{"id":"be8413d0-8570-4917-84a4-843ab6de55e0","arxiv_id":"2607.28244","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Negative thermal expansion in ice I is a collective nuclear-quantum effect of the open tetrahedral H-bond network, not of hexagonal stacking, with a density maximum near 70 K in pure cubic ice Ic.","lead":"Stacking-disorder-free cubic ice shows the same density maximum near 70 K as hexagonal ice, proving negative thermal expansion is a property of the shared tetrahedral hydrogen-bond network. Path-integral simulations reproduce it only with nuclear quantum effects, tying the anomaly to anisotropic proton delocalization and low-frequency transverse modes.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"Embedding proves 0 K volume renormalization needs a quantum environment, but that is not the same as proving NTE itself is collective.","rationale":"The reader correctly flagged the embedding as the soft joint in the strongest claim. Experiment (pure Ic density max ≈ Ih), MB-pol classical-vs-PIMD contrast, and negative-γ transverse modes under quantum statistics are mutually consistent and do not depend on Fig. 4. What depends on Fig. 4 is the sharper assertion that NTE is collective rather than local O–H physics. That assertion mixes a static 0 K volume argument (embedding, coupling omitted) with a finite-T mechanism already explained by QHA/Bose population of network modes. Tightening or replacing that step—e.g. with the local-V_OO vs TA/TO QHA test above—or softening the abstract/conclusion wording from ‘establish … collective’ to ‘consistent with a network-scale quantum-statistical origin’ is the right condition for acceptance. No independent reason to move to REJECT or to full ACCEPT; verdict stays CONDITIONAL, confidence in that call unchanged. Novelty and low circularity assessments stand.","tokens_in":19012,"tokens_out":759,"duration_ms":33756,"concrete_test":"Build a local reference free energy: use the ZPE-corrected V_OO(d_OO) curves from Fig. 4(e) (classical and quantum embeddings) as the sole volume-dependent potential, with no network phonon spectrum, and compute α_V(T) from that 1D effective potential. Separately recompute α_V(T) from the existing MB-pol QHA sum restricted to TA+low-TO branches only. If the local V_OO model produces no density maximum near 70 K while the TA/TO QHA sum does, the collectivity claim for NTE stands without the embedding; if the local model already yields a ~70 K maximum, Fig. 4’s leap from 0 K density to NTE collectivity is unsupported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central wording—that NTE is a collective quantum effect rather than a local O–H zero-point correction—rests on the single-proton embedding of Fig. 4 and the surrounding text (§ after Fig. 3). That construction shows: (i) classical environment + quantum proton shifts 0 K density the wrong way via ZPE(d_OO); (ii) quantum environment recovers a density nearer PIMD/experiment. The paper then concludes collectivity for NTE. Two gaps make this load-bearing step insecure. First, the authors explicitly neglect coupling between the embedded proton and the surroundings, so the classical-vs-quantum contrast is not a controlled isolation of local vs network physics. Second, and more important, the embedding constrains the static 0 K minimum of V_OO, whereas NTE is the finite-T slope of density below ~70 K. That slope is already largely recovered by QHA on network phonons (Pamuk comparison; mode-resolved γ of TA/TO under Bose weighting). A correct 0 K volume from quantum embedding does not, by itself, establish that the thermal contraction is collective rather than a local anharmonic or ZPE effect. The multi-messenger pieces (Ic≈Ih, PIMD vs classical, ADP/gyration anisotropy, negative-γ TA/TO) still support network + nuclear statistics; the specific local-vs-collective proof offered for NTE overreaches what Fig. 4 shows.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports neutron powder diffraction on stacking-disorder-free cubic ice Ic (from topotactic degassing of C2 hydrogen hydrate) together with classical MD and PIMD simulations using MB-pol. Experimentally, Ic shows a density maximum near 70 K that closely matches hexagonal ice Ih. Classical MD yields no density maximum, while PIMD quantitatively reproduces the anomaly (max relative error ~0.2%). The density maximum coincides, within the 15 K grid, with a maximum in the transverse-over-longitudinal proton ring-polymer anisotropy; neutron ADPs independently show enhanced transverse proton displacement. Phonon calculations identify low-frequency TA/TO branches with the most negative Grüneisen parameters, and mode-resolved QHA analysis attributes the cryogenic contraction to Bose-weighted population of those branches. A single-proton embedding construction (classical vs quantum environment PES) is used to argue that the effect is collective rather than a local O–H zero-point correction. The authors conclude that NTE in ice I is an intrinsic, collective nuclear-quantum response of the open tetrahedral hydrogen-bond network, not specific to hexagonal stacking.","tokens_in":19352,"tokens_out":1749,"duration_ms":38573,"significance":"If the result holds, this is a clean and timely contribution: stacking-disorder-free Ic finally allows a controlled comparison of ice-I polytypes that share local tetrahedral coordination but differ in long-range stacking. Demonstrating that the density maximum is essentially the same in Ic and Ih, and that MB-pol PIMD (but not classical MD with the same PES) recovers it quantitatively, substantially strengthens the case that ice-I NTE is a nuclear-quantum network phenomenon. The combination of neutron lattice parameters, ADPs, gyration-radius anisotropy, mode-resolved Grüneisen parameters, and classical-vs-PIMD contrast is a strong multi-messenger package. The work also supplies a chemically transparent reference for quantum-driven anomalies in other open tetrahedral networks. Credit is due for the quantitative MB-pol agreement, the isolation of nuclear statistics via identical PES, and the mode-resolved QHA breakdown in the SM.","major_comments":[{"comment":"The title/abstract claim that NTE is a 'collective quantum effect' rather than a local O–H ZPE correction rests heavily on the single-proton embedding of Fig. 4 and the following paragraphs. That construction shows that a quantum proton in a classical environment shifts the 0 K V_OO minimum the wrong way (higher density via ZPE(d_OO)), whereas a quantum environment recovers a density nearer PIMD/experiment. Two gaps make this an insecure proof of collectivity for NTE itself. (i) The authors explicitly neglect coupling between the embedded proton and the surroundings, so the contrast is not a controlled isolation of local vs network physics. (ii) More importantly, the embedding constrains the static 0 K minimum of V_OO, whereas NTE is the finite-T slope of density below ~70 K. That slope is already largely recovered by QHA on network phonons (Pamuk comparison; SM mode-resolved γ of TA/TO","section":"Fig. 4 and text after Fig. 3"},{"comment":"The claimed 'direct link' between maximal proton-path anisotropy and the density maximum is stated to hold 'within the temperature resolution.' Both quantities are sampled on a coarse grid (experiment ~15 K; PIMD points similarly sparse), and R_g,z/R_g,x shows a broad maximum rather than a sharp peak. Coincidence on this grid is suggestive but not decisive. Either denser temperature sampling of the anisotropy ratio near 50–90 K, or a quantitative cross-correlation / lag analysis, is needed before the coincidence can carry load-bearing weight in the abstract. Otherwise the language should be softened to 'consistent with' rather than 'coincides… directly linking.'","section":"Fig. 3(a) and abstract"},{"comment":"The paper argues that classical MD fails because equipartition erases the selective low-T weighting of negative-γ branches, while QHA with Bose statistics succeeds. This is a central mechanistic claim and is supported by the SM mode-resolved α_V breakdown. However, the main text does not show the mode-resolved thermal-expansion decomposition (only the colored dispersion in Fig. 5 and a qualitative narrative). Given that this is what actually ties nuclear statistics to NTE—more directly than Fig. 4—the mode-resolved α_V(T) panel (SM Fig. S7) or an equivalent should appear in the main text, with explicit cumulative contributions from TA, low-TO, and the rest of the spectrum versus T.","section":"Fig. 5; SM Fig. S7 / mode-resolved α_V"}],"minor_comments":[{"comment":"Mass rescaling of D2O experiment onto H2O simulation is validated in the SM (Figs. S1–S2) and is acceptable, but the main-text Fig. 2 caption and related discussion should state explicitly that the experimental Ic curve is mass-rescaled D2O and point to the SM validation, so readers are not left to infer the procedure.","section":"Fig. 2 caption"},{"comment":"Neutron ADPs vs ring-polymer gyration radii are correctly flagged as only qualitatively comparable, yet Fig. 3(a) overlays them as crosses on the R_g panel. A separate panel or clearer visual distinction would avoid implying a direct numerical identification.","section":"Fig. 3(a,c)"},{"comment":"Ref. [20] (del Rosso et al., arXiv:2602.13053) reports NTE in ice-I polytypes and should be discussed briefly in the introduction or discussion so the novelty relative to that concurrent work is explicit.","section":"Introduction / Discussion"},{"comment":"Typographical issues: '´Ecole Polytechnique F´ ed´ eerale' (double e); 'F d¯3mspace group' missing space; 'I sd' formatting; 'Gr¨ uneisen' spacing inconsistencies; 'oppure about' left in SM neutron-methods paragraph.","section":"Affiliations; SM"},{"comment":"Fig. 5 caption notes imaginary modes near Γ as mesh-interpolation artifacts and missing long-wavelength analytic corrections. A brief statement on whether those artifacts affect the extracted γ of the TA/TO branches used in the NTE argument would help.","section":"Fig. 5"},{"comment":"The SM reports P=48 beads at 100 K with P×T constant and PIGLET; a one-sentence convergence check (density vs P at one low-T point) in the SM would strengthen confidence in the NQE isolation.","section":"SM, PIMD methods"}],"recommendation":"major_revision","confidential_remarks":"The experimental Ic dataset and the classical-vs-PIMD MB-pol contrast are the durable contributions and would stand even if the collectivity rhetoric is toned down. The main risk is overclaiming on Fig. 4; if the authors reframe cleanly, this is a strong paper for a high-quality condensed-matter/chemical-physics venue. Worth checking overlap and citation balance with the concurrent del Rosso et al. arXiv on NTE in ice-I polytypes so priority/novelty is handled transparently."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The real news is a stacking-disorder-free Ic density curve from the C2 route that tracks Ih’s ~70 K maximum, plus MB-pol classical MD vs PIMD that isolates nuclear statistics and matches experiment to ~0.2%. That is new, useful, and low-circularity: the diffraction is external, MB-pol was not fit to this peak, and the same PES fails classically.\n\nWhat they do well is the multi-messenger stack. Gyration anisotropy peaking near the density max, neutron ADPs with large transverse proton amplitude, and mode-resolved Grüneisen maps pointing at TA/low-TO branches under Bose weighting all line up. The QHA comparison and the SM mode-resolved α_V breakdown make the statistical (not just anharmonic) point concrete. Mass-rescaling checks and isotope notes in the SM are careful. Citation pattern is normal for the subfield.\n\nSoft spot, in proportion: the single-proton embedding (Fig. 4) shows that a quantum environment is needed to get the 0 K volume right, and that local ZPE alone does not. That is a fair 0 K argument. It is not a direct proof that the finite-T NTE slope itself is collective rather than local. NTE is already largely recovered by network phonons in QHA; the embedding neglects explicit coupling and constrains the static V_OO minimum, not the thermal contraction. The broader claim still stands on Ic≈Ih, PIMD vs classical, ADPs/gyration, and negative-γ TA/TO under quantum statistics—the embedding wording just overreaches. Temperature coincidence of anisotropy and density max is within their resolution, not a precision lock. Concurrent polytype NTE work means novelty is shared, not sole.\n\nThis is for people who care about water potentials, NQE in ice, and open tetrahedral networks. Serious referee material; tighten the collectivity language and ship the digital artifacts. I would engage and cite the Ic curve and the MB-pol benchmark.","headline":"Clean Ic NTE data plus MB-pol PIMD make the network+NQE case; the embedding step overclaims collectivity for the finite-T slope.","tokens_in":20049,"tokens_out":511,"would_cite":true,"duration_ms":8989,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Cubic ice contracts on warming near 70 K for the same quantum network reason as hexagonal ice.","keywords":["negative thermal expansion","cubic ice","nuclear quantum effects","path-integral molecular dynamics","hydrogen-bond network","Grüneisen parameters","proton anisotropy","MB-pol"],"falsifier":"A stacking-disorder-free cubic-ice density curve (experiment or PIMD) that lacks a maximum near 70 K, or a classical simulation with the same potential that still produces the density maximum.","tokens_in":19806,"feed_emoji":"❄️","tokens_out":933,"duration_ms":13727,"temperature":0.7,"pith_summary":"Ice usually expands when heated, but hexagonal ice shrinks below about 70 K. This paper shows that pure cubic ice does the same thing, even though its long-range stacking of water layers is different. The shared open tetrahedral hydrogen-bond network is what matters. Neutron diffraction on stacking-disorder-free cubic ice finds a density maximum near 70 K that matches hexagonal ice. Path-integral molecular dynamics with the MB-pol potential recovers the anomaly only when nuclei are treated as quantum particles; classical nuclei do not. The density maximum lines up with the peak anisotropy of the proton quantum distribution, neutron data show large transverse proton displacements, and the modes with the most negative Grüneisen parameters are low-frequency transverse network vibrations. The contraction is therefore a collective quantum effect of the network under Bose–Einstein statistics, not a local zero-point correction on single O–H bonds or a stacking-specific feature.","feed_headline":"Cubic ice shrinks on warming for the same quantum reason as hexagonal ice","feed_subtitle":"A density maximum near 70 K is a network effect, not a stacking quirk, and needs nuclear quantum statistics","key_machinery":"The anisotropic gyration radius of the path-integral proton ring polymer (transverse over longitudinal components), whose temperature maximum coincides with the density maximum and is linked to low-frequency transverse acoustic and optical modes that carry the most negative Grüneisen parameters and are weighted by Bose–Einstein statistics.","core_discovery":"Negative thermal expansion in ice I is an intrinsic response of the open tetrahedral hydrogen-bond network and is largely independent of stacking sequence. Stacking-disorder-free cubic ice Ic has a density maximum near 70 K that closely matches hexagonal ice Ih. MB-pol simulations reproduce the anomaly quantitatively only when nuclear quantum effects are included; classical dynamics do not. The temperature of the density maximum coincides with maximal anisotropy of the proton quantum path, transverse proton displacements are strongly enhanced, and the contractive free-energy contributions come from low-frequency transverse modes under quantum statistics.","pith_inferences":["The same transverse-network plus quantum-statistics mechanism should appear in other low-density tetrahedral frameworks (certain clathrates or ices) once stacking disorder and classical nuclear treatments are removed.","Isotope substitution that leaves the anisotropy ratio almost unchanged while shifting absolute delocalization offers a clean experimental dial on whether the density-maximum temperature tracks anisotropy or absolute zero-point amplitude.","If the single-proton embedding is replaced by a fully coupled multi-proton quantum treatment, the classical-versus-quantum environment contrast should still be required to recover the anomalous zero-kelvin density, or the collectivity argument would need revision."],"forward_implications":["NTE in ice I is a property of the shared open tetrahedral network, not of hexagonal stacking.","Classical molecular dynamics cannot capture the cryogenic density maximum of ice I even with an accurate potential; nuclear quantum statistics are required.","Low-frequency transverse network modes, not intramolecular stretches or bends, dominate the contractive response under quantum weighting.","Denser ice phases are expected to lose this cryogenic NTE because compression stiffens the relevant network modes and alters proton delocalization.","Stacking-disorder-free ice Ic is a clean reference system for quantum thermodynamic anomalies in other open tetrahedral networks."],"fun_headline_variants":["Cubic ice shows density max near 70 K from quantum H-bond network","Negative thermal expansion in ice I is a shared network quantum effect","Stacking-free cubic ice shrinks on warming like hexagonal ice","Nuclear quantum effects drive ice Ic density maximum near 70 K","Transverse proton modes make cubic ice contract on warming"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"That comparing a single quantum proton in a classical versus quantum environment, while neglecting explicit coupling to the surroundings, is enough to prove the effect must be collective rather than local.","fun_headline_variants_meta":{"raw":{"variants":["Cubic ice shows density max near 70 K from quantum H-bond network","Negative thermal expansion in ice I is a shared network quantum effect","Stacking-free cubic ice shrinks on warming like hexagonal ice","Nuclear quantum effects drive ice Ic density maximum near 70 K","Transverse proton modes make cubic ice contract on warming"]},"model":"grok-4.5","effort":"low","cost_usd":0.00238,"raw_usage":{"total_tokens":951,"prompt_tokens":777,"num_sources_used":0,"completion_tokens":71,"cost_in_usd_ticks":23804000,"prompt_tokens_details":{"text_tokens":777,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":103,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":777,"tokens_out":71,"duration_ms":3033,"temperature":1.0,"reasoning_tokens":103,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T13:27:45.393513+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A stacking-disorder-free cubic-ice density curve (experiment or PIMD) that lacks a maximum near 70 K, or a classical simulation with the same potential that still produces the density maximum.","supporting_citations":[],"review_version":1}