arXiv subjects

Kazuaki Toyoura

Publications and source records attributed to Kazuaki Toyoura.

11 recordsLinked to original sources

arXiv2026

Metadynamics for Vacancy Dynamics in Crystals

We propose a metadynamics-based (MetaD-based) approach to construct the free energy surface (FES) for vacancy dynamics in crystals. In this approach, the vacancy FES can be constructed without explicitly defining a unique vacancy coordinate by introducing a set of parameters that strongly govern the FES, which is enabled by parallel bias MetaD with partitioned families (PB-MetaDPF). In addition, the proposed approach is made more efficient and effective by multihill strategy that exploits crystallographic symmetry. We demonstrate the validity and efficacy of the proposed approach by its applications to self-diffusion and impurity diffusion via monovacancies and divacancies in metallic and ionic crystals.

cond-mat.mtrl-sci↗

arXiv2026

A Metadynamics-Based Framework for Free Energy Surface Mapping of Multiparticle Diffusion in Crystals

We propose two metadynamics (MetaD)-based methodologies for efficiently mapping free energy surfaces (FESs) of multiple interacting carriers diffusing in crystalline solids. Our approaches circumvent the challenges of high-dimensional collective variables (CVs) by employing replica state exchange MetaD and parallel bias MetaD, both of which decompose the high-dimensional CVs into multiple lower-dimensional ones. As a benchmark, we investigate two-dimensional lithium (Li) diffusion in LixTiS2 across a wide range of Li concentrations. The Li jump frequencies estimated from the obtained FESs exhibit concentration- and temperature-dependent trends consistent with previous kinetic Monte Carlo simulations and nuclear magnetic resonance measurements. Our approaches provide a promising framework for capturing the slow dynamics of diffusive carriers that are typically inaccessible to conventional molecular dynamics simulations.

cond-mat.mtrl-sci↗

arXiv2024

Nuclear Quantum Effects on Proton Diffusivity in Perovskite Oxides

In the present study, the nuclear quantum effects (NQEs) on proton diffusivity in oxides were evaluated by molecular dynamics (MD) simulations with the quantum thermal bath (QTB) based on the Langevin dynamics. We employed the proton diffusion in barium zirconate (BaZrO3) with the cubic perovskite structure as the model system, in which protons migrate by rotation around single oxide ions and hopping between adjacent oxide ions. MD simulations with the standard classical thermal bath (CTB) and phonon calculations were also conducted to verify the conventionally used classical harmonic transition state theory (classical h-TST), in which the transition state theory (TST), the harmonic approximation, and the classical approximation are assumed. As a result, the h-TST are reasonable for the proton rotation, while significantly overestimate the activation energy and the pre-exponential factor of the jump frequency for the proton hopping. Furthermore, the classical approximation makes the proton jump frequencies close to linear in the Arrhenius plots, which should actually be nonlinear by the NQEs in the temperature range of 500-2000 K. This suggests the necessity of the treatment beyond the classical h-TST for accurate evaluation of the proton diffusivity in oxides even in the intermediate temperature range (573-873 K).

cond-mat.mtrl-sci↗

arXiv2024

Quantitative evaluation of nuclear quantum effects on the phase transitions in BaTiO3 using large-scale molecular dynamics simulations based on machine learning potentials

The machine learning potential (MLP) based molecular dynamics (MD) method was applied for constructing the pressure-temperature phase diagram in the barium titanate (BaTiO3) crystals. The nuclear quantum effects (NQEs) on the phase transitions were quantitatively evaluated from the difference in the phase transition pressures between the NQEs-incorporated and classical simulations. In this study, the quantum thermal bath (QTB) method was used for incorporating the NQEs. The constructed phase diagrams verified that the NQEs lower the phase transition temperatures and pressures. The NQEs are more significant at lower temperatures but cannot be ignored even at room temperature. The phase diagram in the QTB-based MLPMD is in good agreement with those of the previous studies based on dielectric measurements and path-integral based simulations. In addition, this study clarified that the large cell size (a 16x16x16 or larger cell) and friction coefficient (>= 15 THz) are required for accurately reproducing the phase transitions during the QTB-MD simulations.

cond-mat.mtrl-sci↗

arXiv2024

First-principles study of phase transition in cadmium titanate by molecular dynamics incorporating nuclear quantum effects

First-principles molecular dynamics (FPMD) simulations were applied for the paraelectric-ferroelectric phase transition in the perovskite-type cadmium titanate, CdTiO3. Since the phase transition is reported to occur at the low temperature around 80 K, the quantum thermal bath (QTB) method was utilized in this study, which incorporates the nuclear quantum effects (NQEs). The structural evolutions in the QTB-FPMD simulations are in reasonable agreement with the experimental results, by contrast in the conventional FPMD simulations using the classical thermal bath (CTB-FPMD). According to our phonon calculations, volume expansion is the key in the stabilization of the ferroelectric phase at low temperatures, which was well reproduced in the QTB-FPMD with the NQEs. Thus, the NQEs are of importance in phase transitions at low temperatures, particularly below the room temperature, and the QTB is useful in that it incorporates the NQEs in MD simulations with low computational costs comparable to the conventional CTB.

cond-mat.mtrl-sci↗

arXiv2023

Multi-Hill Strategy in Metadynamics for Interstitial Diffusion in Crystals

We propose an efficient and general strategy of metadynamics (MetaD) for investigating interstitial diffusion in a crystal by exploiting crystallographic symmetry. Assuming complete ignorance of the diffusion phenomenon of interest, the three-dimensional coordinates of the interstitial atom with the periodic boundaries are chosen as the collective variables (CVs). Multiple potential hills are simultaneously deposited at all crystallographically-equivalent positions on the free energy surface (FES) defined in the CV space. As a result, the proposed multi-hill strategy highly accelerates atomic jumps in comparison with the single-hill strategy in the conventional MetaD. The key features are that the FES estimated from the final bias potential is exactly satisfied with the symmetry of the host crystal and that all elementary processes of interstitial diffusion are obtained by the single MetaD simulation without any prior knowledge on the diffusion mechanism. The high efficiency and efficacy of the multi-hill strategy are demonstrated, taking the proton diffusion in barium zirconate with the cubic perovskite structure as a model case.

cond-mat.mtrl-sci↗

arXiv2020

Machine-learning-based sampling method for exploring local energy minima of interstitial species in a crystal

An efficient machine-learning-based method combined with a conventional local optimization technique has been proposed for exploring local energy minima of interstitial species in a crystal. In the proposed method, an effective initial point for local optimization is sampled at each iteration from a given feasible set in the search space. The effective initial point is here defined as the grid point that most likely converges to a new local energy minimum by local optimization and/or is located in the vicinity of the boundaries between energy basins. Specifically, every grid point in the feasible set is classified by the predicted label indicating the local energy minimum that the grid point converges to. The classifier is created and updated at every iteration using the already-known information on the local optimizations at the earlier iterations, which is based on the support vector machine (SVM). The SVM classifier uses our original kernel function designed as reflecting the symmetries of both host crystal and interstitial species. The most distant unobserved point on the classification boundaries from the observed points is sampled as the next initial point for local optimization. The proposed method is applied to three model cases, i.e., the six-hump camelback function, a proton in strontium zirconate with the orthorhombic perovskite structure, and a water molecule in lanthanum sulfate with the monoclinic structure, to demonstrate the high performance of the proposed method.

physics.comp-ph↗

arXiv2020

A Sampling Strategy in Efficient Potential Energy Surface Mapping for Predicting Atomic Diffusivity in Crystals by Machine Learning

We propose a machine-learning-based (ML-based) method for efficiently predicting atomic diffusivity in crystals, in which the potential energy surface (PES) of a diffusion carrier is partially evaluated by first-principles calculations. To preferentially evaluate the region of interest governing the atomic diffusivity, a statistical PES model based on a Gaussian process (GP-PES) is constructed and updated iteratively from known information on already-computed potential energies (PEs). In the proposed method, all local energy minima (stable & metastable sites) and elementary processes of atomic diffusion (atomic jumps) are explored on the predictive mean of the GP-PES. The uncertainty of jump frequency in each elementary process is then estimated on the basis of the variance of the GP-PES. The acquisition function determining the next grid point to be computed is designed to reflect the impacts of the uncertainties of jump frequencies on the uncertainty of the macroscopic atomic diffusivity. The numerical solution of the master equation is here employed to readily estimate the atomic diffusivity, which enables us to design the acquisition function reflecting the centrality of each elementary process.

cond-mat.mtrl-sci↗

arXiv2019

Active learning for enumerating local minima based on Gaussian process derivatives

We study active learning (AL) based on Gaussian Processes (GPs) for efficiently enumerating all of the local minimum solutions of a black-box function. This problem is challenging due to the fact that local solutions are characterized by their zero gradient and positive-definite Hessian properties, but those derivatives cannot be directly observed. We propose a new AL method in which the input points are sequentially selected such that the confidence intervals of the GP derivatives are effectively updated for enumerating local minimum solutions. We theoretically analyze the proposed method and demonstrate its usefulness through numerical experiments.

stat.ML↗

arXiv2018

Exploring a potential energy surface by machine learning for characterizing atomic transport

We propose a machine-learning method for evaluating the potential barrier governing atomic transport based on the preferential selection of dominant points for the atomic transport. The proposed method generates numerous random samples of the entire potential energy surface (PES) from a probabilistic Gaussian process model of the PES, which enables defining the likelihood of the dominant points. The robustness and efficiency of the method are demonstrated on a dozen model cases for proton diffusion in oxides, in comparison with a conventional nudge elastic band method.

physics.chem-ph↗

arXiv2015

A machine learning-based selective sampling procedure for identifying the low energy region in a potential energy surface: a case study on proton conduction in oxides

In this paper, we propose a selective sampling procedure to preferentially evaluate a potential energy surface (PES) in a part of the configuration space governing a physical property of interest. The proposed sampling procedure is based on a machine learning method called the Gaussian process (GP), which is used to construct a statistical model of the PES for identifying the region of interest in the configuration space. We demonstrate the efficacy of the proposed procedure for atomic diffusion and ionic conduction, specifically the proton conduction in a well-studied proton-conducting oxide, barium zirconate BaZrO3. The results of the demonstration study indicate that our procedure can efficiently identify the low-energy region characterizing the proton conduction in the host crystal lattice, and that the descriptors used for the statistical PES model have a great influence on the performance.

cond-mat.mtrl-sci↗