A 2016 defect-resolved analysis of ab initio liquid water (Gasparotto, Hassanali and Ceriotti) established that coordination defects of the hydrogen-bond network cluster non-randomly, that under-coordinated environments sustain a weak but still directional hydrogen-bond mode, and that defect populations are qualitatively robust across electronic-structure approximations — but the study was deliberately static: the interaction-energy components stabilizing the weak motifs and defect clusters were not decomposed, and hydrogen-bond kinetics were not analyzed. The open problem is to complete this picture at a first-principles level of description: quantify the enthalpic stability and energetic components of the weak directional hydrogen-bond motifs and clustered defect configurations relative to regular tetrahedral environments; determine from time-resolved trajectories how defect clustering controls hydrogen-bond breaking and formation rates and the mechanisms of collective network rearrangement; and establish statistically converged defect populations whose differences between electronic-structure treatments can be judged significant. Later work has advanced the flanks (orbital-level energy analysis of nearby-bond correlations, collective jump dynamics in classical models, machine-learned-potential sampling at hybrid-DFT accuracy, charged and high-pressure defects), but the defect-resolved energetics and kinetics in first-principles water themselves remain undone.

Background

In liquid water each molecule forms a fluctuating network of hydrogen bonds (H bonds), the attractive interactions in which a hydrogen atom covalently bound to one oxygen interacts with a lone pair of a neighboring oxygen. In ice Ih every molecule donates two H bonds and accepts two (a locally tetrahedral, two-donor two-acceptor environment). A coordination defect is a molecule whose donated and accepted H-bond counts deviate from this reference; because every H bond has exactly one donor and one acceptor, defects cannot appear alone but occur in correlated pairs, for example an over-coordinated molecule (two donated, three accepted H bonds) adjacent to under-coordinated ones. These defects and their fluctuations are central to water's static and time-dependent anomalies. Ab initio molecular dynamics (AIMD) computes the interatomic forces on the fly from density-functional theory (DFT), avoiding empirical parametrization but limiting trajectory length; its accuracy depends on the exchange-correlation functional (including whether a fraction of exact Hartree-Fock exchange is used, as in hybrid functionals), on the treatment of van der Waals dispersion (for example the empirical D3 correction scheme), and on whether the nuclei are treated as classical particles or quantum-mechanically via path-integral simulation (nuclear quantum effects, NQE). The study that defines this problem (Gasparotto, Hassanali and Ceriotti, 2016) analyzed replica-exchange AIMD of 128 water molecules under varied modeling conditions — replica exchange (parallel tempering) being a sampling scheme in which several copies of the system simulated at different temperatures periodically swap configurations so that low-temperature trajectories can overcome energy barriers — using the PAMM approach (probabilistic analysis of molecular motifs), a data-driven H-bond definition in which each donor-hydrogen-acceptor triplet is assigned a posterior H-bond probability obtained by fitting a Gaussian mixture model to the distribution of structural fingerprints in the simulation data. They found that defects cluster with characteristic propensities, that under-coordinated environments sustain strong angular correlations traceable to a weaker but still directional H-bonding mode related to the lattice of ice VIII (a high-pressure crystalline ice phase built from two interpenetrating hydrogen-bonded sublattices) expanded to the density of ambient water, and that the qualitative defect structure is remarkably robust across dispersion corrections, hybrid exchange and NQE — although dispersion corrections strongly affect the oxygen-oxygen radial distribution function (RDF), with evidence that this apparent success rests partly on a cancellation of errors between many-body terms. The authors were explicit that their inference was drawn entirely from static, time-independent structural correlation functions. The natural complementary tool is energy decomposition analysis (EDA), which partitions the interaction energy between molecular fragments into physically interpretable components — electrostatics, exchange repulsion, dispersion or correlation, and induction/charge transfer; a condensed-phase DFT implementation based on maximally localized Wannier orbitals (spatially localized orbitals that attach electron pairs to individual molecules or bonds, making fragment-by-fragment energy bookkeeping possible in a periodic liquid) had already been used by Kühne and Khaliullin (2013, 2014) to reveal that most water molecules in DFT-simulated liquid water carry one strong and one substantially weaker H bond on their donor or acceptor side (the local asymmetry of the network). The open question formulated here is the completion that the 2016 study itself deferred: an energetic and kinetic, not merely structural, account of the defective H-bond network in first-principles water.

Problem Statement

The problem, as deferred by the source study and still only partially addressed by later literature, is to give the defective hydrogen-bond network of first-principles liquid water a quantitative energetic and kinetic characterization, jointly serving one objective: replacing the static structural-correlation picture of coordination defects with a full energetic and dynamical account. The characterization is one objective with three jointly required components. Specifically: (1) Energetics — determine which interaction-energy components stabilize the weak directional hydrogen-bond motifs and the correlated defect-pair and defect-cluster configurations identified by defect-resolved structural analysis, and the resulting enthalpic stabilization relative to regular tetrahedral environments, so that the entropic-enthalpic balance of network fluctuations can be rationalized microscopically. (2) Kinetics — from time-resolved first-principles trajectories, determine how the local defect environment controls hydrogen-bond breaking and formation: defect-state transition rates and hydrogen-bond lifetimes as a function of the clustering environment, the degree to which rearrangements are collective rather than localized, and the connection of these rates to transport and reorientational properties. (3) Statistical robustness — establish defect-population and defect-correlation statistics that are converged with respect to trajectory length and system size, with uncertainties small enough that differences between electronic-structure treatments (functional, dispersion treatment, nuclear quantum effects) can be judged significant or not, in particular for the minority defect classes. The task is scoped to the level of description of the source study — ab initio (first-principles) molecular dynamics, including machine-learned potentials demonstrated to reproduce first-principles accuracy, which enable the sampling lengths the original authors could not afford — rather than empirical classical models.

References

  1. Piero Gasparotto, Ali A. Hassanali, Michele Ceriotti (2016). Probing Defects and Correlations in the Hydrogen-Bond Network of ab Initio Water. Journal of Chemical Theory and Computation 12(4), 1953-1964. DOI: 10.1021/acs.jctc.5b01138. URL: https://doi.org/10.1021/acs.jctc.5b01138 (preprint arXiv:1602.03577, https://arxiv.org/abs/1602.03577)
  2. Piero Gasparotto, Michele Ceriotti (2014). Recognizing molecular patterns by machine learning: An agnostic structural definition of the hydrogen bond. The Journal of Chemical Physics 141, 174110. DOI: 10.1063/1.4900655. URL: https://doi.org/10.1063/1.4900655
  3. Thomas D. Kühne, Rustam Z. Khaliullin (2013). Electronic signature of the instantaneous asymmetry in the first coordination shell of liquid water. Nature Communications 4, 1450. DOI: 10.1038/ncomms2459. URL: https://doi.org/10.1038/ncomms2459
  4. Thomas D. Kühne, Rustam Z. Khaliullin (2014). Nature of the Asymmetry in the Hydrogen-Bond Networks of Hexagonal Ice and Liquid Water. Journal of the American Chemical Society 136(9), 3395-3399. DOI: 10.1021/ja411161a. URL: https://doi.org/10.1021/ja411161a
  5. Yifei Shi, Hayden Scheiber, Rustam Z. Khaliullin (2018). Contribution of the Covalent Component of the Hydrogen-Bond Network to the Properties of Liquid Water. The Journal of Physical Chemistry A 122(37), 7482-7490. DOI: 10.1021/acs.jpca.8b06857. URL: https://doi.org/10.1021/acs.jpca.8b06857
  6. Yonghwan Yun, Rustam Z. Khaliullin, Yousung Jung (2022). Correlated Local Fluctuations in the Hydrogen Bond Network of Liquid Water. Journal of the American Chemical Society 144(29), 13127-13136. DOI: 10.1021/jacs.2c02362. URL: https://doi.org/10.1021/jacs.2c02362
  7. Robert A. DiStasio Jr., Biswajit Santra, Zhaofeng Li, Xifan Wu, Roberto Car (2014). The individual and collective effects of exact exchange and dispersion interactions on the ab initio structure of liquid water. The Journal of Chemical Physics 141, 084502. DOI: 10.1063/1.4893377. URL: https://doi.org/10.1063/1.4893377
  8. Francesco Ambrosio, Giacomo Miceli, Alfredo Pasquarello (2016). Structural, Dynamical, and Electronic Properties of Liquid Water: A Hybrid Functional Study. The Journal of Physical Chemistry B 120(30), 7456-7470. DOI: 10.1021/acs.jpcb.6b03876. URL: https://doi.org/10.1021/acs.jpcb.6b03876
  9. Ondrej Marsalek, Thomas E. Markland (2017). Quantum Dynamics and Spectroscopy of Ab Initio Liquid Water: The Interplay of Nuclear and Electronic Quantum Effects. The Journal of Physical Chemistry Letters 8(7), 1545-1551. DOI: 10.1021/acs.jpclett.7b00391. URL: https://doi.org/10.1021/acs.jpclett.7b00391
  10. Michael D. LaCount, François Gygi (2019). Ensemble first-principles molecular dynamics simulations of water using the SCAN meta-GGA density functional. The Journal of Chemical Physics 151, 164107. DOI: 10.1063/1.5124957. URL: https://doi.org/10.1063/1.5124957
  11. Yi Yao, Yosuke Kanai (2020). Temperature dependence of nuclear quantum effects on liquid water via artificial neural network model based on SCAN meta-GGA functional. The Journal of Chemical Physics 153, 064502. DOI: 10.1063/5.0012815. URL: https://doi.org/10.1063/5.0012815
  12. Cecilia Herrero, Michela Pauletti, Gabriele Tocci, Marcella Iannuzzi, Laurent Joly (2022). Connection between water's dynamical and structural properties: Insights from ab initio simulations. Proceedings of the National Academy of Sciences 119(21), e2121641119. DOI: 10.1073/pnas.2121641119. URL: https://doi.org/10.1073/pnas.2121641119
  13. Adu Offei-Danso, Uriel N. Morzan, Alex Rodriguez, Ali Hassanali, Asja Jelic (2023). The collective burst mechanism of angular jumps in liquid water. Nature Communications 14, 1345. DOI: 10.1038/s41467-023-37069-9. URL: https://doi.org/10.1038/s41467-023-37069-9
  14. Nicolás A. Loubet, Alejandro R. Verde, Sebastián R. Accordino, Laureano M. Alarcón, Gustavo A. Appignanesi (2024). Role of hydrogen-bond coordination defects in the structural relaxation of supercooled water. Physical Review E 110, 054601. DOI: 10.1103/PhysRevE.110.054601. URL: https://doi.org/10.1103/PhysRevE.110.054601
  15. Solana Di Pino, Edward Danquah Donkor, Verónica M. Sánchez, Alex Rodriguez, Giuseppe Cassone, Damian Scherlis, Ali Hassanali (2023). ZundEig: The Structure of the Proton in Liquid Water from Unsupervised Learning. The Journal of Physical Chemistry B 127(45), 9822-9832. DOI: 10.1021/acs.jpcb.3c06078. URL: https://doi.org/10.1021/acs.jpcb.3c06078
  16. Austin O. Atsango, Tobias Morawietz, Ondrej Marsalek, Thomas E. Markland (2023). Developing machine-learned potentials to simultaneously capture the dynamics of excess protons and hydroxide ions in classical and path integral simulations. The Journal of Chemical Physics 159, 064101. DOI: 10.1063/5.0162066. URL: https://doi.org/10.1063/5.0162066
  17. Renxi Liu, Mohan Chen (2023). Characterization of the Hydrogen-Bond Network in High-Pressure Water by Deep Potential Molecular Dynamics. Journal of Chemical Theory and Computation 19(16), 5602-5608. DOI: 10.1021/acs.jctc.3c00445. URL: https://doi.org/10.1021/acs.jctc.3c00445
  18. Federico Giberti, Ali A. Hassanali (2017). The excess proton at the air-water interface: The role of instantaneous liquid interfaces. The Journal of Chemical Physics 146, 164701. DOI: 10.1063/1.4986082. URL: https://doi.org/10.1063/1.4986082
  19. Hendrik Vondracek, Sho Imoto, Lukas Knake, Gerhard Schwaab, Dominik Marx, Martina Havenith (2019). Hydrogen-Bonding in Liquid Water at Multikilobar Pressures. The Journal of Physical Chemistry B 123(36), 7748-7753. DOI: 10.1021/acs.jpcb.9b06821. URL: https://doi.org/10.1021/acs.jpcb.9b06821
  20. Lilian Najm Alsayed, Florian Pabst, Giuseppe Cassone, Ali Hassanali, Fabio Novelli (2026). Hydrogen-Bond Donor-Acceptor Imbalance in Low-Frequency Terahertz Water Spectra. Preprint arXiv:2607.06780. URL: https://arxiv.org/abs/2607.06780