Classical Interatomic Potentials Research Articles

Amorphous MoS2 from a machine learning inter-atomic potential.

Amorphous molybdenum disulfide has shown potential as a hydrogen evolution catalyst, but the origin of its high activity is unclear, as is its atomic structure. Here, we have developed a classical inter-atomic potential using the charge equilibration neural network method, and we have employed it to generate atomic models of amorphous MoS2 by melting and quenching processes. The amorphous phase contains an abundance of molybdenum and sulfur atoms in low coordination. Besides the 6-coordinated molybdenum typical of the crystalline phases, a substantial fraction displays coordinations 4 and 5. The amorphous phase is also characterized by the appearance of direct S-S bonds. Density functional theory shows that the amorphous phase is metallic, with a considerable contribution of the 4-coordinated molybdenum to the density of states at the Fermi level. S-S bonds are related to the reduction of sulfur, with the excess electrons spread over several molybdenum atoms. Moreover, S-S bond formation is associated with a distinctive broadening of the 3s states, which could be exploited for experimental characterization of the amorphous phases. The large variety of local environments and the high density of electronic states at the Fermi level may play a positive role in increasing the electrocatalytic activity of this compound.

Interatomic Potential For Carbon Based Quantum‐Technology Applications

AbstractTo optimize parameters for laser processing of quantum‐technology relevant materials, such as diamond, precise atomistic simulations of the light‐matter interaction on large scales (on the order of atoms) are essential. Classical empirical interatomic potentials are commonly employed for simulating such a large number of atoms, however they fail to accurately capture all relevant effects of light‐matter interaction. Conversely, ab initio methods like Density Functional Theory (DFT) can effectively incorporate quantum properties arising from photon excitations, but their applicability is limited to small systems containing at most approximately atoms. Consequently, bridging the gap between achieving DFT precision and handling millions of atoms necessitates the development of innovative classes of interatomic potentials. In this paper, the construction of a highly accurate interatomic potential for diamond is presented, derived from an extensive dataset of DFT calculations. The parameters of the interatomic potential depend on the electronic temperature (). The findings demonstrate that this newly developed interatomic potential can aptly describe the laser processing of diamond for nanophotonic applications, achieving accuracy comparable to ab initio methods for large systems.

Accurate machine learning force fields via experimental and simulation data fusion

Machine Learning (ML)-based force fields are attracting ever-increasing interest due to their capacity to span spatiotemporal scales of classical interatomic potentials at quantum-level accuracy. They can be trained based on high-fidelity simulations or experiments, the former being the common case. However, both approaches are impaired by scarce and erroneous data resulting in models that either do not agree with well-known experimental observations or are under-constrained and only reproduce some properties. Here we leverage both Density Functional Theory (DFT) calculations and experimentally measured mechanical properties and lattice parameters to train an ML potential of titanium. We demonstrate that the fused data learning strategy can concurrently satisfy all target objectives, thus resulting in a molecular model of higher accuracy compared to the models trained with a single data source. The inaccuracies of DFT functionals at target experimental properties were corrected, while the investigated off-target properties were affected only mildly and mostly positively. Our approach is applicable to any material and can serve as a general strategy to obtain highly accurate ML potentials.

Comparison of Interatomic Potentials for Modeling Defects in Graphene Using Molecular Dynamics

Comparison of Interatomic Potentials for Modeling Defects in Graphene Using Molecular Dynamics

Spline-based neural network interatomic potentials: Blending classical and machine learning models

While machine learning (ML) interatomic potentials (IPs) are able to achieve accuracies nearing the level of noise inherent in the first-principles data to which they are trained, it remains to be shown if their increased complexities are strictly necessary for constructing high-quality IPs. In this work, we introduce a new MLIP framework which blends the simplicity of spline-based MEAM (s-MEAM) potentials with the flexibility of a neural network (NN) architecture. The proposed framework, which we call the spline-based neural network potential (s-NNP), is a simplified version of the traditional NNP that can be used to describe complex datasets in a computationally efficient manner. We demonstrate how this framework can be used to probe the boundary between classical and ML IPs, highlighting the benefits of key architectural changes. Furthermore, we show that using spline filters for encoding atomic environments results in a readily interpreted embedding layer which can be coupled with modifications to the NN to incorporate expected physical behaviors and improve overall interpretability. Finally, we test the flexibility of the spline filters, observing that they can be shared across multiple chemical systems in order to provide a convenient reference point from which to begin performing cross-system analyses.

Molecular dynamics simulations of magnesium aluminosilicate glass structure: High‐coordinated alumina and oxygen tricluster formation

AbstractPrevious research has shown a consistent discrepancy in the reported structure of alkaline earth aluminosilicate glasses using molecular dynamics (MD) simulations versus nuclear magnetic resonance (NMR) experiments. Past MD results have consistently shown less than 5% five‐coordinated Al units (Al[5]) in peraluminous glass compositions, but with high fractions of triple‐bonded oxygens (TBO, i.e., triclusters). Experimental results have shown a high fraction of Al[5] with no direct evidence for TBO. One of the main criticisms associated with high TBO content found in MD‐generated glass structures is the use of classical interatomic potentials. To investigate this issue, we analyze the formation of both TBO and Al[5] using three independently developed potentials with varying silica content and [Al2O3]/[MgO] ratios for the magnesium aluminosilicate (MAS) system. We specifically choose compositions with high ratios of alumina to magnesium oxide as this region is not as commonly explored. Results indicate that Al[5] charge compensates the Al network in metaluminous compositions (compositions with more Mg than Al) while both TBO and Al[5] are prevalent in peraluminous ranges (high Al content compositions) to charge balance Al units. From the literature, NMR experiments report MAS glasses with varying Al[5] fractions and show significant differences for the same reported compositions. When comparing MD results from this work, the fraction of calculated Al[5] is within the experimental variation found in the literature. This indicates that classical potentials can accurately capture alumina environments and that both Al[5] and TBO can coexist in relatively high fractions. From the consistency in our results, we conclude that TBOs are inherent to the aluminosilicate glass system and are not simulation artifacts.

Machine-Learning Accelerated First-Principles Accurate Modeling of the Solid-Liquid Phase Transition in MgO under Mantle Conditions.

While accurate measurements of MgO under extreme high-pressure conditions are needed to understand and model planetary behavior, these studies are challenging from both experimental and computational modeling perspectives. Herein, we accelerate density functional theory (DFT) accurate calculations using deep neural network potentials (DNPs) trained over multiple phases and study the melting behavior of MgO via the two-phase coexistence (TPC) approach at 0-300 GPa and ≤9600 K. The resulting DNP-TPC melting curve is in excellent agreement with existing experimental studies. We show that the mitigation of finite-size effects that typically skew the predicted melting temperatures in DFT-TPC simulations in excess of several hundred kelvin requires models with ∼16 000 atoms and >100 ps molecular dynamics trajectories. In addition, the DNP can successfully describe MgO metallization well at increased pressures that are captured by DFT but missed by classical interatomic potentials.

N-body potential for simulating lattice defects and diffusion in copper

We present a new classical interatomic potential for molecular dynamics simulations of copper. The potential was developed within the N-body approach, which, in addition to pair and many-body interactions, can take into account three-body ones with the required precision. Our potential accurately predicts a wide range of properties of copper: structural, elastic, point and planar, and thermal such as melting point. The latter is essential for the proper reproduction of the homologous temperature and comparison of the calculated results with the corresponding experimental data, such as diffusion coefficients. The potential predicts the intrinsic stacking fault energy within the range of experimental measurements, which indicates its ability to correctly describe the deformation behavior of copper. In this work, we performed simulations of lattice and grain-boundary diffusion of copper with our potential. It shows a moderate overestimation of the lattice diffusivity with a deviation less than one order of magnitude. To simulate the grain boundary diffusion, we built a general-type grain boundary employing the original method. The calculated grain-boundary diffusivity is in remarkable agreement with the literature data and almost coincides with them at higher temperatures in the considered temperature range. The constructed potential can be effectively used for molecular dynamics simulations of defect properties and diffusion phenomena and composing the interatomic potentials within the N-body approach for alloys containing copper.

Formation Energy Profiles of Oxygen Vacancies at Grain Boundaries in Perovskite‐Type Electroceramics

Oxygen vacancy formation energies play a major role in the electric field‐assisted abnormal grain growth of technologically relevant polycrystalline perovskite phases. The underlying effect on the atomic scale is assumed to be a redistribution of cationic and anionic point defects between grain boundaries (GBs) and the bulk interior regions of the grains due to different defect formation energies in the structurally different regions, accompanied by the formation of space charge zones. Using atomistic calculations based on classical interatomic potentials, optimized structures of the symmetric tilt GBs Σ5(210)[001] and Σ5(310)[001], and of the asymmetric tilt GB (430)[001]||(100)[001] in the electroceramic perovskite materials SrTiO3, BaTiO3, and BaZrO3, are derived and discussed. Profiles of oxygen vacancy formation energies across those GBs are presented and their dependence on composition and GB type is discussed.

Evaluating the transferability of machine-learned force fields for material property modeling

Machine-learned force fields have generated significant interest in recent years as a tool for molecular dynamics (MD) simulations, with the aim of developing accurate and efficient models that can replace classical interatomic potentials. However, before these models can be confidently applied to materials simulations, they must be thoroughly tested and validated. The existing tests on the radial distribution function and mean-squared displacements are insufficient in assessing the transferability of these models. Here we present a more comprehensive set of benchmarking tests for evaluating the transferability of machine-learned force fields. We use a graph neural network (GNN)-based force field coupled with the OpenMM package to carry out MD simulations for Argon as a test case. Our tests include computational X-ray photon correlation spectroscopy (XPCS) signals, which capture the density fluctuation at various length scales in the liquid phase, as well as phonon density-of-states in the solid phase and the liquid-solid phase transition behavior. Our results show that the model can accurately capture the behavior of the solid phase only when the configurations from the solid phase are included in the training dataset. This underscores the importance of appropriately selecting the training data set when developing machine-learned force fields. The tests presented in this work provide a necessary foundation for the development and application of machine-learned force fields for materials simulations.

The Highly Accurate Interatomic Potential of CsPbBr3 Perovskite with Temperature Dependence on the Structure and Thermal Properties

CsPbBr3 perovskite has excellent optoelectronic properties and many important application prospects in solar cells, photodetectors, high-energy radiation detectors and other fields. For this kind of perovskite structure, to theoretically predict its macroscopic properties through molecular dynamic (MD) simulations, a highly accurate interatomic potential is first necessary. In this article, a new classical interatomic potential for CsPbBr3 was developed within the framework of the bond-valence (BV) theory. The optimized parameters of the BV model were calculated through first-principle and intelligent optimization algorithms. Calculated lattice parameters and elastic constants for the isobaric-isothermal ensemble (NPT) by our model are in accordance with the experimental data within a reasonable error and have a higher accuracy than the traditional Born-Mayer (BM) model. In our potential model, the temperature dependence of CsPbBr3 structural properties, such as radial distribution functions and interatomic bond lengths, was calculated. Moreover, the temperature-driven phase transition was found, and the phase transition temperature was close to the experimental value. The thermal conductivities of different crystal phases were further calculated, which agreed with the experimental data. All these comparative studies proved that the proposed atomic bond potential is highly accurate, and thus, by using this interatomic potential, the structural stability and mechanical and thermal properties of pure inorganic halide and mixed halide perovskites can be effectively predicted.

Universality of grain boundary phases in fcc metals: Case study on high-angle [111] symmetric tilt grain boundaries

Grain boundaries often exhibit ordered atomic structures. Increasing amounts of evidence have been provided by transmission electron microscopy and atomistic computer simulations that different stable and metastable grain boundary structures can occur. Meanwhile, theories to treat them thermodynamically as grain boundary phases have been developed. Whereas atomic structures were identified at particular grain boundaries for particular materials, it remains an open question if these structures and their thermodynamic excess properties are material specific or generalizable to, e.g., all fcc metals. In order to elucidate that question, we use atomistic simulations with classical interatomic potentials to investigate a range of high-angle [111] symmetric tilt grain boundaries in Ni, Cu, Pd, Ag, Au, Al, and Pb. We could indeed find two families of grain boundary phases in all of the investigated grain boundaries, which cover most of the standard fcc materials. Where possible, we compared the atomic structures to atomic-resolution electron microscopy images and found that the structures match. This poses the question if the grain boundary phases are simply the result of sphere-packing geometry or if material-specific bonding physics play a role. We tested this using simple model pair potentials and found that medium-ranged interactions are required to reproduce the atomic structures, while the more realistic material models mostly affect the grain boundary (free) energy. In addition to the structural investigation, we also report the thermodynamic excess properties of the grain boundaries, explore how they influence the thermodynamic stability of the grain boundary phases, and detail the commonalities and differences between the materials.

Automated determination of grain boundary energy and potential-dependence using the OpenKIM framework

We present a systematic methodology, built within the Open Knowledgebase of Interatomic Models (OpenKIM) framework (https://openkim.org), for quantifying properties of grain boundaries (GBs) for arbitrary interatomic potentials (IPs), GB character, and lattice structure and species. The framework currently generates results for symmetric tilt GBs in cubic materials, but can be readily extended to other types of boundaries. In this paper, GB energy data are presented that were generated automatically for Al, Ni, Cu, Fe, and Mo with 225 IPs; the system is installed on openkim.org and will continue to generate results for all new IPs uploaded to OpenKIM. The results from the atomistic calculations are compared to the lattice matching model, which is a semi-analytic geometric model for approximating GB energy. It is determined that the energy predicted by all IPs (that are stable for the given boundary type) correlate closely with the energy from the model, up to a multiplicative factor. It thus is concluded that the qualitative form of the GB energy versus tilt angle is dominated more by geometry than the choice of IP, but that the IP can strongly affect the energy level. The spread in GB energy predictions across the ensemble of IPs in OpenKIM provides a measure of uncertainty for GB energy predictions by classical IPs.

Variable thermal transport in black, blue, and violet phosphorene from extensive atomistic simulations with a neuroevolution potential

Phosphorus has diverse chemical bonds, and even in its two-dimensional form, there are three stable allotropes: black phosphorene (Black-P), blue phosphorene (Blue-P), and violet phosphorene (Violet-P). Due to the complexity of these structures, no efficient and accurate classical interatomic potential has been developed for them. In this paper, we develop an efficient machine-learned neuroevolution potential model for these allotropes and apply it to study thermal transport in them via extensive molecular dynamics (MD) simulations. Based on the homogeneous nonequilibrium MD method, the thermal conductivities are predicted to be 12.5±0.2 (Black-P in armchair direction), 78.4±0.4 (Black-P in zigzag direction), 128±3 (Blue-P), and 2.36±0.05 (Violet-P) Wm−1K−1. The underlying reasons for the significantly different thermal conductivity values in these allotropes are unraveled through spectral decomposition, phonon eigenmodes, and phonon participation ratio. Under external tensile strain, the thermal conductivity in black-P and violet-P are finite, while that in blue-P appears unbounded due to the linearization of the flexural phonon dispersion that increases the phonon mean free paths in the zero-frequency limit.

Classical and machine learning interatomic potentials for BCC vanadium

BCC transition metals (TMs) exhibit complex temperature and strain-rate dependent plastic deformation behavior controlled by individual crystal lattice defects. Classical empirical and semiempirical interatomic potentials have limited capability in modeling defect properties such as the screw dislocation core structures and Peierls barriers in the BCC structure. Machine learning (ML) potentials, trained on DFT-based datasets, have shown some successes in reproducing dislocation core properties. However, in group VB TMs, the most widely used DFT functionals produce erroneous shear moduli ${C}_{44}$ which are undesirably transferred to machine-learning interatomic potentials, leaving current ML approaches unsuitable for this important class of metals and alloys. Here, we develop two interatomic potentials for BCC vanadium (V) based on (i) an extension of the partial electron density and screening parameter in the classical semiempirical modified embedded-atom method (XMEAM-V) and (ii) a recent hybrid descriptor in the ML Deep Potential framework (DP-HYB-V). We describe distinct features in these two disparate approaches, including their dataset generation, training procedure, weakness and strength in modeling lattice and defect properties in BCC V. Both XMEAM-V and DP-HYB-V reproduce a broad range of defect properties (vacancy, self-interstitials, surface, dislocation) relevant to plastic deformation and fracture. In particular, XMEAM-V reproduces nearly all mechanical and thermodynamic properties at DFT accuracies and with ${C}_{44}$ near the experimental value. XMEAM-V also naturally exhibits the anomalous slip at 77 K widely observed in group VB and VIB TMs and outperforms all existing, publically available interatomic potentials for V. The XMEAM thus provides a practical path to developing accurate and efficient interatomic potentials for nonmagnetic BCC TMs and possibly multiprincipal element TM alloys.

Effect of crystal defects in iron on carbon diffusivity: Analytical model married to atomistics

We devise a scheme for simulating diffusion of interstitial solutes in metals when the diffusivity is modified by the presence of defects. The method is quite general and may be used to model the effects of more than one defect in combination. In addition, it is amenable to being described by a simple analytic model fitted to a small number of parameters that can be determined by computer experiments. Using classical interatomic potentials and molecular dynamics, we calculate the diffusivity of carbon in iron and study the effects of a substitutional alloying element and an edge dislocation. Some of our results are validated against published experimental findings and previous simulations, and we are able to apply our theoretical model to rationalize our results and to show how, once the model is fitted to the computer experiments, it may be used in a predictive manner.

Developing Classical Interatomic Potentials for Solid Electrolytes

ADVERTISEMENT RETURN TO ISSUEViewpointNEXTDeveloping Classical Interatomic Potentials for Solid ElectrolytesZhewen ZhuZhewen ZhuSchool of Materials Science and Engineering, Zhejiang University, Hangzhou, Zhejiang 310027, ChinaSchool of Engineering, Westlake University, Hangzhou, Zhejiang 300024, ChinaMore by Zhewen ZhuView Biography and Yizhou Zhu*Yizhou ZhuResearch Center for Industries of the Future, Westlake University, 600 Dunyu Road, Hangzhou, Zhejiang 310030, ChinaSchool of Engineering, Westlake University, Hangzhou, Zhejiang 300024, China*Email: [email protected]More by Yizhou ZhuView Biographyhttps://orcid.org/0000-0002-5819-7657Cite this: Acc. Mater. Res. 2022, 3, 11, 1101–1105Publication Date (Web):October 12, 2022Publication History Received13 September 2022Published online12 October 2022Published inissue 25 November 2022https://doi.org/10.1021/accountsmr.2c00184Copyright © 2022 Accounts of Materials Research. Co-published by ShanghaiTech University and American Chemical Society. All rights reserved.RIGHTS & PERMISSIONSArticle Views651Altmetric-Citations1LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InReddit Read OnlinePDF (2 MB) Get e-AlertsSUBJECTS:Ab initio molecular dynamics,Diffusion,Materials,Potential energy,Solid electrolytes Get e-Alerts

Developing a force field for the Ba1−xCaxZrO3 ferroelectric alloy: Prediction of a ferroelectric superlattice structure

In this study, we develop a classical interatomic potential for the ${\mathrm{Ba}}_{1\ensuremath{-}x}{\mathrm{Ca}}_{x}{\mathrm{ZrO}}_{3}$ alloy based on the bond-valence theory. The bond-valence model enables rapid and large-scale molecular dynamics simulations of alloys with variable compositions. Molecular dynamics simulations based on this force field can reproduce the experimentally observed composition-dependent dielectric responses and lattice constants very well, indicating the validity and robustness of this force field. Based on molecular dynamics simulations, we demonstrate that ${\mathrm{Ba}}_{1\ensuremath{-}x}{\mathrm{Ca}}_{x}{\mathrm{ZrO}}_{3}$ can adopt a ferroelectric phase under a tensile strain, while the ${\mathrm{Ba}}_{1\ensuremath{-}x}{\mathrm{Ca}}_{x}{\mathrm{ZrO}}_{3}$ solid solution under zero strain is paraelectric. In addition, we find that strain can invoke a phase transition from antiferroelectric to ferroelectric in the ${\mathrm{BaZrO}}_{3}/{\mathrm{CaZrO}}_{3}$ superlattice. These results successfully explain previous experimental results and first-principles calculations. This interatomic potential provides a powerful tool for simulating the physical properties of perovskite alloys at nanoscale and thus has great significance in understanding the underlying physical mechanisms and designing functional materials.

Investigation of the adhesive and abrasive wear mechanisms at the atomic scale using molecular dynamic simulations

In this study, the effect of adhesion on material removal is numerically investigated at the nanoscale level. In this regard, Molecular Dynamics (MD) simulations are conducted to distinguish the contribution of adhesive and abrasive wear mechanisms during a scratching process in terms of the degree of interfacial adhesion and scratching depth. The numerical model simulates the scratching of a flat workpiece made of a single crystalline aluminum using a rigid conical indenter with a blunted spherical tip at four different indentation depths (i.e. 0, 3, 7, and 10 Å). The classical Leonard-Jones interatomic potential is used to mimic the degree of adhesion by varying the adhesion parameter between 5–70% of the aluminum bonding energy. It is shown that, at shallow scratching depths, the contribution of adhesive work to the frictional work is much larger than the ploughing work. On the contrary, the ratio of adhesive to ploughing work reduces by increasing the scratching depth.

Bayesian, frequentist, and information geometric approaches to parametric uncertainty quantification of classical empirical interatomic potentials

In this paper, we consider the problem of quantifying parametric uncertainty in classical empirical interatomic potentials (IPs) using both Bayesian (Markov Chain Monte Carlo) and frequentist (profile likelihood) methods. We interface these tools with the Open Knowledgebase of Interatomic Models and study three models based on the Lennard-Jones, Morse, and Stillinger-Weber potentials. We confirm that IPs are typically sloppy, i.e., insensitive to coordinated changes in some parameter combinations. Because the inverse problem in such models is ill-conditioned, parameters are unidentifiable. This presents challenges for traditional statistical methods, as we demonstrate and interpret within both Bayesian and frequentist frameworks. We use information geometry to illuminate the underlying cause of this phenomenon and show that IPs have global properties similar to those of sloppy models from fields, such as systems biology, power systems, and critical phenomena. IPs correspond to bounded manifolds with a hierarchy of widths, leading to low effective dimensionality in the model. We show how information geometry can motivate new, natural parameterizations that improve the stability and interpretation of uncertainty quantification analysis and further suggest simplified, less-sloppy models.