GeTe, a prototypical phase‐change material, has attracted pronounced attention for next‐generation storage and neuromorphic computing, yet its nucleation mechanism remains an ongoing pursuit. The challenge stems from the limited scales of conventional simulations and the intrinsic structural disorder of nucleation. To bridge the scale gap, we developed a machine learning potential (ML potential) trained on an extensive density functional theory dataset, enabling large‐scale molecular dynamics simulations that capture the complete crystallization pathway with quantum‐level fidelity. By employing a unified short‐ and medium‐range structural framework, we simplified the complexity and disorder inherent to nucleation, allowing us to unravel the intricate atomic environment, pinpoint essential structural motifs, and track their dynamic evolution. Through this approach, our simulations reveal an unconventional nucleation process: the initial formation of Ge‐rich clusters with defective octahedral coordination, followed by their Te‐mediated assembly into the final rock‐salt structure. This sequential mechanism presents a different scenario from classical nucleation theory's expectation of alternating Ge/Te incorporation, wherein the Te sublattice preferentially forms a face‐centered‐cubic structure ahead of Ge ordering. The observed two‐stage nucleation process adds a valuable perspective on the structural ordering kinetics in PCMs, helping to explain their fast crystallization characteristics at the atomic level.
The assembly of nanodiamonds (NDs) dictates their emergent structural and functional states, yet the atomistic mechanisms governing this process remain largely unresolved. In this work, the facet-dependent interactions and temperature-regulated aggregation of NDs are investigated through large-scale deep potential molecular dynamics simulations using a newly developed machine-learning potential. Utilizing dimerized truncated-octahedral C1126 clusters as a model system, we identify a fundamental dichotomy in the assembly motifs: the {100} facets promote spontaneous covalent fusion even at ambient conditions, whereas graphitized {111} surfaces interact predominantly through noncovalent π–π stacking. Under thermal activation, a progressive transition is observed from nonbonded or weakly interacting contacts to increasingly reconstructed interfaces, followed by extensive and ultimately complete coalescence, establishing temperature as a critical parameter for tuning interfacial morphology. At 1500 K, where clear fusion occurs, the two interfaces follow distinct pathways: a center-initiated bidirectional zippering for {100}-{100} and an edge-initiated unidirectional zippering for {111}-{111}. In both cases, bonding is dominated by outer-shell atoms, and the resulting peanut-like core–shell structures contain undercoordinated and curved interfacial motifs that may serve as candidate anchoring environments for metal species.
Rui He, Jingshuang Dang· Journal of Physical Chemistr...· 0 citations
Molecular dynamics (MD) is essential for investigating atomic‐scale processes in materials and molecular systems, but the cost of high‐accuracy machine learning force field simulations still limits accessible system sizes and timescales. Here, we propose a practical model‐switching strategy for Deep Potential (DP)‐based MD simulations that alternates between independently trained DP models with different cutoff radii: a standard 6 Å model for higher accuracy and a reduced‐cutoff 4 Å model for faster inference. The method was implemented in LAMMPS/DeePMD and evaluated using solid‐phase anatase TiO
2
and liquid‐phase polyethylene glycol (PEG). For anatase TiO
2
, the 1:3 4–6 Å switching scheme preserved radial distribution function (RDF) correlations of 0.996 or higher relative to the 6 Å baseline while achieving a 1.24‐fold speedup. For PEG, the switching scheme maintained RDF correlations of 0.996 or higher with a 1.18‐fold speedup. Additional optimization using network‐size reduction and mixed‐precision inference achieved a 2.53‐fold speedup with RDF correlations of 0.975–0.988. Constant particle‐number, pressure, and temperature (NPT) simulations remained stable, whereas constant particle‐number, volume, and energy (NVE) simulations revealed system‐dependent energy‐drift behavior, particularly for aggressively optimized models. These results demonstrate that DP model switching provides a simple and practical route for accelerating structural MD simulations while highlighting the need for validation when strict energy conservation is required.
R. Kanda, Megumu Yamazaki, Yuta Yoshimoto et al.· Advanced Intelligent Discove...· 0 citations
Complex microstructural pattern formation, such as dendrite growth, occurs across a wide range of materials and plays a crucial role in determining their properties and functional performance. While the phase-field method is a powerful computational approach for modeling microstructure dynamics, its substantial computational cost limits its integration into practical materials design workflows. Here, we introduce a machine-learning framework that employs autoregressive deep surrogates, trained on short trajectories from quantitative phase-field simulations of alloy solidification within limited spatial domains. Once trained, these surrogates accurately predict dendritic evolution over extended length and time scales, achieving speed-ups exceeding two orders of magnitude. We demonstrate the effectiveness of this approach through examples of isothermal growth and directional solidification of a dilute Al–Cu alloy, confirming its capability to predict complex microstructural pattern formation. Quantitative comparisons with phase-field benchmarks reveal excellent agreement in the tip-selection constant, morphological symmetry, and primary spacing evolution, further validating our approach.
Kai-Hua Ji, Luning Sun, Shusen Liu et al.· Machine Learning: Science an...· 0 citations
This study presents a machine‐learning‐based method to predict the spatiotemporal evolution of dust particle configurations in two‐dimensional dusty plasmas. Evolutionary datasets are generated via molecular dynamics simulations under the Yukawa potential approximation, and particle configurations are encoded as fixed‐resolution grayscale image sequences. A spatiotemporal recurrent neural network, PredRNN, is adopted to model the structural evolution of dust particles and conduct rolling predictions from a limited number of initial frames. Results demonstrate that PredRNN effectively captures the dynamical behavior of dust particles and yields predictions consistent with molecular dynamics simulations for key structural features, including cluster formation, coalescence, and long‐term evolution. Quantitative evaluations using MSE, SSIM, and LPIPS further validate the model's predictive performance at both structural and perceptual levels. The proposed framework accurately reproduces the self‐organized structural evolution of dust systems and offers a data‐driven perspective for understanding collective self‐organization dynamics in dusty plasmas. It also provides an efficient computational approach for long‐timescale extrapolation.
Unknown authors· Contributions to Plasma Phys...· 0 citations
Cubic boron nitride (c-BN) nanoparticles are promising for extreme-condition applications, yet their atomistic evolution remains poorly understood. Here, we develop a high-fidelity machine learning potential and perform large-scale deep potential molecular dynamics simulations to investigate their high-temperature behavior. A universal reconstruction pathway is revealed, involving defect formation, inward-to-outward atomic migration, and progressive healing into multilayer hexagonal BN (h-BN). This mechanism is validated across multiple morphologies and exposed facets and is found to be strongly dependent on facet and termination. Furthermore, temperature-programmed dynamics identify ∼1800 K as the critical threshold for activating large-scale atomic flux, driving the transformation from core-shell architectures to multishell fullerene-like h-BN structures. At extreme temperatures (>3300 K), chemical segregation emerges, leading to the formation of boron clusters and polynitrogen chains, consistent with experimental observations. We further compared the reconstruction behaviors of isoelectronic nanodiamond and c-BN nanoparticles, revealing that c-BN exhibits superior thermal stability and enhanced self-healing capability, originating from the higher kinetic barriers associated with partially ionic B-N bonds relative to covalent C-C bonds.
This perspective examines three interconnected issues, namely, glass formation procedures, interatomic potential development, and machine learning applications, which emerged from the 5th International Workshop on Challenges of Atomistic Simulations of Glasses and Amorphous Materials.
N. A. Anoop Krishnan, A. Pedone, Xiaonan Lu et al.· Journal of The American Cera...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
