled '1', two labelled '5', and similar issues for '10', '11', and '12'); the bibliography should be regenerated with unique numbers.","section":"References"},{"comment":"The text says simulations are performed 'with an anisotropic thermostat' for an NPT ensemble; this should presumably read 'anisotropic barostat'.","section":"SI Section S1C"}],"recommendation":"major_revision","confidential_remarks":"The missing conductivity is the main obstacle to publication. The structural and mechanistic analyses are careful and the SI contains useful sensitivity tests. If the authors can supply an explicit conductivity value or clearly anchor their classification to a quoted prior result, the central claim would be substantially strengthened. I do not see grounds for rejection, but the current overreach from mechanism to the 'superionic' label needs to be fixed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real content here is the mechanism, and the mechanism is credible. The paper shows that nanoconfined superionic water conducts via correlated, chain-like Grotthuss proton transfers through intact water molecules, and it isolates two enabling features: short O–O contacts that lower the proton-transfer barrier and dangling hydrogen bonds that let the network rearrange. The two-step diffusion model and the comparison against ice VII are the strongest parts—ice VII has proton transfer but a rigid H-bond network, which cleanly supports the argument that both ingredients are needed. The ICOBI bonding analysis and the chain-length statistics with sensitivity checks over cutoff and time window are careful. The SI hybrid-functional cross-check for bulk superionic water is a genuine effort, and the authors lean on previously validated ML potentials rather than fitting anything to their conclusion. No circularity problem; the design rules are inductive but honestly drawn from the simulation data and benchmarked against bulk ice and AgI.\n\nThe soft spot is exactly where the stress-test puts it: the 0.1 S/cm threshold they invoke is never evaluated. No conductivity is computed, quoted from prior work, or even estimated from the defect MSDs. That matters because the title and the Discussion make a classification claim, not just a mechanistic one. If the conductivity of this phase is below threshold, the label 'molecular superionic' overreaches, and the design rules lose some of their force. This is fixable—either compute conductivity at these state points or explicitly anchor the classification to a published number—but as written it is a gap in the central claim. The missing error bars on MSDs and free-energy surfaces are a smaller issue: they do not break the qualitative picture but do hurt confidence in the relative defect mobilities. The pressure labels are sloppy in places (Table S1 has '50000 GPa' and '500 GPa' where it clearly means 50 and 0.5), easily corrected but annoying. No data or code deposit is a real problem for a simulation paper of this type, especially with custom chain and correlation analysis.\n\nReliance on the revPBE0-D3 neural network and the implicit Morse-wall graphene model is a legitimate limitation, but I would not call it fatal. The potential and confinement model were validated in the prior Nature paper, and the physical picture is internally consistent. The missing conductivity number is the load-bearing issue.\n\nThis paper deserves a serious referee. Send it out, but flag the conductivity gap and the missing UQ in the review. As it stands, the mechanistic insight will survive; the classification claim needs a number.","headline":"Solid mechanistic paper on nanoconfined superionic water; the 'superionic' label needs a conductivity number it never produces.","tokens_in":19438,"tokens_out":1788,"would_cite":true,"duration_ms":63619,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-07T15:39:02.761124+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}