{"id":"86ed0c5a-31ed-4273-b514-f9e167cc6d52","arxiv_id":"2506.13169","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Hydrogen bond symmetrization in the C2 and C3 hydrogen hydrate phases occurs as a continuous crossover at lower pressures than in pure ice, driven by quantum fluctuations and guest-host coupling.","lead":"Researchers found that in two high-pressure forms of hydrogen hydrate (hydrogen stuffed ice), water's hydrogen bonds become symmetric at lower pressures than in pure ice. The change is a smooth crossover, not a sharp phase transition, and trapped hydrogen molecules plus quantum effects control it.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The predicted crossover pressures rely on a NequIP potential trained only to 80 GPa while the C3 static symmetrization is quoted at 110 GPa, with DFT benchmarks only at 20 and 30 GPa; direct DFT at the claimed crossover pressures is needed to support the quantitative central claims.","rationale":"The reader's conditional verdict is justified. I checked the full text and methods: the ML potential is the sole source of the pressure values cited in the abstract and conclusions, while the experimental Raman comparison is qualitative and terminates before the claimed crossover. The training-range and benchmark limitations are explicitly stated in the Appendix; this is not a hidden assumption but an unvalidated extrapolation that directly affects the main quantitative results. I therefore identify it as the most load-bearing concern. A second-order caveat is the proton-ordered model, which primarily affects the continuous-crossover characterization rather than the raw pressure numbers. Because the potential could still be accurate and the SSCHA methodology is established, the appropriate response is to keep the paper conditional and request direct DFT checks at the crossover pressures. The final verdict should remain CONDITIONAL.","tokens_in":11015,"tokens_out":12311,"duration_ms":136958,"concrete_test":"Use the same DFT functional, pseudopotentials, and supercell as the training set to perform static and SSCHA calculations for C3 at 60, 80, 100, 110, and 120 GPa, and for C2 at 20, 24, and 28 GPa. Compare the OH bond length, O-O distance, and proton potential profile against NequIP predictions. If the ML and DFT OH bond lengths differ by more than about 0.02 Å at 110 GPa, or if direct DFT places the static C3 symmetrization more than about 10 GPa from 110 GPa, the central quantitative claims and the universal O-O-distance picture are unsupported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"All of the paper's quantitative crossover pressures—C2 near 24 GPa, C3 static 110 GPa versus quantum 70 GPa, and the critical O-O distances of 2.41 and 2.39 Å—are outputs of the NequIP ML potential, not of direct DFT or of a unique experimental observable. As stated in the Numerical Methods appendix, the potential was trained on DFT data from 5–40 GPa (C2) and 40–80 GPa (C3), and the only explicit DFT benchmark reported is at 20 and 30 GPa. The static C3 symmetrization at ~110 GPa is therefore an extrapolation 30 GPa beyond the training window, and the quantum C3 value at 70 GPa sits at the upper edge of that window. If the potential's description of the OH potential-energy surface drifts with compression—for example, by slightly stabilizing the centered hydrogen position—the predicted crossover pressures and the 110→70 GPa quantum shift could be off by many GPa. The Raman data cannot independently pin the crossover because the hydrate OH mode is obscured above roughly 10–17 GPa, so the experiment validates only the pre-crossover redshift. In addition, the continuous-crossover characterization is obtained with a proton-ordered water sublattice; proton disorder is not tested and could alter the order parameter and the 'no overall symmetry change' conclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a combined experimental and computational study of hydrogen bond symmetrization in the filled-ice hydrogen hydrate phases C2 and C3. Using Raman spectroscopy and SSCHA simulations driven by a NequIP machine-learned potential, the authors find that the OH stretching mode in C2 softens and disappears, and that the quantum-renormalized OH bond lengths undergo a continuous crossover around 24 GPa in C2 and around 70 GPa (quantum) versus 110 GPa (static) in C3. They argue that symmetrization occurs without a change in crystal symmetry and that the critical O–O distance is approximately 2.41 Å for C2 and pure ice, but 2.39 Å for C3, reflecting suppression of transverse proton fluctuations by the dense H2 sublattice.","tokens_in":11287,"tokens_out":5911,"duration_ms":63202,"significance":"If the quantitative predictions are correct, the work significantly advances the field: it extends hydrogen-bond symmetrization studies to hydrogen-filled ices, demonstrates a large isotope/quantum effect, and identifies guest–host interactions as a control parameter that shifts the critical geometry. The combination of state-of-the-art anharmonic quantum simulations with experimental Raman data, including the isotopic comparison, is a strength. The C3 prediction (110→70 GPa quantum shift) and the critical O–O distance ordering are falsifiable. However, the ML-potential extrapolation beyond 80 GPa and the use of a proton-ordered water sublattice leave the quantitative central claims in need of further validation.","major_comments":[{"comment":"The NequIP potential is trained on DFT data from 5–40 GPa (C2) and 40–80 GPa (C3), with direct DFT benchmarks reported only at 20 and 30 GPa. The static C3 symmetrization pressure of ~110 GPa is therefore an extrapolation 30 GPa above the training range, and the quantum C3 result at ~70 GPa sits at the upper edge of that range. Since the central quantitative claims for C3 (110 GPa static, 70 GPa quantum, and the 2.39 Å critical O–O distance) are outputs of this potential, please add direct DFT validation at and around 70 and 110 GPa—for example, the OH bond length, the proton potential profile along the hydrogen bond, and the SSCHA renormalized phonons at those pressures—or provide a systematic uncertainty estimate for the ML potential at those compressions.","section":"Appendix: Numerical Methods"},{"comment":"The claim that symmetrization is a continuous crossover with no change in overall crystal symmetry is derived from a proton-ordered model of the water sublattice ('as in ice VIII and XI'). The actual C2 and C3 phases are based on an expanded ice Ic framework, whose proton disorder is not represented in the simulations. Proton disorder could alter the order parameter, broaden or smear the crossover, and affect the interpretation of 'no overall symmetry change.' Please test at least one disordered supercell or explicitly discuss the expected effect of proton disorder on the crossover character and on the symmetry statement in the abstract.","section":"II, paragraph beginning 'To shed light on the crossover character...'"},{"comment":"The experimental Raman data for C2 are confined to pressures below about 12–17 GPa, where the hydrate OH mode becomes obscured by the diamond signal and broadening; no experimental observable tracks the mode to zero frequency or confirms the crossover at ~24 GPa. The conclusion in Section III that 'Raman spectroscopy ... reveals a clear softening and eventual disappearance of the OH stretching mode, consistent with a continuous symmetrization process occurring around 24 GPa' overstates the experimental reach. Please rephrase to attribute the 24 GPa crossover and its continuity to the simulations, and to state explicitly that the experiment validates only the pre-crossover redshift.","section":"II, Fig. 3 and Conclusions"}],"minor_comments":[{"comment":"The caption states that the top and bottom right panels correspond to D2O-D2, while the main text says the top and bottom left panels are H2:D2O; please correct this inconsistency.","section":"Figure 1 caption"},{"comment":"The caption uses 'SCHA' but the method abbreviation used elsewhere is 'SSCHA'; please make this uniform.","section":"Figure 2 caption"},{"comment":"The word 'losange' appears in the caption; this should be 'diamonds' in English.","section":"Figure 3 caption"},{"comment":"Please define the plotted quantity explicitly, including the order parameter Δ and the meaning of the third-root-of-volume normalization; the current caption is difficult to parse.","section":"Figure 5 caption"},{"comment":"The phrase 'Raman spectroscopy ... reveals a clear softening and eventual disappearance' should be qualified, as the experimental Raman data do not extend through the disappearance; the simulations provide the disappearance signature.","section":"III, Conclusions"},{"comment":"Please add a sentence clarifying that the 270 cm⁻¹ rigid frequency shift is applied only to improve the frequency comparison and does not affect the predicted crossover pressure or the O–O distance analysis.","section":"III, text near Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a solid contribution, but the C3 predictions rest on an ML potential extrapolated 30 GPa beyond its training set. I would encourage the editor to request direct DFT checks at 70 and 110 GPa and a discussion of proton disorder; with those, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the good news. This is the first report of hydrogen-bond symmetrization in the C2 and C3 phases of hydrogen hydrate, and the qualitative physics is credible: the C2 phase symmetrizes continuously around 24 GPa, far below pure ice's ~55 GPa, and the C3 phase's delayed symmetrization (despite the same water framework) due to dense H2 guests is a genuinely new observation. The isotope shift between H2O and D2O is a nice fingerprint, and the Raman spectra, while limited, do show the expected pre-crossover softening.\n\nThe method is solid as far as it goes: SSCHA is the right tool for anharmonic quantum protons, and the NequIP potential is benchmarked at 20 and 30 GPa with reasonable errors. The paper is transparent about the rigid 270 cm-1 frequency shift and about the experimental obscuration above 10-17 GPa.\n\nNow the soft spots, in proportion. The load-bearing numbers—especially the C3 static symmetrization at 110 GPa—are extrapolations. The potential was trained on 5-80 GPa data, with the C3 branch covering 40-80 GPa. 110 GPa is 30 GPa outside that window, and the only direct DFT benchmarks quoted are at 20 and 30 GPa. If the potential's proton potential-energy surface drifts with compression, the quantitative crossover pressures and the 110→70 GPa quantum shift could be off by many GPa. The C2 result at ~24 GPa is safer, being inside the training range and consistent with the Raman trend, but the C3 numbers should be treated as tentative.\n\nThe second soft spot is the proton-ordered water sublattice. The continuous-crossover claim and the 'no overall symmetry change' conclusion are demonstrated only for this ordered model. Proton disorder is not tested, and it could alter the order parameter. The experiment, moreover, cannot track the OH mode to zero frequency, so the continuous character is not directly confirmed by data.\n\nNone of this sinks the qualitative central claim—that symmetrization in these hydrates is a smooth crossover occurring at lower pressure than pure ice, and that guest H2 density shifts the critical O-O distance in C3. But a serious referee should ask for DFT benchmarks at 70-110 GPa, or at least a softening of the quantitative language, before those numbers enter the literature.\n\nWho is this for? The high-pressure water/hydrogen and planetary-ice community. It deserves a serious referee and likely publication after revision. I'd bring it to reading group, but I wouldn't cite the C3 pressure numbers in my own work until the extrapolation is checked.","headline":"First look at hydrogen-bond symmetrization in hydrogen-filled ices, but the quantitative crossover pressures hinge on an ML potential extrapolation and a proton-ordered model.","tokens_in":11854,"tokens_out":3454,"would_cite":false,"duration_ms":34501,"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":"Hydrogen bonds in hydrogen-filled ice clathrates symmetrize at much lower pressure than in pure ice, through a continuous crossover.","keywords":["hydrogen bond symmetrization","filled ice","ice clathrates","hydrogen hydrate","quantum nuclear effects","SSCHA","Raman spectroscopy","high pressure"],"falsifier":"A pressure-quenched or in-situ neutron or X-ray diffraction experiment on the C2 phase between 20 and 30 GPa that locates hydrogen positions would show either a smooth increase of $\\Delta = d_{\\mathrm{OH}} - d_{\\mathrm{HO}}$ toward zero (confirming the crossover) or a jump-like discontinuity (refuting it). For C3, recomputing the static symmetrization with an interatomic potential explicitly benchmarked at 100–110 GPa would test whether the predicted 110 GPa value is physical or an extrapolation artifact.","tokens_in":10832,"feed_emoji":"🧊","tokens_out":6230,"duration_ms":58203,"temperature":0.7,"pith_summary":"The paper sets out to establish that hydrogen bond symmetrization—the pressure-driven moment when a water molecule's two O–H bonds become equal—happens differently in hydrogen-filled ice clathrates than in pure ice. Using Raman spectroscopy backed by quantum atomistic simulations, the authors argue that in the C2 and C3 filled-ice phases of hydrogen hydrate, symmetrization occurs as a continuous crossover with no change in crystal symmetry, and at markedly lower pressures than in pure ice: around 24 GPa in C2 versus about 55 GPa in ice VII/VIII, and around 70 GPa in C3 once quantum fluctuations are included. The central insight is that the dense H2 guest molecules and the quantum motion of protons together control the symmetrization geometry, shortening the critical O–O distance in C3 and even shaping the spectroscopic fingerprint. If true, this means hydrogen-rich planetary ice layers and any compressed water–hydrogen mixture cannot be described with pure-ice phase boundaries.","feed_headline":"Hydrogen-filled ice symmetrizes bonds at 24 GPa, vs 55 in pure ice","feed_subtitle":"Quantum proton motion and trapped H2 molecules lower the symmetrization pressure and turn it into a smooth crossover.","key_machinery":"The analysis rests on two tools. The Stochastic Self-Consistent Harmonic Approximation (SSCHA) is a method that replaces the quantum ionic wavefunction with an optimized Gaussian density matrix, producing anharmonic phonons and an average structure under pressure; it is what lets the authors include zero-point motion and anharmonicity, and it was previously used to describe symmetrization in pure ice. The interatomic forces come from an equivariant neural-network potential trained on DFT data. The central diagnostic quantity is the order parameter $\\Delta = d_{\\mathrm{OH}} - d_{\\mathrm{HO}}$, the difference between the covalent and hydrogen-bond O–H distances; plotting $\\Delta$ against pressure reveals the continuous character of the crossover, while the distance of the structure to half- and fully-symmetrized configurations confirms that the transition is gradual even in the proton-ordered model.","core_discovery":"In both the C2 and C3 phases of hydrogen hydrate, hydrogen bond symmetrization proceeds as a smooth, continuous crossover rather than a sharp phase transition, and it takes place without any change in the overall crystal symmetry. In C2 the crossover appears near 24 GPa—roughly half the pressure of the ice VII–X transition—while in C3 it appears near 70 GPa with quantum fluctuations (110 GPa in static calculations). The two phases share the same expanded ice-Ic water sublattice, yet they symmetrize at different critical O–O distances: about 2.41 Å in C2 and pure ice, but about 2.39 Å in C3, because the denser H2 population suppresses the transverse quantum fluctuations of the protons. The paper therefore claims that guest–host interaction and nuclear quantum effects, not just lattice geometry, determine where and how symmetrization occurs.","pith_inferences":["If the continuous-crossover picture holds, the notion of a sharp ice VII→X-type transition should be replaced by a crossover line in the C2/C3 region of the water–hydrogen phase diagram; the exact boundary may then depend on temperature and isotope in a way that a single transition pressure cannot capture.","A direct test would be a DFT calculation of the C3 static transition near 110 GPa, or an experimental compression of C3 beyond 90 GPa with a probe sensitive to proton positions, both of which would check whether the predicted quantum 70 GPa value is robust.","The same protocol could be applied to other filled ices, such as methane hydrates or salty ice, to ask whether guest-induced suppression of transverse proton fluctuations generalizes the C3 result.","A testable prediction is that the proton's inelastic neutron scattering spectrum in C3 should show reduced transverse librational amplitude compared to C2 at the same O–O distance."],"forward_implications":["In the C2 phase the OH stretching mode softens, broadens, and disappears near 24 GPa, giving a clear Raman signature of the crossover even though no ice-X-like T2g mode appears.","Deuterating the water framework shifts the C2 crossover up by about 5 GPa, showing that nuclear quantum motion lowers the symmetrization pressure.","In C3 the symmetrization-related stretching mode is so strongly mixed with low-frequency translational modes that it is spectroscopically silent; only simulations reveal the crossover near 70 GPa (quantum) or 110 GPa (static).","The critical O–O distance is not universal: C3 symmetrizes at 2.39 Å, about 0.02 Å shorter than C2 and pure ice, because dense H2 guests suppress transverse proton fluctuations.","Static calculations fail to symmetrize C2 at all in the studied range, implying that quantum fluctuations are essential for the low-pressure crossover."],"supporting_citations":[{"why":"provides the pure-ice quantum benchmark (ice VII-X, ~55 GPa) that the C2 and C3 results are compared against.","marker":"[5]"},{"why":"establishes the C3 phase and contributes DFT data used to train the interatomic potential.","marker":"[11]"},{"why":"characterizes the C2 filled ice and supplies DFT training data for the potential.","marker":"[12]"},{"why":"defines the SSCHA method used to include quantum and anharmonic effects.","marker":"[27]"},{"why":"supplies the equivariant neural-network interatomic potential that provides the forces.","marker":"[31]"},{"why":"justifies the 270 cm^-1 rigid shift applied to simulated OH/OD stretching frequencies.","marker":"[37]"},{"why":"provides neutron-diffraction evidence of hydrogen-bond symmetrization in pure D2O ice, the experimental baseline.","marker":"[4]"},{"why":"documents a universal critical O-O distance in hydrous minerals, the context that C3's shorter 2.39 Å value deviates from.","marker":"[10]"}],"fun_headline_variants":["Ice clathrate bond symmetrization: smooth crossover at 24 GPa","Quantum protons enable continuous ice bond symmetrization","Trapped H2 molecules alter symmetrization pressure in ice clathrates","Hydrogen hydrate shows continuous bond symmetrization without symmetry change","2.41 Å critical distance: ice clathrate symmetrization via quantum effects"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results assume the machine-learned potential trained on DFT data from 5–80 GPa stays accurate at the predicted crossover pressures, including the static C3 value of 110 GPa that lies outside the training range, and that the proton-ordered ice sublattice used in the simulations faithfully represents the real phases.","fun_headline_variants_meta":{"raw":{"variants":["Ice clathrate bond symmetrization: smooth crossover at 24 GPa","Quantum protons enable continuous ice bond symmetrization","Trapped H2 molecules alter symmetrization pressure in ice clathrates","Hydrogen hydrate shows continuous bond symmetrization without symmetry change","2.41 Å critical distance: ice clathrate symmetrization via quantum effects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000395,"raw_usage":{"total_tokens":2068,"prompt_tokens":938,"completion_tokens":1130,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":1035}},"tokens_in":554,"tokens_out":1130,"duration_ms":12389,"temperature":1.0,"reasoning_tokens":1035,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:37:03.887759+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A pressure-quenched or in-situ neutron or X-ray diffraction experiment on the C2 phase between 20 and 30 GPa that locates hydrogen positions would show either a smooth increase of $\\Delta = d_{\\mathrm{OH}} - d_{\\mathrm{HO}}$ toward zero (confirming the crossover) or a jump-like discontinuity (refuting it). For C3, recomputing the static symmetrization with an interatomic potential explicitly benchmarked at 100–110 GPa would test whether the predicted 110 GPa value is physical or an extrapolation artifact.","supporting_citations":[{"cited_title":"Quantum effects in H-bond symmetriza- tion and in thermodynamic properties of high pressure ice,","cited_arxiv_id":null,"evidence_quote":"provides the pure-ice quantum benchmark (ice VII-X, ~55 GPa) that the C2 and C3 results are compared against."},{"cited_title":"Observation of the most H2-dense filled ice under high pressure,","cited_arxiv_id":null,"evidence_quote":"establishes the C3 phase and contributes DFT data used to train the interatomic potential."},{"cited_title":"Giant splitting of the hydrogen rota- tional eigenenergies in the c2 filled ice,","cited_arxiv_id":null,"evidence_quote":"characterizes the C2 filled ice and supplies DFT training data for the potential."},{"cited_title":"The stochastic self-consistent harmonic approximation: calculating vibrational proper- ties of materials with full quantum and anharmonic ef- fects,","cited_arxiv_id":null,"evidence_quote":"defines the SSCHA method used to include quantum and anharmonic effects."},{"cited_title":"E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials,","cited_arxiv_id":null,"evidence_quote":"supplies the equivariant neural-network interatomic potential that provides the forces."},{"cited_title":"The mi- croscopic origin of the anomalous isotopic properties of ice relies on the strong quantum anharmonic regime of atomic vibration,","cited_arxiv_id":null,"evidence_quote":"justifies the 270 cm^-1 rigid shift applied to simulated OH/OD stretching frequencies."},{"cited_title":"Hydrogen bond sym- metrisation in D2O ice observed by neutron diffraction,","cited_arxiv_id":null,"evidence_quote":"provides neutron-diffraction evidence of hydrogen-bond symmetrization in pure D2O ice, the experimental baseline."},{"cited_title":"Structural independence of hydrogen- bond symmetrisation dynamics at extreme pressure con- ditions,","cited_arxiv_id":null,"evidence_quote":"documents a universal critical O-O distance in hydrous minerals, the context that C3's shorter 2.39 Å value deviates from."}],"review_version":1}