Development of transferable interaction potentials for water. V. Extension of the flexible, polarizable, Thole-type model potential (TTM3-F, v. 3.0) to describe the vibrational spectra of water clusters and liquid water

Implementation of Geometry-Dependent Charge Flux into the Polarizable AMOEBA+ Potential

Molecular dynamics (MD) simulations employing classical force fields (FFs) have been widely used to model molecular systems. The important ingredient of the current FFs, atomic charge, remains fixed during MD simulations despite the atomic environment or local geometry changes. This approximation hinders the transferability of the potential being used in multiple phases. Here we implement a geometry-dependent charge flux (GDCF) model into the multipole-based AMOEBA+ polarizable potential. The CF in the current work explicitly depends on the local geometry (bond and angle) of the molecule. To our knowledge, this is the first study that derives energy and force expressions due to GDCF in a multipole-based polarizable FF framework. Due to the inclusion of GDCF, the AMOEBA+ water model is noticeably improved in terms of describing the monomer properties, cluster binding/interaction energy, and a variety of liquid properties, including the infrared spectra that previous flexible water models were not able to capture.

Insights into the Gas Adsorption Mechanisms in Metal-Organic Frameworks from Classical Molecular Simulations

Classical molecular simulations can provide significant insights into the gas adsorption mechanisms and binding sites in various metal-organic frameworks (MOFs). These simulations involve assessing the interactions between the MOF and an adsorbate molecule by calculating the potential energy of the MOF-adsorbate system using a functional form that generally includes nonbonded interaction terms, such as the repulsion/dispersion and permanent electrostatic energies. Grand canonical Monte Carlo (GCMC) is the most widely used classical method that is carried out to simulate gas adsorption and separation in MOFs and identify the favorable adsorbate binding sites. In this review, we provide an overview of the GCMC methods that are normally utilized to perform these simulations. We also describe how a typical force field is developed for the MOF, which is required to compute the classical potential energy of the system. Furthermore, we highlight some of the common analysis techniques that have been used to determine the locations of the preferential binding sites in these materials. We also review some of the early classical molecular simulation studies that have contributed to our working understanding of the gas adsorption mechanisms in MOFs. Finally, we show that the implementation of classical polarization for simulations in MOFs can be necessary for the accurate modeling of an adsorbate in these materials, particularly those that contain open-metal sites. In general, molecular simulations can provide a great complement to experimental studies by helping to rationalize the favorable MOF-adsorbate interactions and the mechanism of gas adsorption.

Assessing cluster models of solvation for the description of vibrational circular dichroism spectra: synergy between static and dynamic approaches

Solvation effects are essential for defining the shape of vibrational circular dichroism (VCD) spectra.

Atlas of putative minima and low-lying energy networks of water clusters n = 3-25

We report a database consisting of the putative minima and ∼3.2 × 10⁶ local minima lying within 5 kcal/mol from the putative minima for water clusters of sizes n = 3-25 using an improved version of the Monte Carlo temperature basin paving (MCTBP) global optimization procedure in conjunction with the ab initio based, flexible, polarizable Thole-Type Model (TTM2.1-F, version 2.1) interaction potential for water. Several of the low-lying structures, as well as low-lying penta-coordinated water networks obtained with the TTM2.1-F potential, were further refined at the Møller-Plesset second order perturbation (MP2)/aug-cc-pVTZ level of theory. In total, we have identified 3 138 303 networks corresponding to local minima of the clusters n = 3-25, whose Cartesian coordinates and relative energies can be obtained from the webpage https://sites.uw.edu/wdbase/. Networks containing penta-coordinated water molecules start to appear at n = 11 and, quite surprisingly, are energetically close (within 1-3 kcal/mol) to the putative minima, a fact that has been confirmed from the MP2 calculations. This large database of water cluster minima spanning quite dissimilar hydrogen bonding networks is expected to influence the development and assessment of the accuracy of interaction potentials for water as well as lower scaling electronic structure methods (such as different density functionals). Furthermore, it can also be used in conjunction with data science approaches (including but not limited to neural networks and machine and deep learning) to understand the properties of water, nature's most important substance.

Observation of the Low-Frequency Spectrum of the Water Dimer as a Sensitive Test of the Water Dimer Potential and Dipole Moment Surfaces

Using the helium nanodroplet isolation setup at the ultrabright free-electron laser source FELIX in Nijmegen (BoHeNDI@FELIX), the intermolecular modes of water dimer in the frequency region from 70 to 550 cm^-1 were recorded. Observed bands were assigned to donor torsion, acceptor wag, acceptor twist, intermolecular stretch, donor torsion overtone, and in-plane and out-of-plane librational modes. This experimental data set provides a sensitive test for state-of-the-art water potentials and dipole moment surfaces. Theoretical calculations of the IR spectrum are presented using high-level quantum and approximate quasiclassical molecular dynamics approaches. These calculations use the full-dimensional ab initio WHHB potential and dipole moment surfaces. Based on the experimental data, a considerable increase of the acceptor switch and a bifurcation tunneling splitting in the librational mode is deduced, which is a consequence of the effective decrease in the tunneling barrier.

Quantum mechanical static dipole polarizabilities in the QM7b and AlphaML showcase databases

While density functional theory (DFT) is often an accurate and efficient methodology for evaluating molecular properties such as energies and multipole moments, this approach often yields larger errors for response properties such as the dipole polarizability (α), which describes the tendency of a molecule to form an induced dipole moment in the presence of an electric field. In this work, we provide static α tensors (and other molecular properties such as total energy components, dipole and quadrupole moments, etc.) computed using quantum chemical (QC) and DFT methodologies for all 7,211 molecules in the QM7b database. We also provide the same quantities for the 52 molecules in the AlphaML showcase database, which includes the DNA/RNA nucleobases, uncharged amino acids, several open-chain and cyclic carbohydrates, five popular pharmaceutical molecules, and 23 isomers of C8Hn. All QC calculations were performed using linear-response coupled-cluster theory including single and double excitations (LR-CCSD), a sophisticated approach for electron correlation, and the d-aug-cc-pVDZ basis set to mitigate basis set incompleteness error. DFT calculations employed the B3LYP and SCAN0 hybrid functionals, in conjunction with d-aug-cc-pVDZ (B3LYP and SCAN0) and d-aug-cc-pVTZ (B3LYP).

Pressure response of the THz spectrum of bulk liquid water revealed by intermolecular instantaneous normal mode analysis

The radial distribution functions of liquid water are known to change significantly their shape upon hydrostatic compression from ambient conditions deep into the kbar pressure regime. It has been shown that despite their eye-catching changes, the fundamental locally tetrahedral fourfold H-bonding pattern that characterizes ambient water is preserved up to about 10 kbar (1 GPa), which is the stability limit of liquid water at 300 K. The observed increase in coordination number comes from pushing water molecules into the first coordination sphere without establishing an H-bond, resulting in roughly two such additional interstitial molecules at 10 kbar. THz spectroscopy has been firmly established as a powerful experimental technique to analyze H-bonding in aqueous solutions given that it directly probes the far-infrared lineshape and thus the prominent H-bond network mode around 180 cm^-1. We, therefore, set out to assess pressure effects on the THz response of liquid water at 10 kbar in comparison to the 1 bar (0.1 MPa) reference, both at 300 K, with the aim to trace back the related lineshape changes to the structural level. To this end, we employ the instantaneous normal mode approximation to rigorously separate the H-bonding peak from the large background arising from the pronounced librational tail. By exactly decomposing the total molecular dynamics into hindered translations, hindered rotations, and intramolecular vibrations, we find that the H-bonding peak arises from translation-translation and translation-rotation correlations, which are successively decomposed down to the level of distinct local H-bond environments. Our utmost detailed analysis based on molecular pair classifications unveils that H-bonded double-donor water pairs contribute most to the THz response around 180 cm^-1, whereas interstitial waters are negligible. Moreover, short double-donor H-bonds have their peak maximum significantly shifted toward higher frequencies with respect to such long H-bonds. In conjunction with an increasing relative population of these short H-bonds versus the long ones (while the population of other water pair classes is essentially pressure insensitive), this explains not only the blue-shift of the H-bonding peak by about 20-30 cm^-1 in total from 1 bar to 10 kbar but also the filling of the shallow local minimum of the THz lineshape located in between the network peak and the red-wing of the librational band at 1 bar. Based on the changing populations as a function of pressure, we are also able to roughly estimate the pressure-dependence of the H-bond network mode and find that its pressure response and thus the blue-shifting are most pronounced at low kbar pressures.

Accurate molecular polarizabilities with coupled cluster theory and machine learning

The molecular dipole polarizability describes the tendency of a molecule to change its dipole moment in response to an applied electric field. This quantity governs key intra- and intermolecular interactions, such as induction and dispersion; plays a vital role in determining the spectroscopic signatures of molecules; and is an essential ingredient in polarizable force fields. Compared with other ground-state properties, an accurate prediction of the molecular polarizability is considerably more difficult, as this response quantity is quite sensitive to the underlying electronic structure description. In this work, we present highly accurate quantum mechanical calculations of the static dipole polarizability tensors of 7,211 small organic molecules computed using linear response coupled cluster singles and doubles theory (LR-CCSD). Using a symmetry-adapted machine-learning approach, we demonstrate that it is possible to predict the LR-CCSD molecular polarizabilities of these small molecules with an error that is an order of magnitude smaller than that of hybrid density functional theory (DFT) at a negligible computational cost. The resultant model is robust and transferable, yielding molecular polarizabilities for a diverse set of 52 larger molecules (including challenging conjugated systems, carbohydrates, small drugs, amino acids, nucleobases, and hydrocarbon isomers) at an accuracy that exceeds that of hybrid DFT. The atom-centered decomposition implicit in our machine-learning approach offers some insight into the shortcomings of DFT in the prediction of this fundamental quantity of interest.

Using a monomer potential energy surface to perform approximate path integral molecular dynamics simulation of ab initio water at near-zero added cost

It is now established that nuclear quantum motion plays an important role in determining water's hydrogen bonding, structure, and dynamics. Such effects are important to include in density functional theory (DFT) based molecular dynamics simulation of water. The standard way of treating nuclear quantum effects, path integral molecular dynamics (PIMD), multiplies the number of energy/force calculations by the number of beads required. In this work we introduce a method whereby PIMD can be incorporated into a DFT simulation with little extra cost and little loss in accuracy. The method is based on the many body expansion of the energy and has the benefit of including a monomer level correction to the DFT energy. Our method calculates intramolecular forces using the highly accurate monomer potential energy surface developed by Partridge-Schwenke, which is cheap to evaluate. Intermolecular forces and energies are calculated with DFT only once per timestep using the centroid positions. We show how our method may be used in conjunction with a multiple time step algorithm for an additional speedup and how it relates to ring polymer contraction and other schemes that have been introduced recently to speed up PIMD simulations. We show that our method, which we call "monomer PIMD", correctly captures changes in the structure of water found in a full PIMD simulation but at much lower computational cost.

On the geometric dependence of the molecular dipole polarizability in water: A benchmark study of higher-order electron correlation, basis set incompleteness error, core electron effects, and zero-point vibrational contributions

In this work, we investigate how geometric changes influence the static dipole polarizability (α) of a water molecule by explicitly computing the corresponding dipole polarizability surface (DPS) across 3125 total (1625 symmetry-unique) geometries using linear response coupled cluster theory including single, double, and triple excitations (LR-CCSDT) and the doubly augmented triple-ζ basis set (d-aug-cc-pVTZ). Analytical formulae based on power series expansions of this ab initio surface are generated using linear least-squares analysis and provide highly accurate estimates of this quantity as a function of molecular geometry (i.e., bond and angle variations) in a computationally tractable manner. An additional database, which consists of 25 representative molecular geometries and incorporates a more thorough treatment of both basis sets and core electron effects, is provided as a current benchmark for this quantity and the corresponding leading-order C ₆ dispersion coefficient. This database has been utilized to assess the importance of these effects as well as the relative accuracy that can be obtained using several quantum chemical methods and a library of density functional approximations. In addition to high-level electron correlation methods (like CCSD) and our analytical least-squares formulae, we find that the SCAN0, PBE0, MN15, and B97-2 hybrid functionals yield the most accurate descriptions of the molecular polarizability tensor in H₂O. Using first-order perturbation theory, we compute the zero-point vibrational correction to α at the CCSDT/d-aug-cc-pVTZ level and find that this correction contributes approximately 3% to the isotropic (α _iso) and nearly 50% to the anisotropic (α _aniso) polarizability values. In doing so, we find that α _iso = 9.8307 bohr³, which is in excellent agreement with the experimental value of 9.83 ± 0.02 bohr³ provided by Russell and Spackman. The DPS reported herein provides a benchmark-quality quantum mechanical estimate of this fundamental quantity of interest and should find extensive use in the development (and assessment) of next-generation force fields and machine-learning based approaches for modeling water in complex condensed-phase environments.

Unique Structures and Vibrational Spectra of Protic Ionic Liquids Confined in TiO2 Slits: The Role of Interfacial Hydrogen Bonds

The ionic liquid (IL)/titanium dioxide (TiO₂) interface exists in many application systems, such as nanomaterial synthesis, catalysis, and electrochemistry systems. The nanoscale interfacial properties in the above systems are a common issue. However, directly detecting the interfacial properties of nanoconfined ILs by experimental methods is still challenging. To help better learn about the interfacial issue, molecular dynamics simulations have been performed to explore the structures, vibration spectra, and hydrogen bond (HB) properties at the IL/TiO₂ interface. Ethylammonium nitrate (EAN) ILs confined in TiO₂ slit pores with different pore widths were studied. A unique vibrational spectrum appeared for EAN ILs confined in a 0.7 nm TiO₂ slit, and this phenomenon is related to interfacial hydrogen bonds (HBs). An analysis of the HB types indicated that the interfacial NH₃⁺ group of the cations was in an asymmetric HB environment in the 0.7 nm TiO₂ slit, which led to the disappearance of the symmetric N-H stretching mode. In addition, the significant increase in the HB strength between NH₃⁺ groups and the TiO₂ surface slowed down the stretching vibration of the N-H bond, resulting in one peak in the vibrational spectra at a lower frequency. For the first time, our simulation work establishes a molecular-level relationship between the vibrational spectrum and the local HB environment of nanoconfined ILs at the IL/TiO₂ interface, and this relationship is helpful for interface design in related systems.

Approximating Matsubara dynamics using the planetary model: Tests on liquid water and ice

Matsubara dynamics is the quantum-Boltzmann-conserving classical dynamics which remains when real-time coherences are taken out of the exact quantum Liouvillian [T. J. H. Hele et al., J. Chem. Phys. 142, 134103 (2015)]; because of a phase-term, it cannot be used as a practical method without further approximation. Recently, Smith et al. [J. Chem. Phys. 142, 244112 (2015)] developed a "planetary" model dynamics which conserves the Feynman-Kleinert (FK) approximation to the quantum-Boltzmann distribution. Here, we show that for moderately anharmonic potentials, the planetary dynamics gives a good approximation to Matsubara trajectories on the FK potential surface by decoupling the centroid trajectory from the locally harmonic Matsubara fluctuations, which reduce to a single phase-less fluctuation particle (the "planet"). We also show that the FK effective frequency can be approximated by a direct integral over these fluctuations, obviating the need to solve iterative equations. This modification, together with use of thermostatted ring-polymer molecular dynamics, allows us to test the planetary model on water (gas-phase, liquid, and ice) using the q-TIP4P/F potential surface. The "planetary" fluctuations give a poor approximation to the rotational/librational bands in the infrared spectrum, but a good approximation to the bend and stretch bands, where the fluctuation lineshape is found to be motionally narrowed by the vibrations of the centroid.

Correlation effects and many-body interactions in water clusters

Background: The quantum-chemical description of the interactions in water clusters is an essential basis for deriving accurate and physically sound models of the interaction potential for water to be used in molecular simulations. In particular, the role of many-body interactions beyond the two-body interactions, which are often not explicitly taken into account by empirical force fields, can be accurately described by quantum chemistry methods on an adequate level, e.g., random-phase approximation electron correlation methods. The relative magnitudes of the different interaction energy contributions obtained by accurate ab initio calculations can therefore provide useful insights that can be exploited to develop enhanced force field methods. Results: In line with earlier theoretical studies of the interactions in water clusters, it has been found that the main contribution to the many-body interactions in clusters with a size of up to N = 13 molecules are higher-order polarisation interaction terms. Compared to this, many-body dispersion interactions are practically negligible for all studied sytems. The two-body dispersion interaction, however, plays a significant role in the formation of the structures of the water clusters and their stability, since it leads to a distinct compression of the cluster sizes compared to the structures optimized on an uncorrelated level. Overall, the many-body interactions amount to about 13% of the total interaction energy, irrespective of the cluster size. The electron correlation contribution to these, however, amounts to only about 30% to the total many-body interactions for the largest clusters studied and is repulsive for all structures considered in this work. Conclusion: While this shows that three- and higher-body interactions can not be neglected in the description of water complexes, the electron correlation contributions to these are much smaller in comparison to the two-body electron correlation effects. Efficient quantum chemistry approaches for describing intermolecular interactions between water molecules may therefore describe higher-body interactions on an uncorrelated Hartree-Fock level without a serious loss in accuracy.

Nuclear quantum effects on the vibrational dynamics of liquid water

Based on quantum-mechanical path-integral molecular dynamics simulations, the impact of nuclear quantum effects on the vibrational and hydrogen bond dynamics in liquid water is investigated. The instantaneous fluctuations in the frequencies of the O-H stretch modes are calculated using the wavelet method of time-series analysis, while the time scales of the vibrational spectral diffusion are determined from frequency-time correlation functions, joint probability distributions, and the slope of three-pulse photon echo. We find that the inclusion of nuclear quantum effects leads not only to a redshift of the vibrational frequency distribution by around 130 cm^-1 but also to an acceleration of the vibrational dynamics by as much as 30%. In addition, quantum fluctuations also entail a significantly faster decay of correlation in the initial diffusive regime, which is in agreement with recent vibrational echo experiments.

Assessing Intermolecular Interactions in Guest-Free Clathrate Hydrate Systems

Recently, empty hydrate structures sI, sII, sH, and others have been proposed as low-density ice structures by both experimental observations and computer simulations. Some of them have been synthesized in the laboratory, which motivates further investigations on the stability of such guest-free clathrate structures. Using semiempirical and ab initio-based water models, as well as dispersion-corrected density functional theory approaches, we predict their stability, including cooperative many-body effects, in comparison with reference data from converged wave function-based DF-MP2 electronic structure calculations. We show that large basis sets and counterpoise corrections are required to improve convergence in the interaction/binding energies for such systems. Therefore, extrapolation schemes based on triple/quadruple and quadruple/quintuple ζ quality basis sets are used to reach high accuracy. Eleven different water structures corresponding to dodecahedron, edge sharing, face sharing, and fused cubes, as a part of the WATER27 database, as well as cavities from the sI, sII, and sH clathrate hydrates formed by 20, 24, 28, and 36 water molecules, are employed, and new benchmark energies are reported. Using these benchmark sets of interaction energies, we assess the performance of both analytical models and direct DFT calculations for such clathrate-like systems. In particular, seven popular water models (TIP4P/ice, TIP4P/2005, q-TIP4P/F, TTM2-F, TTM3-F, TTM4-F, and MB-pol) available in the literature, and nine density functional approximations (3 meta-GGAs, 3 hybrids, and 3 range separated functionals) are used to investigate their accuracy. By including dispersion corrections, our results show that errors in the interaction energies are reduced for most of the DFT functionals. Despite the difficulties faced by current water models and DFT functionals to accurately describe the interactions in such water systems, we found some general trends that could serve to extend their applicability to larger systems.

Symmetry-Adapted Machine Learning for Tensorial Properties of Atomistic Systems

Statistical learning methods show great promise in providing an accurate prediction of materials and molecular properties, while minimizing the need for computationally demanding electronic structure calculations. The accuracy and transferability of these models are increased significantly by encoding into the learning procedure the fundamental symmetries of rotational and permutational invariance of scalar properties. However, the prediction of tensorial properties requires that the model respects the appropriate geometric transformations, rather than invariance, when the reference frame is rotated. We introduce a formalism that extends existing schemes and makes it possible to perform machine learning of tensorial properties of arbitrary rank, and for general molecular geometries. To demonstrate it, we derive a tensor kernel adapted to rotational symmetry, which is the natural generalization of the smooth overlap of atomic positions kernel commonly used for the prediction of scalar properties at the atomic scale. The performance and generality of the approach is demonstrated by learning the instantaneous response to an external electric field of water oligomers of increasing complexity, from the isolated molecule to the condensed phase.

Strategy for Modeling the Infrared Spectra of Ion-Containing Water Drops

Hydrated ions are ubiquitous in environmental and biological media. Understanding the perturbation exerted by an ion on the water hydrogen bond network is possible in the nanodrop regime by recording vibrational spectra in the O-H bond stretching region. This has been achieved experimentally in recent years by forming gaseous ions containing tens to hundreds of water molecules and recording their infrared photodissociation spectra. In this paper, we demonstrate the capabilities of a modeling strategy based on an extension of the AMOEBA polarizable force field to implement water atomic charge fluctuations along with those of intramolecular structure along the dynamics. This supplementary flexibility of nonbonded interactions improves the description of the hydrogen bond network and, therefore, the spectroscopic response. Finite temperature IR spectra are obtained from molecular dynamics simulations by computing the Fourier transform of the dipole moment autocorrelation function. Simulations of 1-2 ns are required for extensive sampling in order to reproduce the experimental spectra. Furthermore, bands are assigned with the driven molecular dynamics approach. This method package is shown to compare successfully with experimental spectra for 11 ions in water drops containing 36-100 water molecules. In particular, band frequency shifts of the free O-H stretching modes at the cluster surface are well reproduced as a function of both ion charge and drop size.

On the aqueous solvation of AsO(OH)3vs. As(OH)3. Born–Oppenheimer molecular dynamics density functional theory cluster studies

BOMD simulations were used to reveal the hydration features of As(OH)₃ and (for the first time) AsO(OH)₃ in aqueous solution.

The long-range convergence of the energetic properties of the water monomer in bulk water at room temperature

The Interacting Quantum Atoms (IQA) energy partitioning scheme has been applied to a set of liquid water largely spherical clusters (henceforth called spheres) of up to 9 Å radius, with a maximum cluster size of 113 molecules. This constitutes half of the commonly used 216 molecules in a typical simulation box of a liquid water box, and to our knowledge is the largest analysis of this kind ever undertaken. As well as demonstrating the topological analysis of large systems, which has only recently become computationally feasible, important long range properties of liquid water are obtained. The full topological partitioning of each sphere into atomic basins is used to consider the long-range convergence of the energetic and multipolar properties of the water molecule at the centre of each sphere. It is found that the total molecular energy converges to its 9 Å value after 7 Å, which corresponds to approximately the first three solvation shells, while the molecular dipole and quadrupole moments approximately converge after 5.5 Å, which corresponds to approximately the first two solvation shells. The effect of water molecule flexibility is also considered.