{"id":"5c0f121a-6d4b-411d-8190-9851759171e1","arxiv_id":"1908.11589","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Path integral simulations from 300 K to 1.67 K show that oxygen quantum delocalization can match or exceed proton delocalization in finite water clusters at ultra-low temperatures, while the effect is absent in ice.","lead":"This paper uses path integral simulations to track hydrogen-bond structure and nuclear quantum delocalization in water clusters, protonated water clusters, and ice from 300 K down to 1.67 K. It finds that at ultra-low temperatures the quantum spread of oxygen atoms can rival or exceed that of protons in weakly constrained clusters, while strong hydrogen bonds are hardly perturbed by temperature.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central O/H delocalization crossover in finite clusters rests solely on fitted potentials without ab initio validation, so a force-field error in low-frequency modes could change the result.","rationale":"The paper is a methodologically careful computational study that systematically maps hydrogen-bond properties from 300 K to 1 K. The PIGLET path integral technique is validated, and the ice Ih result is independently supported by RPBE-D3 PIMD. The physical interpretation via coordination is plausible. However, the central claim of an O/H quantum delocalization crossover in finite clusters is only computed with fitted potentials, with no ab initio or experimental cross-check at the relevant temperature. The radius of gyration at ultra-low temperature is extremely sensitive to the zero-point motion in low-frequency modes; these modes are difficult to fit and are not covered by the ice validation because ice does not show the crossover. The absence of error bars makes it impossible to judge the significance of the small O/H differences. This is precisely the concern the reader identified, and it is the most load-bearing uncertainty in the paper. Since the reader has already marked the verdict CONDITIONAL, our stress-test does not change that verdict; it sharpens the requirement: an ab initio calculation for at least the Zundel cation is needed to confirm the headline effect.","tokens_in":13103,"tokens_out":6803,"duration_ms":61062,"concrete_test":"Run ab initio PIMD at 1.67 K for the Zundel cation using a high-level electronic structure method (e.g., MP2 or CCSD(T) with a sufficiently large basis), with the same PIGLET thermostat and path integral discretization, and compute the averaged radius of gyration of the two oxygen atoms and the shared proton. Compare the O/H ordering and magnitudes with the NNP results in Fig. 3; if the oxygen delocalization still exceeds the proton's, the model-dependence concern is resolved, otherwise the central claim fails at least for the Zundel case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The claim that oxygen delocalization approaches or exceeds proton delocalization at ~1 K is computed for finite clusters exclusively with q-TIP4P/F for neutral systems and a CCSD(T)-fitted neural network potential for cationic systems. In contrast to ice Ih, which is cross-checked with RPBE-D3 PIMD, no finite cluster is validated at the target temperature. At 1 K, the radius of gyration is governed by the zero-point curvature of low-frequency intermolecular modes; for oxygens, this involves large-amplitude motion far from the fitting configurations. If q-TIP4P/F or the NNP underestimates the restoring force on oxygen (or overconfines the proton), the O/H crossover in Fig. 3 is an artifact. No statistical error bars are given, so the small O/H differences in the dimer and trimer cannot be assessed for significance. This unresolved model dependence is the weakest link in the argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses path integral molecular dynamics with colored-noise (PIGLET) thermostatting to study hydrogen-bonded systems from 300 K down to 1.67 K, explicitly including nuclear quantum effects. The systems are the water dimer, water hexamer, protonated Zundel, trimer, and Eigen clusters, and hexagonal ice Ih. Structural distributions (rOO, δ, ∠HOO) and the path-integral radius of gyration (Eq. 1) are computed as a function of temperature. The central claim is that at around 1 K the quantum delocalization of oxygen atoms approaches or exceeds that of protons in finite clusters, while in ice Ih the usual ordering (H more delocalized than O) persists. The crossover is attributed to the weaker spatial constraints (lower coordination) in small clusters.","tokens_in":13129,"tokens_out":3011,"duration_ms":29135,"significance":"If the central claim is correct, the paper demonstrates a qualitative breakdown of the mass-based expectation that light atoms are always more quantum-delocalized than heavy atoms in hydrogen-bonded systems, with direct relevance for interpreting ultra-cold cluster spectroscopy. The systematic temperature ladder from ambient to 1 K and the comparison across neutral, cationic, finite, and condensed systems is a valuable contribution. The ice Ih results are cross-validated with ab initio RPBE-D3 PIMD at 1.67 K, which is a concrete strength. The use of a neural-network potential fitted to CCSD(T) for the protonated clusters is also a step beyond empirical force fields. However, the central O/H crossover in finite clusters is not independently validated at the target temperature, and no statistical error bars are given, so the robustness of the headline finding remains uncertain.","major_comments":[{"comment":"The central claim that oxygen delocalization approaches or exceeds proton delocalization at ~1 K in finite clusters is supported only by q-TIP4P/F (neutral clusters) and an NNP fitted to CCSD(T) (cationic clusters). The only ab initio validation, RPBE-D3 PIMD, is performed for ice Ih (triangles in the right panel of Fig. 3), where the effect is absent. At 1 K the radius of gyration is dominated by zero-point motion in low-frequency intermolecular modes, which are far from the typical fitting configurations of both potentials. A force-field error in these modes could change the O/H ordering. Please provide an explicit ab initio PIMD cross-check for at least one finite cluster (e.g., the Zundel cation or the water dimer) at 1.67 K, or alternatively quantify the expected uncertainty in the NNP and q-TIP4P/F radii of gyration for the relevant modes.","section":"Section III, Fig. 3"},{"comment":"No statistical error bars are reported for the averaged radius of gyration ⟨r_g⟩. The differences between O and H shown for the dimer and the protonated trimer at 1.67 K appear to be on the order of 0.05 Å or less, comparable to typical block-averaging uncertainties for a 125 ps simulation with 256 replicas. Without error bars, the claimed crossover cannot be distinguished from statistical noise. Please report standard errors (e.g., from block averaging) and specify the number of independent samples after decorrelation.","section":"Fig. 3 and Section II"},{"comment":"The manuscript states that below 100 K the water hexamer remained in its starting isomer and that only the ordered hexagonal ring is used below 100 K, with the statement that other isomers behave similarly. This limits the generality of the finite-cluster conclusion for the hexamer. Since the 'interaction induced localization effect' is argued to depend on coordination, a single isomer is a reasonable starting point, but the claim that the crossover would also occur for other isomers should either be demonstrated or explicitly labeled as a tentative extension rather than a result of the present simulations.","section":"Section II, water hexamer"}],"minor_comments":[{"comment":"The phrase 'neutron defraction' appears to be a typo for 'neutron diffraction'.","section":"Introduction"},{"comment":"The phrase 'seamlessly extending our insights into noncovalent interactions down to ultra-low temperatures' is vague; consider specifying what the 'seamless' extension adds beyond the temperature range studied.","section":"Abstract"},{"comment":"The binding energy per water monomer is estimated by dissecting protonated clusters into a hydronium core and remaining waters; this definition is a reasonable rough measure, but the caption of Fig. 2 should state explicitly that the quoted hydrogen-bond strengths are approximate and system-dependent.","section":"Section II, binding energies"},{"comment":"The horizontal dashed lines are described as 'the average values of all systems at the highest considered temperatures'; please state in the caption that these are constants for orientation only, since the individual system values scatter around them.","section":"Fig. 3"},{"comment":"The statement 'it is tempting to speculate that the crossover might set in at still lower temperatures' for the hexamer and Eigen cation is a reasonable speculation, but it should be clearly separated from the directly computed results to avoid overstatement.","section":"Section III"}],"recommendation":"major_revision","confidential_remarks":"The paper's core conclusion depends on potentials (q-TIP4P/F and the CCSD(T)-fitted NNP) that are validated outside the ultra-low-temperature regime. The lack of any ab initio cluster-level check at 1.67 K is the key issue. The heavy reliance on the authors' own previous work (refs 39, 40, 44, 72, 73) is not itself a flaw, but the combination with the absence of independent validation for the finite-cluster crossover warrants a careful revision. If the authors can provide one ab initio PIMD data point for a finite cluster at 1.67 K, or a robust error estimate for the force-field/NNP radii of gyration, the paper would be publishable; as is, the central claim is not yet fully supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper for two reasons. First, it is the most systematic temperature scan to date of nuclear quantum delocalization in hydrogen-bonded systems, covering six systems from 300 K down to 1.67 K with a consistent path-integral framework. Second, its central claim—that at ~1 K the quantum delocalization of oxygen can approach or exceed that of protons in small clusters, most dramatically in the Zundel cation—is genuinely striking and will get cited. The companion observation that this effect is absent in ice Ih, where coordination suppresses it, is the real new result.\n\nWhat the paper does well: the simulations use a validated PIGLET thermostat for ultra-low temperatures, with path-integral discretizations checked in earlier work. The neutral systems use q-TIP4P/F, the cations a neural network potential fitted to CCSD(T), and ice is independently cross-checked with ab initio RPBE-D3 PIMD at 1.67 K. The structural analysis (rOO, proton sharing coordinate, angles) is thorough and the temperature trends are physically sensible: strong hydrogen bonds in protonated clusters are nearly temperature-independent, while weak bonds in the neutral clusters soften further with warming.\n\nThe soft spots are real but not crushing. No statistical error bars appear anywhere, so the small O/H differences in the dimer and trimer at 1.67 K—the quantitative basis for the crossover—cannot be distinguished from noise. More importantly, the finite-cluster crossover is computed only with q-TIP4P/F and the NNP; there is no ab initio check on any finite cluster at the target temperature, unlike ice. The authors argue, fairly, that the NNP is essentially converged to coupled-cluster, but that still leaves the low-frequency intermolecular modes at 1 K dependent on fitted potential shapes. The coordination explanation is qualitative, not a controlled test. None of these issues sinks the paper; they just mean the boldest conclusion is not fully nailed down.\n\nThe citation pattern leans heavily on the authors' own prior work for the effect itself, the thermostat, the NNP, and convergence. That is not a flaw here, because those are indeed the relevant references and the effect was first reported by them. The new contribution is the systematic temperature behavior and the ice contrast.\n\nWho is this for? Anyone working on ultra-cold spectroscopy of water clusters or on nuclear quantum effects in hydrogen bonding. It deserves serious peer review. I would send it to review, with a request for error bars and, ideally, one ab initio validation on a finite cluster at 1 K.","headline":"A careful systematic PIMD study showing oxygen delocalization can match or exceed proton delocalization in small hydrogen-bonded clusters near 1 K, absent in ice—but the finite-cluster crossover rests on fitted potentials and lacks error bars.","tokens_in":13822,"tokens_out":2206,"would_cite":true,"duration_ms":23002,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"At around 1 K, oxygen nuclei in small hydrogen-bonded clusters delocalize as much as—or more than—the protons they bond to, reversing the usual mass-based ordering of quantum spread.","keywords":["hydrogen bond","nuclear quantum effects","path integral molecular dynamics","ultra-low temperature","quantum delocalization","Zundel cation","water clusters","ice Ih"],"falsifier":"Compute the averaged radius of gyration of oxygen and hydrogen in the Zundel cation at 1.67 K using a converged ab initio path integral simulation (for example, with a CCSD(T)-level potential energy surface or a different machine-learned potential). If the oxygen $\\langle r_g \\rangle$ is not larger than the proton's, the central reversal claim fails. A complementary observable test: measure the zero-point kinetic energy of oxygen versus hydrogen in cryogenic protonated water clusters via neutron Compton scattering; the predicted reversal implies an anomalously high oxygen kinetic energy at 1 K.","tokens_in":12711,"feed_emoji":"❄️","tokens_out":6007,"duration_ms":51976,"temperature":0.7,"pith_summary":"This paper asks how the quantum fuzziness of atoms in hydrogen bonds changes when temperature drops from room temperature to about 1 K, the regime of helium-nanodroplet and tagging spectroscopies. Using path integral simulations that explicitly include nuclear quantum effects, it compares the water dimer, water hexamer, three protonated water clusters (Zundel, trimer, Eigen), and hexagonal ice. The central finding is that in the small finite clusters the quantum delocalization of the heavy oxygen nuclei, measured by the averaged radius of gyration of the path integral, approaches and in some cases exceeds that of the lighter protons as the temperature approaches 1 K. In ice, by contrast, the usual ordering is preserved at all temperatures. The paper argues that coordination—how many hydrogen bonds surround each atom—controls the effect, since stronger spatial constraints suppress the delocalization of the heavy atoms.","feed_headline":"At 1 K, oxygen nuclei spread wider than protons in some water clusters","feed_subtitle":"Path-integral simulations show the mass-based ordering of quantum delocalization reverses in small clusters but not in ice.","key_machinery":"The central object is the imaginary-time path integral representation of the quantum nuclei, and the averaged radius of gyration of the ring polymer, $r_g^2 = \\frac{1}{P}\\sum_{s=1}^P \\langle (R_s - R_c)^2 \\rangle$, which measures each nucleus's instantaneous quantum delocalization and equals the quantum contribution $\\langle \\Delta x^2 \\rangle_q$ to the total position fluctuations. The PIGLET thermostat allows converged path integral simulations down to 1.67 K with affordable replica counts. Two potential energy surfaces carry the systems: q-TIP4P/F for neutral water clusters and ice, and a neural network potential fitted to coupled cluster reference calculations for the protonated clusters, which rules out force-field artifacts. The coordination number of each hydrogen-bonding atom is the explanatory variable that distinguishes finite clusters from the condensed phase.","core_discovery":"The paper's central claim is that the spatial quantum delocalization of oxygen nuclei in hydrogen-bonded clusters can match or exceed that of the protons they bond to at temperatures near 1 K, reversing the naive $1/\\sqrt{M}$ scaling of the thermal de Broglie wavelength. For the Zundel cation, whose hydrogen bond is centered and termed 'ultra-strong', the average radius of gyration of the oxygen atoms significantly exceeds that of the shared proton at 1.67 K; the water dimer and protonated trimer show near-degeneracy, while the hexamer and Eigen cation show a strongly reduced gap. In hexagonal ice Ih, the oxygen–hydrogen delocalization difference stays nearly constant down to 1 K and the reversal never occurs. The authors trace this difference to the coordination of the hydrogen-bonded atoms: increased coordination in the condensed phase constrains translational and rotational quantum delocalization and quenches the 'interaction induced localization' of the protons.","pith_inferences":["The paper does not simulate clusters inside helium nanodroplets, but its coordination argument suggests that the helium solvent, by adding an external confining potential, could quench or shift the oxygen/hydrogen crossover in the same way ice does—an experimentally testable prediction.","A direct test of potential dependence would be to run the same 1 K path integral protocol on the Zundel cation using a different high-level potential energy surface (for example, a CCSD(T) grid or a second machine-learned potential) and compare $\\langle r_g \\rangle$ for oxygen and hydrogen; the paper's neural network has coupled cluster accuracy, but only the ice results are cross-validated with a","Since the radius of gyration maps to the quantum kinetic energy through the virial estimator, the predicted reversal might also be observable in path-integral estimates of isotope fractionation or in neutron Compton scattering line shapes of cryogenic clusters, if such measurements become feasible.","The temperature at which the oxygen and hydrogen radii cross (if any) is left open for some systems; the paper speculates that the hexamer and Eigen cation might cross below 1 K. A systematic scan to 0.1 K with a converged potential would map the crossover locus as a function of coordination and hydrogen bond strength."],"forward_implications":["Ultra-cold spectroscopic experiments on isolated water clusters probe a regime in which heavy atoms are not classical: the oxygen framework itself is quantum-delocalized on a scale comparable to the proton, so assigning spectra with fixed heavy-atom geometries may mislead.","For strong, centered hydrogen bonds like Zundel's, structural observables ($r_{OO}$, $\\delta$, $\\angle HOO$) are essentially temperature-independent from 1 K to 250 K, meaning ground-state quantum effects dominate all the way up to ambient conditions.","For weak hydrogen bonds (water dimer, hexamer), results at 1 K cannot be directly transferred to ambient conditions; temperature substantially weakens and bends these bonds.","In condensed-phase ice, nuclear quantum delocalization differences between oxygen and hydrogen remain stable down to 1 K, so models that work at ambient conditions may remain valid at ultra-low temperature, at least for this property.","The coordination argument predicts that any finite cluster with low coordination should show the oxygen/hydrogen delocalization crossover at sufficiently low temperature, generalizing beyond water to other hydrogen-bonded dimers."],"supporting_citations":[{"why":"First described the 'interaction induced localization' of protons in hydrogen-bonded dimers at ultra-low temperatures, the effect this paper systematically extends and explains.","marker":"[72]"},{"why":"Provides the neural network potential fitted to coupled cluster reference calculations for the protonated water clusters, showing the effect is not an artifact of simple force fields.","marker":"[44]"},{"why":"Introduces the q-TIP4P/F force field used for the neutral water clusters and ice, explicitly parameterized for path integral simulations.","marker":"[37]"},{"why":"Extends PIGLET thermostatting to ultra-low temperatures, making the 1 K path integral simulations computationally feasible.","marker":"[39]"},{"why":"Validates the path integral discretization at ultra-low temperatures for the Zundel cation hydrogen bond, justifying the replica counts used here.","marker":"[40]"},{"why":"Early work on protonated water clusters that first indicated temperature independence of the strong hydrogen bond, confirmed here for Zundel.","marker":"[26]"},{"why":"The PIGLET thermostat method itself, which underpins all reported path integral molecular dynamics runs.","marker":"[53]"},{"why":"Shows helium–solute interactions in nanodroplets increase localization, providing context for the confinement argument the paper applies to coordination.","marker":"[73]"}],"fun_headline_variants":["At 1 K, oxygen atoms can out-spread protons in water clusters","Quantum reversal: oxygen delocalization exceeds protons at 1 K","Heavy oxygen spreads more than protons in cold water clusters","Mass ordering flips: oxygen delocalization beats protons near 1 K","Cold clusters flip quantum delocalization order, but ice doesn't"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fitted potentials—q-TIP4P/F for neutral systems and the coupled-cluster-fitted neural network for protonated clusters—correctly describe how oxygen and hydrogen nuclei spread out at about 1 K in the finite clusters; only the ice results are checked against ab initio path integral simulation at that temperature.","fun_headline_variants_meta":{"raw":{"variants":["At 1 K, oxygen atoms can out-spread protons in water clusters","Quantum reversal: oxygen delocalization exceeds protons at 1 K","Heavy oxygen spreads more than protons in cold water clusters","Mass ordering flips: oxygen delocalization beats protons near 1 K","Cold clusters flip quantum delocalization order, but ice doesn't"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000883,"raw_usage":{"total_tokens":3850,"prompt_tokens":1019,"completion_tokens":2831,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":635,"completion_tokens_details":{"reasoning_tokens":2738}},"tokens_in":635,"tokens_out":2831,"duration_ms":18319,"temperature":1.0,"reasoning_tokens":2738,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:10:42.831596+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the averaged radius of gyration of oxygen and hydrogen in the Zundel cation at 1.67 K using a converged ab initio path integral simulation (for example, with a CCSD(T)-level potential energy surface or a different machine-learned potential). If the oxygen $\\langle r_g \\rangle$ is not larger than the proton's, the central reversal claim fails. A complementary observable test: measure the zero-point kinetic energy of oxygen versus hydrogen in cryogenic protonated water clusters via neutron Compton scattering; the predicted reversal implies an anomalously high oxygen kinetic energy at 1 K.","supporting_citations":[{"cited_title":"Automated Fitting of Neural Network Potentials at Coupled Cluster Accuracy: Protonated Water Clusters as Testing Ground","cited_arxiv_id":"1908.08734","evidence_quote":"Provides the neural network potential fitted to coupled cluster reference calculations for the protonated water clusters, showing the effect is not an artifact of simple force fields."},{"cited_title":"Uhl , author D","cited_arxiv_id":null,"evidence_quote":"Extends PIGLET thermostatting to ultra-low temperatures, making the 1 K path integral simulations computationally feasible."},{"cited_title":"Schran , author F","cited_arxiv_id":null,"evidence_quote":"Validates the path integral discretization at ultra-low temperatures for the Zundel cation hydrogen bond, justifying the replica counts used here."}],"review_version":1}