{"id":"9568fb04-7b67-4bd6-9ccb-c9563a7050b1","arxiv_id":"2411.12545","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Using CCSD(T)-accurate path integral simulations of water, the authors predict the 1H NMR relaxation rate at 28 MHz as 0.278 s^-1, within the 0.280 ± 0.003 s^-1 experimental value.","lead":"This paper computes the 1H NMR relaxation rate of liquid water from high-accuracy quantum simulations and matches the 1966 experiment almost exactly. It shows that nuclear quantum effects and a corrected intramolecular hydrogen-hydrogen distance are what make the prediction work.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The final R1 agreement rests on the untested transferability of the TIP4P/2005-derived intermolecular difference function ΔJ_inter to CCMD and experimental structures; since ΔJ_inter encodes the short-time librational and jump motions that NQE demonstrably alter, this assumption is load-bearing.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing concern: the intermolecular difference function is taken from classical TIP4P/2005 simulations and applied unchanged to CCMD and experimental datasets. This is a genuine soft spot because ΔJ_inter is defined as the short-time deviation from Hwang-Freed diffusion, dominated by librations and jump reorientation, and the paper itself shows that quantum nuclear effects significantly quench the analogous intramolecular reorientational correlation function. The mathematical structure of the argument makes the assumption testable: only G_inter(t) from CCMD trajectories is needed to replace the transferred function. If the resulting R1_inter shifts by more than the 0.003 s^-1 experimental tolerance, the headline 'close to perfect agreement' is no longer supported. This concern does not invalidate the paper; it means the central claim is conditional on a check that has not been performed. Missing error bars on Table I are a secondary issue that would also need attention, but the transferability of ΔJ_inter is the more incisive issue. The reader's conditional verdict remains appropriate, so no verdict change is recommended.","tokens_in":18716,"tokens_out":4376,"duration_ms":48369,"concrete_test":"Compute G_inter(t) directly from the 256-molecule CCMD (RPMD) trajectories, form ΔG_inter(t) = G_CCMD(t) - G_HF(t) using the CCMD dHH and D0 with the same ttr and Equation 45-style finite-size scaling (or an equivalent extrapolation), and recompute R1_inter at 28 MHz. If R1_inter changes by more than about 0.003 s^-1 relative to the reported 0.1378 s^-1, the transferability assumption fails and the total moves outside the experimental uncertainty.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV.C computes all intermolecular relaxation rates using \"the same TIP4P/2005 derived ΔJ_inter functions for all three datasets.\" This transfers a classical short-time dynamical correction to CCMD and experimental structures without testing it. ΔJ_inter (Eqs. 40-45) is the deviation of the MD dipole-dipole correlation function from Hwang-Freed diffusion, and the text attributes it specifically to librational motions and Laage-Hynes jump reorientation. Those are precisely the motions that Section IV.B shows are quenched by nuclear quantum effects: C2^HH(t) from CCMD lies significantly below TIP4P/2005, τ2^HH drops by 12.8%, and D0×τHH changes from 0.570 Ų to 0.506 Ų. Applying an unmodified classical ΔJ_inter therefore assumes that the short-time intermolecular spectral density is insensitive to the same NQE that the paper demonstrates matter for intramolecular reorientation. The robustness argument offered (R1_inter ∝ 1/dHH and small dHH sensitivity) does not test this assumption: it only varies the static distance of closest approach while keeping the dynamical difference function fixed. The final prediction (0.2782 vs 0.280 ± 0.003 s^-1) sits within about 0.7% of experiment, so even a modest shift in R1_inter arising from a CCMD-corrected ΔJ_inter could move the total outside the reported agreement. This is the least secure link between the CCMD trajectories and the headline claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript computes the 1H NMR dipole-dipole relaxation rate R1 of liquid water at 298 K and 28 MHz using a combination of coupled-cluster molecular dynamics (CCMD) trajectories that include nuclear quantum effects, structural data from neutron scattering and from the experimental consensus H-H distance of Faux et al., and a previously introduced framework that separates intermolecular relaxation into a Hwang-Freed diffusion part and a short-time difference function computed from MD. With CCMD-derived structural parameters (intramolecular H-H distance 154.1 pm, DCA 195.68 pm) and a rescaling of the intramolecular correlation time by the ratio D0(CCMD)/D0(expt), the authors obtain R1 = 0.2782 s-1 at 28 MHz, to be compared with the experimental value of (0.280 ± 0.003) s-1. They also find that the intramolecular and intermolecular contributions are nearly equal in magnitude, in contrast to earlier classical simulations.","tokens_in":19085,"tokens_out":5258,"duration_ms":55119,"significance":"If the result holds, the paper would constitute a quantitative, essentially parameter-free prediction of a fundamental NMR relaxation rate in liquid water, and it would demonstrate that nuclear quantum effects and accurate intermolecular structure are both needed for that prediction. The manuscript's strengths are its use of multiple independent structural inputs (CCMD, Soper's neutron data, Faux et al.'s experimental distance) that agree on the key intramolecular H-H distance, the explicit finite-size correction for the intermolecular part, and the open-source availability of the software (MDorado and FreeDRelax). However, the central claim rests on an untested transferability assumption for the intermolecular short-time difference function, and the reported agreement is sensitive to the small rebalancing of rotational and translational dynamics. These issues need to be resolved before the claim of quantitative prediction can be considered fully supported.","major_comments":[{"comment":"The intermolecular relaxation rate uses the same TIP4P/2005-derived difference function ΔJ_inter for the TIP4P/2005, CCMD, and experimental structural datasets. This is an explicit assumption: the manuscript states 'using the same TIP4P/2005 derived Δ J n inter(ω) functions for all three datasets.' The difference function encodes short-time librational and jump-reorientation dynamics, and Section IV.B shows that these dynamics are significantly altered by nuclear quantum effects (τ_HH drops by 12.8% and D0×τHH changes from 0.570 to 0.506 Ų in CCMD relative to TIP4P/2005). The robustness argument in Section IV.C, based on R1_inter ∝ 1/dHH, only varies the static distance of closest approach and does not test whether the dynamic correction is transferable. Because the final total R1 agrees with experiment to within about 0.7%, an unquantified change in R1_inter can move the prediction outside the reported agreement. The authors should either compute ΔG_inter and ΔJ_inter directly from CCMD trajectories or provide a quantitative sensitivity estimate that bounds the error introduced by using the classical difference function.","section":"Section IV.C, Eqs. (43)-(45)"},{"comment":"The rebalancing step τG ≈ τ_HH × D0(CCMD)/D0(expt) is essential to the final agreement: without it the intramolecular rate is 0.1380 s-1 and the total at 28 MHz is about 0.2758 s-1, which is outside the experimental lower bound of 0.277 s-1. Although the rescaling is small (about 1.8%), it is introduced because the CCMD diffusion coefficient overestimates experiment by a comparable amount. The manuscript should justify why this procedure is a principled correction rather than an empirical adjustment, for example by showing that the same ratio correctly describes D2O or by propagating the uncertainty in D0 and demonstrating that the result is stable across the experimental and computed D0 values. As written, the headline agreement depends on this rescaling, so its status as a first-principles prediction needs clarification.","section":"Section IV.D, Eq. (39)"},{"comment":"The 'experimental' dataset combines the Soper neutron-scattering gHH with the Faux et al. consensus intramolecular H-H distance. While both are reasonable independent sources, they are not measured on the same sample at the same thermodynamic state, and the manuscript does not discuss possible systematic differences between the two. This is a relatively minor concern because the dHH sensitivity is small, but the authors should at least state the assumptions implied by combining these two experimental sources.","section":"Section IV.C, 'experimental structure' dataset"}],"minor_comments":[{"comment":"The phrase 'structural and dynamical informations' should be 'structural and dynamical information'; similar grammatical slips appear elsewhere (e.g., 'reorienational' in Section IV.B and 'empricial' in Section IV.B).","section":"Abstract"},{"comment":"The name 'Gonz´ ales' in the text should be 'González' to match the cited reference (González and Abascal).","section":"Section IV.A"},{"comment":"The phrase 'empricial Kohlrausch-Williams Watts (KWW)' should read 'empirical Kohlrausch-Williams-Watts (KWW)'.","section":"Section IV.B, Eq. (28)"},{"comment":"The caption of Fig. 4 would benefit from stating explicitly which system size (8192 molecules) and which DCA value are used for the Hwang-Freed reference curve, since the figure is central to understanding ΔG_inter.","section":"Section IV.C, Fig. 4"},{"comment":"The rows for νH=0 and νH=28 MHz show R1,intra values of 0.1404 s-1 in both cases, which is consistent with extreme narrowing but could be clarified in the text.","section":"Section IV.D, Table I"},{"comment":"The statement that data are available 'upon reasonable request' is weaker than the open-source software statement; archiving the key correlation functions and fitted parameters would improve reproducibility.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and presents a clearly described framework with reproducible software. The main concern is the transferability of the TIP4P/2005-derived intermolecular difference function to CCMD and experimental structures; if the authors can address this with a CCMD-based ΔJ_inter or a convincing sensitivity analysis, the paper will likely be acceptable. The rebalancing procedure also deserves a clearer justification, but it is a smaller issue that can be fixed in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result is real: Table I gives R1 = 0.2782 s⁻¹ at 28 MHz versus the experimental 0.280 ± 0.003 s⁻¹, and the path to it is mostly clean. The genuinely new piece is that the intramolecular and intermolecular contributions come out nearly equal, which contradicts earlier classical-MD pictures that put most weight on the intramolecular part. That difference traces to the larger, more realistic H–H distance from CCMD and to the rebalancing of translation versus rotation through D0 × τHH. The paper also does several things right: the target R1 is never used as input, the τHH rescaling uses the experimental self-diffusion coefficient as an external benchmark, structural inputs are checked against Soper's neutron data and Faux et al.'s consensus distance, and the analysis software is on GitHub. This is a strong contribution for the NMR-relaxation and water-simulation communities.\n\nThe soft spots are real but not fatal. The most important is the one the paper states explicitly in Section IV.C: the intermolecular difference function ΔJ_inter is computed from classical TIP4P/2005 simulations and applied unchanged to the CCMD and experimental structural datasets. That function is supposed to capture librational and jump-type short-time dynamics, and the paper's own Figure 3 shows that nuclear quantum effects measurably quench those dynamics. The robustness argument offered—that R1_inter depends only weakly on dHH—only varies the static distance of closest approach; it does not test whether the short-time dynamical correction transfers. Given that the final agreement sits within about 0.7% of experiment, a modest change in the intermolecular spectral density could move the total outside the experimental error. A referee should ask for a CCMD-derived ΔJ_inter or at least a sensitivity estimate.\n\nTwo smaller issues: the final rates in Table I have no error bars, which is surprising for a paper whose argument depends on matching an experimental error bar; and the paper leans on the earlier TIP4P/2005 work from the same group for the framework and the θ parameter, which is fine conceptually but leaves the finite-size correction somewhat self-referential until independently reproduced.\n\nOverall, the central claim is plausible and the paper deserves a serious referee. It will be useful to anyone benchmarking water potentials or NMR relaxation models, and it should be published after the transferability assumption and error bars are addressed. I would send it to review, not desk reject.","headline":"The paper makes a quantitatively successful prediction of the 1H NMR relaxation rate of water using CCSD(T)-level path integral simulations, with one untested dynamical assumption that a referee should pin down.","tokens_in":676,"tokens_out":2220,"would_cite":true,"duration_ms":35503,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["76.60.-k","31.15.Qg"],"model":"deepseek-v4-flash","headline":"The 1H NMR spin-lattice relaxation rate of neat liquid water can be computed from quantum-accurate molecular dynamics and match the measured value within experimental error.","keywords":["NMR relaxation","liquid water","molecular dynamics","nuclear quantum effects","CCSD(T)","Hwang-Freed theory","dipolar relaxation","TIP4P/2005"],"falsifier":"Recompute $\\Delta J_{\\mathrm{inter}}$ directly from CCMD trajectories (or from experimental-structure simulations) rather than importing the TIP4P/2005 difference function; if the total $R_1$ at 28 MHz then leaves the window $0.277$–$0.283$ s$^{-1}$, the transferability assumption fails. As a second check, field-cycling measurements above 1 GHz would test the predicted roughly 15% drop in the intermolecular contribution.","tokens_in":18532,"feed_emoji":"💧","tokens_out":7571,"duration_ms":67649,"temperature":0.7,"pith_summary":"The paper sets out to answer a 58-year-old question: can theory predict the measured 1H NMR spin-lattice relaxation rate of neat liquid water? It computes the dipolar relaxation rate using trajectories from Coupled Cluster Molecular Dynamics at CCSD(T) accuracy with nuclear quantum effects, combined with a recently introduced framework that separates structural and dynamical contributions and corrects for finite-size effects. The central result is that when structure, intramolecular H-H distance, and a rebalanced rotational-translational dynamics (the product $D_0 \\times \\tau_{\\mathrm{HH}}$) are taken from those trajectories, the computed rate at 28 MHz is 0.2782 s$^{-1}$, inside the experimental value $0.280 \\pm 0.003$ s$^{-1}$. The paper also concludes that nuclear quantum effects reduce the intramolecular contribution, and that the inter- and intramolecular parts are nearly equal in magnitude.","feed_headline":"Quantum-accurate theory nails water's NMR relaxation rate","feed_subtitle":"Computed 0.2782 s^-1 vs measured 0.280 ± 0.003 s^-1 at 28 MHz, once rotation and translation are rebalanced.","key_machinery":"The argument is carried by a decomposition of the dipolar correlation function into a structural prefactor (the $r^{-6}$-weighted average) and a normalized dynamical correlation function. For the intermolecular part, the normalized function is written as the analytical Hwang-Freed diffusive correlation function, describing two hard spheres diffusing past each other, plus a short-time 'difference function' $\\Delta G_{\\mathrm{inter}}(t)$ computed from MD, which carries the librational and hydrogen-bond-jump corrections; the difference function is scaled to infinite system size with a Yeh-Hummer-type correction. The intramolecular part uses the reorientational correlation function of the H-H vector $C_2^{\\mathrm{HH}}(t)$, fitted to a Kohlrausch-Williams-Watts stretched exponential for long times, multiplied by the effective distance $\\langle r_{\\mathrm{HH}}^{-3}\\rangle^{-2}$. The key dynamical control is the product $D_0 \\times \\tau_{\\mathrm{HH}}$; the authors treat this rotation-translation balance as the quantity that nuclear quantum effects shift, and they rescale $\\tau_{\\mathrm{HH}}$ by the ratio $D_0(\\mathrm{CCMD})/D_0(\\mathrm{expt})$ to bring the computed rate into agreement.","core_discovery":"In the paper's own terms, the discovery is that the 1H dipolar relaxation rate of liquid water at 298 K is quantitatively predictable from theory. The full calculation, using CCMD structures and dynamics, an effective intramolecular H-H distance of 154.1 pm, and a rebalancing of $D_0 \\times \\tau_{\\mathrm{HH}}$, gives $R_1(28\\,\\mathrm{MHz}) = 0.2782$ s$^{-1}$; the experimental value is $0.280 \\pm 0.003$ s$^{-1}$. The paper attributes the previous 13% overestimation from the TIP4P/2005 model mainly to the intramolecular H-H distance and the missing quantum effects on reorientation. It reports that the intermolecular and intramolecular contributions are each about 0.14 s$^{-1}$, so the intermolecular part is roughly as large as the intramolecular part, contrary to earlier classical predictions.","pith_inferences":["Beyond the paper: if the transferred short-time difference function holds at other temperatures, the same framework should reproduce the measured temperature dependence of $R_1$ from 0 to 110 °C, a testable prediction with existing experimental data.","Beyond the paper: the product $D_0 \\times \\tau_{\\mathrm{HH}}$ could serve as a cheap diagnostic for water models—force fields that reproduce this balance without rescaling are likely to predict relaxation rates correctly, and those that do not will need the same correction.","Beyond the paper: the near-equality of inter- and intramolecular contributions suggests that NMR relaxation studies of other hydrogen-bonded liquids, such as alcohols and amides, may need to treat both channels explicitly rather than assuming intramolecular dominance.","Beyond the paper: because the difference function is borrowed from classical TIP4P/2005 dynamics, an explicit path-integral-based computation of $\\Delta J_{\\mathrm{inter}}$ would remove the main residual model dependence; if it changed the rate beyond $0.003$ s$^{-1}$, the agreement would weaken."],"forward_implications":["Theory can serve as a quantitative benchmark for interpreting 1H NMR relaxation in water without relying on analytical models of molecular motion.","The intramolecular H-H distance of rigid water models (151.4 pm) is too short; using an effective distance near 154 pm largely removes the systematic overestimation of $R_1$.","Nuclear quantum effects materially change the relaxation prediction by quenching the reorientational correlation function and reducing $D_0 \\times \\tau_{\\mathrm{HH}}$, so classical simulations alone misattribute the relaxation balance.","The near-equality of inter- and intramolecular contributions revises earlier claims that intramolecular relaxation dominates, and implies both channels must be treated on equal footing in future analyses.","The predicted frequency dependence, with the intermolecular contribution dropping roughly 15% at 1.2 GHz, is a concrete signature testable by field-cycling relaxometry."],"supporting_citations":[{"why":"Supplies the computational framework: Hwang-Freed splitting, the MD difference function, and finite-size corrections for intermolecular relaxation.","marker":"[30,31]"},{"why":"Generates the CCMD trajectories at CCSD(T) accuracy with path-integral nuclear quantum effects used for structure and dynamics.","marker":"[21]"},{"why":"Provides the CCMD self-diffusion coefficients and the analysis of nuclear quantum effects on structure versus dynamics.","marker":"[34]"},{"why":"Supplies experimental radial distribution functions used to derive an alternative distance-of-closest-approach $d_{\\mathrm{HH}}$.","marker":"[35]"},{"why":"Supplies the consensus experimental intramolecular H-H distance and the Levy rotor reorientational picture.","marker":"[36]"},{"why":"Provides the experimental $R_1 = 0.280 \\pm 0.003$ s$^{-1}$ at 28 MHz that the calculation targets.","marker":"[1]"},{"why":"Gives the analytical Hwang-Freed translational diffusion model that is the baseline for the intermolecular spectral density.","marker":"[9]"},{"why":"Supplies the system-size correction for diffusion coefficients used to obtain size-independent $D_0$.","marker":"[25]"}],"fun_headline_variants":["Water's 1H NMR relaxation rate predicted to 0.6%","Quantum-accurate simulation matches water NMR experiment","Intermolecular and intramolecular water NMR contributions equal","CCSD(T)-level dynamics reproduce water 1H relaxation","Theory nails water NMR relaxation rate within error"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes the short-time part of the intermolecular relaxation—the librations and hydrogen-bond jumps that deviate from plain diffusion—can be taken from a classical water model and applied unchanged to the quantum-accurate and experimental structures; if that short-time behavior differs under nuclear quantum effects or at CCSD(T) quality, the computed intermolecular rate shifts.","fun_headline_variants_meta":{"raw":{"variants":["Water's 1H NMR relaxation rate predicted to 0.6%","Quantum-accurate simulation matches water NMR experiment","Intermolecular and intramolecular water NMR contributions equal","CCSD(T)-level dynamics reproduce water 1H relaxation","Theory nails water NMR relaxation rate within error"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000913,"raw_usage":{"total_tokens":3957,"prompt_tokens":1019,"completion_tokens":2938,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":635,"completion_tokens_details":{"reasoning_tokens":2858}},"tokens_in":635,"tokens_out":2938,"duration_ms":22203,"temperature":1.0,"reasoning_tokens":2858,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:23:15.564909+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute $\\Delta J_{\\mathrm{inter}}$ directly from CCMD trajectories (or from experimental-structure simulations) rather than importing the TIP4P/2005 difference function; if the total $R_1$ at 28 MHz then leaves the window $0.277$–$0.283$ s$^{-1}$, the transferability assumption fails. As a second check, field-cycling measurements above 1 GHz would test the predicted roughly 15% drop in the intermolecular contribution.","supporting_citations":[{"cited_title":"MDorado” which is available via GitHub (github.com/Paschek-Lab/MDorado). Our open source software “FreeDRelax","cited_arxiv_id":null,"evidence_quote":"Generates the CCMD trajectories at CCSD(T) accuracy with path-integral nuclear quantum effects used for structure and dynamics."},{"cited_title":"When Theory Meets Experiment: What Does it Take to Accurately Predict $^1$H NMR Dipolar Relaxation Rates in Neat Liquid Water from Theory?","cited_arxiv_id":"2411.12545","evidence_quote":"Provides the experimental $R_1 = 0.280 \\pm 0.003$ s$^{-1}$ at 28 MHz that the calculation targets."}],"review_version":1}