Machine Learning Interatomic Potentials Research Articles

A machine learning interatomic potential for high entropy alloys

High entropy alloys (HEAs) possess a vast compositional space, providing exciting prospects for tailoring material properties yet also presenting challenges in their rational design. Efficiently achieving a well-designed HEA often necessitates the aid of atomistic simulations, which rely on the availability of high-quality interatomic potentials. However, such potentials for most HEA systems are missing due to the complex interatomic interaction. To fundamentally resolve the challenge of the rational design of HEAs, we propose a strategy to build a machine learning (ML) interatomic potential for HEAs and demonstrate this strategy using CrFeCoNiPd as a model material. The fully trained ML model can achieve remarkable prediction precision (>0.92 R2) for atomic forces, comparable to the ab initio molecular dynamics (AIMD) simulations. To further validate the accuracy of the ML model, we implement the ML potential for CrFeCoNiPd in parallel molecular dynamics (MD) code. The MD simulations can predict the lattice constant (1 % error) and stacking fault energy (10 % error) of CrFeCoNiPd HEAs with high accuracy compared to experimental results. Through systematic MD simulations, for the first time, we reveal the atomic-scale deformation mechanisms associated with the stacking fault formation and dislocation cross-slips in CrFeCoNiPd HEAs under uniaxial compression, which are consistent with experimental observations. This study can help elucidate the underlying deformation mechanisms that govern the exceptional performance of CrFeCoNiPd HEAs. The strategy to establish ML interatomic potentials could accelerate the rational design of new HEAs with desirable properties.

Dual-channel phonon transport leads to low thermal conductivity in pyrochlore La2Hf2O7

The pyrochlore oxide La2Hf2O7 (LHO) has attracted extensive attention in thermal management applications. Accurate determination and comprehensive understanding of thermal conductivity (κ) stand as imperative prerequisites for the effective utilization and manipulation of LHO in various applications. In this study, with the assistance of machine learning interatomic potential, we investigate the microscopic mechanisms of phonon thermal transport in LHO crystals. Our findings show that the thermal transport in LHO defies a precise explication through the conventional framework of the Boltzmann transport equation (BTE) due to the emergence of diffusive phonon modes, engendered by strong anharmonicity. Using the state-of-the-art dual-phonon model incorporating both propagating and diffusive modes, we successfully reproduce the experimentally measured results in both magnitude and temperature dependence. It is also shown that the thermal conduction in LHO crystals is dominated by the propagating modes at low temperatures while by the diffusive modes at high temperatures, which synergistically lead to a low κ for LHO crystals. The novel insights into phonon thermal transport provide a stepping stone for regulating the thermal transport properties of LHO crystals.

Thermoelectric properties of lead-free anti-perovskites X3BN (B = Bi, Sb, X = Mg, Ca, Sr): A theoretical study based on first-principles calculations and machine learning interatomic potential

The search for non-toxic and low-cost perovskites with high thermoelectric performance is still a challenge despite low thermal conductivity. The thermoelectric properties of nitride anti-perovskites X3BN (B = Bi, Sb, X = Mg, Ca, Sr) with a cubic structure were investigated using density functional theory and machine learning interatomic potential. The low Debye temperature and thermal conductivity were obtained due to strong lattice anharmonicity, and the phonon vibration modes were also analyzed. The high band degeneracy and suitable bandgap lead to a large power factor. The maximum power factor is 7.54 mW/mk2 for Mg3BiN at 900 K. We obtained a maximum ZT of 1.49 for p-type Sr3BiN, and it is 1.22 for n-type doping at 900 K. The ZT for Mg3BiN is 1.18 and 1.19 for p-type and n-type doping, respectively. Our calculations reveal that these anti-perovskites are excellent materials for non-toxic and low-cost thermoelectric applications.

Neural network molecular dynamics study of LiGe2(PO4)3: Investigation of structure

Machine learning interatomic potential for LiGe2(PO4)3 in a form of a set of neural networks (DeePMD-model) was trained on DFT data. DFT simulations in GGA PBE approximation were performed for ordered and disordered LiGe2(PO4)3 supercell with 432 atoms which allowed the developed potential to describe crystalline, molten and glassy states. The developed DeePMD-potential was verified using literature and our experimental data demonstrating good agreement with it. Neural network molecular dynamic simulations indicate that P atoms have tetrahedral oxygen environment in both crystalline and glassy state. Coordination environment of Ge atoms is more complex. They have octahedral oxygen environment in crystalline state and mixed environment in glassy state: 4-, 5- and 6-coordinated Ge atoms were found. It was shown by calculation of orientational order parameters that in glassy LiGe2(PO4)3 4-coordinated Ge atoms have tetrahedral oxygen environment, while 6-coordinated Ge atoms have octahedral oxygen environment. Analysis of O-Ge-O angle distribution for 5-coordinated Ge atoms in glassy LiGe2(PO4)3 demonstrates that the environment of Ge atoms in this case is represented by various structures: square pyramids, trigonal bipyramids and transitional structures from 5-coordinated Ge atoms to tetrahedrons.

Transferable Water Potentials Using Equivariant Neural Networks.

Machine learning interatomic potentials (MLIPs) have emerged as a technique that promises quantum theory accuracy for reduced cost. It has been proposed [J. Chem. Phys. 2023, 158, 084111] that MLIPs trained on solely liquid water data cannot accurately transfer to the vapor-liquid equilibrium while recovering the many-body decomposition (MBD) analysis of gas-phase water clusters. This suggests that MLIPs do not directly learn the physically correct interactions of water molecules, limiting transferability. In this work, we show that MLIPs using equivariant architecture and trained on 3200 liquid water structures reproduces liquid-phase water properties (e.g., density within 0.003 g/cm3 between 230 and 365 K), vapor-liquid equilibrium properties up to 550 K, the MBD analysis of gas-phase water cluster up to six-body interactions, and the relative energy and the vibrational density of states of ice phases. We show that potentials developed using equivariant MLIPs allow transferability for arbitrary phases of water that remain stable in nanosecond long simulations.

Deep-potential enabled multiscale simulation of gallium nitride devices on boron arsenide cooling substrates

High-efficient heat dissipation plays critical role for high-power-density electronics. Experimental synthesis of ultrahigh thermal conductivity boron arsenide (BAs, 1300 W m−1K−1) cooling substrates into the wide-bandgap semiconductor of gallium nitride (GaN) devices has been realized. However, the lack of systematic analysis on the heat transfer across the GaN-BAs interface hampers the practical applications. In this study, by constructing the accurate and high-efficient machine learning interatomic potentials, we perform multiscale simulations of the GaN-BAs heterostructures. Ultrahigh interfacial thermal conductance of 260 MW m−2K−1 is achieved, which lies in the well-matched lattice vibrations of BAs and GaN. The strong temperature dependence of interfacial thermal conductance is found between 300 to 450 K. Moreover, the competition between grain size and boundary resistance is revealed with size increasing from 1 nm to 1000 μm. Such deep-potential equipped multiscale simulations not only promote the practical applications of BAs cooling substrates in electronics, but also offer approach for designing advanced thermal management systems.

Mechanical Properties of Single and Polycrystalline Solids from Machine Learning

AbstractCalculating the elastic and mechanical characteristics of non‐crystalline solids can be challenging due to the high computational cost of ab initio methods and the low accuracy of empirical potentials. This paper proposes a computational technique for efficient calculations of mechanical properties of polycrystals, composites, and multi‐phase systems from atomistic simulations with high accuracy and reasonable computational cost. The calculated elastic moduli of polycrystalline diamond and their dependence on grain size are determined using a developed approach based on actively learned machine learning interatomic potentials (MLIPs). These potentials are trained on local fragments of the polycrystalline system, and ab initio calculations are used to compute forces, stresses, and energies. This technique allows researchers to perform extensive calculations of the mechanical properties of complex solids with different compositions and structures, achieving high accuracy and facilitating the transition from ideal (single crystal) systems to more realistic ones.

Enhancing the Quality and Reliability of Machine Learning Interatomic Potentials through Better Reporting Practices

Enhancing the Quality and Reliability of Machine Learning Interatomic Potentials through Better Reporting Practices

Studies of Ni-Cr complexation in FLiBe molten salt using machine learning interatomic potentials

In nuclear and/or solar applications that involve molten salts, impurities frequently enter the salt as either fission products or via corrosion. Impurities can interact and make complexes, but the impact of such complexation on the properties of the salts and corrosion rates has not been understood. Common impurities in molten salts, such as FLiBe, include Cr, Ni, and Fe. Here, we investigate the complexation of Cr and Ni in FLiBe using molecular dynamics based on a machine learning interatomic potential (MLIP) fitted using the atomic cluster expansion (ACE) method. The MLIP allows us to overcome the challenges of simultaneously needing accurate energetics and long time scale to study complexation. We demonstrate that impurity behavior is more difficult to capture than that of concentrated elements with MLIPs due to less sampling in training data, but that this can be overcome by using active learning strategies to obtain a robust fit. Our findings suggest that there is a weak but potentially significant binding free energy between CrF2 and NiF2 in eutectic FLiBe of -0.112eV. Under certain conditions this binding creates a significant short-range order between the species and lowers the redox potential of NiF2 in the presence of CrF2 in FLiBe, making Ni dissolution more favorable in the presence of Cr as compared to its dissolution in pure FLiBe. However, we find little impact of this complexation on the diffusivity of Ni and Cr. Overall, the methodology presented here suggests an approach to modeling complexation with MLIPs and suggests that interactions between dissolved cations could be playing a significant role in some salt thermophysical properties.

Thermal conductivity of crystalline Ge2Sb2Te5: lattice contribution and size effects in the cubic phase quantified by approach-to-equilibrium molecular dynamics

Approach-to-equilibrium molecular dynamics simulations are carried out for the cubic phase of crystalline Ge2Sb2Te5, using interatomic forces derived from a machine learning interatomic potential (MLIP) trained with ab initio calculations. The use of this MLIP potential significantly reduces the computational burden, allowing for the study of systems over 70 nm in length. Above 20 nm, the thermal conductivity of the lattice plateaus at 0.37±0.01 W K−1 m−1 in agreement with measurements reported in the literature. However, below 20 nm, size effects lead to a reduction in thermal conductivity which is systematically calculated.

Recent advances in machine learning interatomic potentials for cross-scale computational simulation of materials

Recent advances in machine learning interatomic potentials for cross-scale computational simulation of materials

Unphysical grain size dependence of lattice thermal conductivity in Mg3(Sb, Bi)2: An atomistic view of concentration dependent segregation effects

Grain boundary (GB) engineering is one of the most common strategies to reduce the lattice thermal conductivity and improve thermoelectric materials. However, in the case of the promising thermoelectric material Mg3(Sb1−xBix)2, an abnormal dependence of lattice thermal conductivity on grain size was observed at the concentration of x = 0.25, where smaller grain polycrystalline materials exhibited higher lattice thermal conductivity compared to larger grain materials. The proposed theory about the overestimation of lattice thermal conductivity in inhomogeneous materials with GBs cannot clarify the concentration dependence of this anomaly. Here we elucidate the interplay between segregation, concentration, and lattice thermal conductivity in Mg3(Sb1−xBix)2 alloys through atomistic simulations with a highly accurate machine learning interatomic potential. We find the largest segregation of Bi atoms to GBs in Mg3(Sb1−xBix)2 at the concentration of x = 0.25 for both twist and tilt GB structures due to the combination effects of site spectrality and solute interactions. Our molecular dynamic simulations demonstrate that the pronounced segregation of heavier Bi atoms, particularly at x = 0.25, leads to a substantial increase in the lattice thermal resistance at GBs, thus contributing to the degree of inhomogeneity. The concentration dependent segregation reveals the atomic origin of the observed unphysical inverse relationship between grain size and lattice thermal conductivity at the specific concentration of x = 0.25. These results highlight the need to design the alloy concentration to tune the atomic segregation and tailor the thermal properties of thermoelectric materials with GBs.

Moment tensor potential for static and dynamic investigations of screw dislocations in bcc Nb

A new machine-learning interatomic potential, specifically a moment tensor potential (MTP), is developed for the study of screw-dislocation properties in body-centered-cubic (bcc) Nb in the thermally- and stress-assisted temperature regime. Importantly, configurations with straight screw dislocations and with kink pairs are included in the training set. The resulting MTP reproduces with near density-functional theory (DFT) accuracy a broad range of physical properties of bcc Nb, in particular, the Peierls barrier and the compact screw-dislocation core structure. Moreover, it accurately reproduces the energy of the easy core and the twinning-anti-twinning asymmetry of the critical resolved shear stress (CRSS). Thereby, the developed MTP enables large-scale molecular dynamics simulations with near DFT accuracy of properties such as for example the Peierls stress, the critical waiting time for the onset of screw dislocation movement, atomic trajectories of screw dislocation migration, as well as the temperature dependence of the CRSS. A critical assessment of previous results obtained with classical embedded atom method potentials thus becomes possible.

Exploring the frontiers of condensed-phase chemistry with a general reactive machine learning potential.

Atomistic simulation has a broad range of applications from drug design to materials discovery. Machine learning interatomic potentials (MLIPs) have become an efficient alternative to computationally expensive ab initio simulations. For this reason, chemistry and materials science would greatly benefit from a general reactive MLIP, that is, an MLIP that is applicable to a broad range of reactive chemistry without the need for refitting. Here we develop a general reactive MLIP (ANI-1xnr) through automated sampling of condensed-phase reactions. ANI-1xnr is then applied to study five distinct systems: carbon solid-phase nucleation, graphene ring formation from acetylene, biofuel additives, combustion of methane and the spontaneous formation of glycine from early earth small molecules. In all studies, ANI-1xnr closely matches experiment (when available) and/or previous studies using traditional model chemistry methods. As such, ANI-1xnr proves to be a highly general reactive MLIP for C, H, N and O elements in the condensed phase, enabling high-throughput in silico reactive chemistry experimentation.

Predicting Spin-Dependent Phonon Band Structures of HKUST-1 Using Density Functional Theory and Machine-Learned Interatomic Potentials.

The present study focuses on the spin-dependent vibrational properties of HKUST-1, a metal-organic framework with potential applications in gas storage and separation. Employing density functional theory (DFT), we explore the consequences of spin couplings in the copper paddle wheels (as the secondary building units of HKUST-1) on the material's vibrational properties. By systematically screening the impact of the spin state on the phonon bands and densities of states in the various frequency regions, we identify asymmetric -COO- stretching vibrations as being most affected by different types of magnetic couplings. Notably, we also show that the DFT-derived insights can be quantitatively reproduced employing suitably parametrized, state-of-the-art machine-learned classical potentials with root-mean-square deviations from the DFT results between 3 cm-1 and 7 cm-1. This demonstrates the potential of machine-learned classical force fields for predicting the spin-dependent properties of complex materials, even when explicitly considering spins only for the generation of the reference data used in the force-field parametrization process.

Incorporating long-range electrostatics in neural network potentials via variational charge equilibration from shortsighted ingredients

We present a new approach to construct machine-learned interatomic potentials including long-range electrostatic interactions based on a charge equilibration scheme. This new approach can accurately describe the potential energy surface of systems with ionic and covalent interactions as well as systems with multiple charge states. Moreover, it can either be regressed against known atomic charge decompositions or trained without charge targets, without compromising the accuracy of energy and forces. We benchmark our approach against other state-of-the-art models and prove it to have equivalent performances on a set of simple reference systems while being less computationally expensive. Finally, we demonstrate the accuracy of our approach on complex systems: solid and liquid state sodium chloride. We attain accuracy in energy and forces better than the model based on local descriptors and show that our electrostatic approach can capture the density functional theory tail of the potential energy surface of the isolated Na-Cl dimer, which the local descriptor-based model fails to describe.

Machine learning accelerated discovery of corrosion-resistant high-entropy alloys

Corrosion has a wide impact on society, causing catastrophic damage to structurally engineered components. An emerging class of corrosion-resistant materials are high-entropy alloys. However, high-entropy alloys live in high-dimensional composition and configuration space, making materials designs via experimental trial-and-error or brute-force ab initio calculations almost impossible. Here we develop a physics-informed machine-learning framework to identify corrosion-resistant high-entropy alloys. Three metrics are used to evaluate the corrosion resistance, including single-phase formability, surface energy and the compactness of oxide films formed on an alloy surface evaluated by Pilling–Bedworth ratios. We used random forest models to predict the single-phase formability, trained on an experimental dataset. Machine learning inter-atomic potentials were employed to calculate surface energies and Pilling–Bedworth ratios, which are trained on first-principles data fast sampled using embedded atom models. A combination of random forest models and high-fidelity machine learning potentials represents the first of its kind to relate chemical compositions to corrosion resistance of high-entropy alloys, paving the way for automatic design of materials with superior corrosion protection. This framework was demonstrated on AlCrFeCoNi high-entropy alloys and we identified composition regions with high corrosion resistance from a wide range of compositions. Machine learning predicted lattice constants and surface energies are consistent with values by first-principles calculations. The predicted single-phase formability and corrosion-resistant compositions of AlCrFeCoNi agree well with experiments. This framework provides a computationally efficient approach to navigate high-dimensional composition space of high-entropy alloys. It is general in its application and applicable to other complex materials, enabling high-throughput screening of material candidates and potentially accelerating the iteration of integrated computational materials engineering.

Robust training of machine learning interatomic potentials with dimensionality reduction and stratified sampling

Machine learning interatomic potentials (MLIPs) enable accurate simulations of materials at scales beyond that accessible by ab initio methods and play an increasingly important role in the study and design of materials. However, MLIPs are only as accurate and robust as the data on which they are trained. Here, we present DImensionality-Reduced Encoded Clusters with sTratified (DIRECT) sampling as an approach to select a robust training set of structures from a large and complex configuration space. By applying DIRECT sampling on the Materials Project relaxation trajectories dataset with over one million structures and 89 elements, we develop an improved materials 3-body graph network (M3GNet) universal potential that extrapolates more reliably to unseen structures. We further show that molecular dynamics (MD) simulations with the M3GNet universal potential can be used instead of expensive ab initio MD to rapidly create a large configuration space for target systems. We combined this scheme with DIRECT sampling to develop a reliable moment tensor potential for titanium hydrides without the need for iterative augmentation of training structures. This work paves the way for robust high-throughput development of MLIPs across any compositional complexity.

Predictions of Boron Phase Stability Using an Efficient Bayesian Machine Learning Interatomic Potential.

Thermodynamic phase stability of three elemental boron allotropes, i.e., α-B, β-B, and γ-B, was investigated using a Bayesian interatomic potential trained via a sparse Gaussian process (SGP). SGP potentials trained with data sets from on-the-fly active learning achieve quantum mechanical level accuracy when employed in molecular dynamics (MD) simulations to predict wide-ranging thermodynamic, structural, and vibrational properties. The simulated phase diagram (500-1400 K and 0-16 GPa) agrees with experimental measurements. The SGP-based MD simulations also successfully predicted that the B13 defect is critical in stabilizing β-B below 700 K. At higher temperatures, the entropy becomes the dominant factor, making β-B the more stable phase over α-B. This letter demonstrates that SGP potentials based on a training set consisting of defect-free-only systems could make correct predictions of defect-related phenomena in solid-state crystals, paving the path to investigate crystal phase stability and transitions.

Modeling Temperature-Dependent Electron Thermal Diffuse Scattering via Machine-Learned Interatomic Potentials and Path-Integral Molecular Dynamics.

Electron thermal diffuse scattering is shown to be sensitive to subtle changes in atomic vibrations and shows promise in assessing lattice dynamics at nanometer resolution. Here, we demonstrate that machine-learned interatomic potentials (MLIPs) and path-integral molecular dynamics can accurately capture the potential energy landscape and lattice dynamics needed to describe electron thermal diffuse scattering. Using SrTiO_{3} as a test bed at cryogenic and room temperatures, we compare electron thermal diffuse scattering simulations using different approximations to incorporate thermal motion. Only when the simulations are based on quantum mechanically accurate MLIPs in combination with path-integral molecular dynamics that include nuclear quantum effects is there excellent agreement with experiments.