Rare earth (RE) Mg alloy is a thriving branch of high modulus Mg alloys, since there exist various precipitations to strengthen them. In this work, the structural and mechanical properties of ten different kinds of binary Mg-RE intermetallic compounds (IMCs), including B2, C14, C15, D0(19), D0(3), Mg7RE, Mg12RE, Mg17RE2, Mg24RE5, and Mg41RE5, were systematically investigated by first-principles calculations. The accuracy of the current work was validated by comparing it with limited previous work. The stability of those IMCs was evaluated by accurate convex hull (stable in energy) and elastic constants (mechanical stability). The atomic volume is linear with the ionic radii of corresponding RE elements in each IMC. Regarding the strengthening effect, both D0(3) and Mg12RE show considerable Young's modulus when compared with pure Mg (about 44.6 GPa), and the maximum Young's modulus reaches 74 GPa. For all ten IMCs, the strengthening effect follows the sequence: Mg12RE approximate to D0(3) > Mg17RE2 approximate to Mg41RE5 > B2 > C15 > Mg7RE > Mg24RE5 > C14 > D0(19) (for light rare earth elements), and D0(3) approximate to Mg12RE > B2 > Mg17RE2 approximate to Mg41RE5 > C14 > C15 > Mg24RE5 > D0(19) > Mg7RE (for heavy rare earth elements). In addition, the rule of mixture was validated for atomic volume, formation energy, and Young's modulus in multicomponent Mg-m(REs)(n) systems. Hence, this work provides abundant and fundamental data for designing and understanding novel precipitation-strengthening Mg alloys.

Graph neural networks (GNN) have emerged as a promising machine learning method for microstructure simulations such as grain growth. However, accurate modeling of realistic grain boundary networks requires large simulation cells, which GNN has difficulty scaling up to. To alleviate the computational costs and memory footprint of GNN, we propose a hybrid architecture combining a convolutional neural network (CNN) based bijective autoencoder to compress the spatial dimensions, and a GNN that evolves the microstructure in the latent space of reduced spatial sizes. Our results demonstrate that the new design significantly reduces computational costs with using fewer message passing layer (from 12 down to 3) compared with GNN alone. The reduction in computational cost becomes more pronounced as the spatial size increases, indicating strong computational scalability. For the largest mesh evaluated (1603), our method reduces memory usage and runtime in inference by 117 & times; and 115 & times;, respectively, compared with GNN-only baseline. More importantly, it shows higher accuracy and stronger spatiotemporal capability than the GNN-only baseline, especially in long-term testing. Such combination of scalability and accuracy is essential for simulating realistic material microstructures over extended time scales. The improvements can be attributed to the bijective autoencoder's ability to compress information losslessly from spatial domain into a high dimensional feature space, thereby producing more expressive latent features for the GNN to learn from, while also contributing its own spatiotemporal modeling capability. Training data are generated from stochastic grain growth simulations, providing realistic variability for learning robust microstructure evolution. Comprehensive system validation confirms that the model is accurate, robust, and scalable.

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.

In order to appropriately capture large-scale material features and emergent phenomena via atomistic simulations, such as Molecular Dynamics (MD), the system scale can range up to hundreds of millions of atoms. However, the force-field models that drive those simulations are generally trained with Density Functional Theory (DFT) reference data, limited to relatively small configurations on the order of 100s or 1000s of atoms. To compute DFT forces on atoms in regions of interest, for example for active-learning or on-the-fly training of interatomic potentials, one needs to extract a small set of atoms from the larger simulation box, and typical work with periodic boundary conditions for DFT; however, methods to select the shape and size of this extracted set of atoms, as well as to generate a potentially necessary passivating envelope, have not been systematically analyzed. In this work, we present a benchmark of various techniques to extract atomic environments from large, bulk configurations and embed them into smaller configurations suitable for DFT calculations with periodic boundary conditions. We test with a diverse set of material systems, which includes amorphous SiO_2, Ta with screw dislocations, and molten C. We demonstrated a notably simple procedure, a method we refer to as deletions, yields superior performance over an array of alternative extraction methods.

Refractory complex concentrated alloys (RCCA) offer exceptionally high-temperature strength compared to pure metals and dilute alloys, but predictive theory for RCCA design is lacking. We present large-scale molecular Dynamics (MD) simulations of crystal plasticity to explore alloy compositions for maximum mechanical strength, focusing on Fe-Ta-W and Nb-Ta-Mo-W alloy families modeled with Embedded Atom Model (EAM) and Spectral Neighbor Analysis Potentials (SNAP). To efficiently guide the search for strong alloy compositions, we employ iterative optimization using Gaussian process regression. Many simulated RCCA compositions exhibit pronounced cocktail strengthening, with strengths surpassing their strongest constituent metal, tungsten. Contrary to expectations, the highest strength is found on binary edges of the RCCA composition space. Detailed analyses of atomistic simulations reveal that, similar to pure BCC metals, plastic response in RCCA is primarily governed by screw dislocations. However, at large strains, dislocation multiplication and interactions (Taylor hardening) become the dominant mechanisms contributing to RCCA strength.

A simulation can stand its ground against an experiment only if its prediction uncertainty is known. The unknown accuracy of interatomic potentials (IPs) is a major source of prediction uncertainty, severely limiting the use of large-scale classical atomistic simulations in a wide range of scientific and engineering applications. Here we explore covariance between predictions of metal plasticity, from 178 large-scale (~108 atoms) molecular dynamics (MD) simulations, and a variety of indicator properties computed at small-scales (≤102 atoms). All simulations use the same 178 IPs. In a manner similar to statistical studies in public health, we analyze correlations of strength with indicators, identify the best predictor properties, and build a cross-scale “strength-on-predictors” regression model. This model is then used to estimate regression error over the statistical pool of IPs. Small-scale predictors found to be highly covariant with strength are computed using expensive quantum-accurate calculations and used to predict flow strength, within the statistical error bounds established in our study.

The weak interaction between transition metal selenides and carbon materials caused difficulties for their combination. In this study, a novel strategy based on lignin and Co nanoparticles was used for modifying electrospun carbon nanofibers to boost the growth of Ni3Se4. Interestingly, various types of lignin could induce the aggregation and distribution of Co nanoparticles, leading to different crystal structures, morphologies, and amounts of Ni-based selenides. Among them, sodium lignosulphonate brought about Co nanoparticles inside nanofibers and the formation of S-Co bonds, leading to the massive growth of Ni3Se4 nanosheets with high redox activities and abundant active sites. As a result, the sample modified by Co and sodium lignosulphonate (CFLS@NiSe) with a diffusion-controlled hybrid capacity displayed the highest value of 384.00 C g-1 (853.33 F g-1) at 0.5 A g-1 and acceptable rate performance (capacity retention of 39 %, from 0.5 to 10 A g-1), which exceeded many reported transition metal selenide-carbon composites. Furthermore, a hybrid solid-state super-capacitor (CFLS@NiSe//bamboo derived active carbon) with a high energy density of 61.33 Wh kg-1 was assembled, which could supply energy for light emitting diode and deliver a capacity retention of 94.15 % as well as Coulombic efficiency of 92.29 % after 10,000 cycles. Compared with traditional modification strategies, this work proposed a novel, facile, feasible, low cost and environmentally friendly method to modify carbon nano-fibers to regulate the growth of selenides, which will offer ideas for the design of composite materials.

Lithium-ion batteries (LIBs) face significant safety challenges due to the inherent limitations of conventional polyolefin separators, such as poor mechanical strength and inadequate thermal stability, which heighten the risk of thermal runaway episodes. This study presents a robust, high-temperature-resistant aramid nanofiber (ANF) separator with a hierarchical lamellar interconnected network structure, synthesized via a bottom-up low-temperature polycondensation strategy. The resulting ANF separator exhibits exceptional mechanical strength (192 MPa), outstanding thermal stability (initial degradation temperature of ≈510 °C), and negligible thermal shrinkage even at 300 °C. Electrochemical evaluations reveal superior Li⁺ transference number (0.536) and high temperature cycling stability (94.6% capacity retention after 100 cycles at 100 °C), outperforming commercial polypropylene (PP) separators. Compared with PP, the pouch lithium cell assembled by ANF separator can effectively maintain structural integrity even under harsh thermal conditions (150 °C). This work demonstrates a scalable, efficient method to fabricate advanced separators with a multiscale structure composed of ANF, addressing critical safety concerns and enhancing the performance of high-energy-density LIBs.

An accurate description of information is relevant for a range of problems in atomistic machine learning (ML), such as crafting training sets, performing uncertainty quantification (UQ), or extracting physical insights from large datasets. However, atomistic ML often relies on unsupervised learning or model predictions to analyze information contents from simulation or training data. Here, we introduce a theoretical framework that provides a rigorous, model-free tool to quantify information contents in atomistic simulations. We demonstrate that the information entropy of a distribution of atom-centered environments explains known heuristics in ML potential developments, from training set sizes to dataset optimality. Using this tool, we propose a model-free UQ method that reliably predicts epistemic uncertainty and detects out-of-distribution samples, including rare events in systems such as nucleation. This method provides a general tool for data-driven atomistic modeling and combines efforts in ML, simulations, and physical explainability.

All-solid-state lithium batteries (ASSLBs) employing Li-rich layered oxide (LLO) cathodes are regarded as promising next-generation energy storage systems owing to their outstanding energy density and intrinsic safety. Polymer-in-salt solid electrolytes (PISSE) offer advantages such as low processing costs, high ionic conductivity, and good anode compatibility; however, their practical deployment is hindered by poor oxidative stability especially under high-voltage. In this study, we report the rational design of a bilayer electrolyte architecture featuring an in situ solidified LiClO₄-doped succinonitrile (LiClO₄–SN) plastic-crystal interlayer between a Li₁.₂Mn₀.₆Ni₀.₂O₂ (LMNO) cathode and a PVDF-HFP-based PISSE. This PISSE/SN–LiClO₄ configuration exhibits a wide electrochemical stability window up to 4.7 V vs. Li+/Li and delivers a high ionic conductivity of 2.38 × 10-4 S cm-1 at 25 °C. The solidified LiClO₄-SN layer serves as an effective physical barrier, shielding the PVDF-HFP matrix from direct interfacial contact with LMNO and thereby suppressing its oxidative decomposition at elevated potentials. As a result, the bilayer polymer-based cells with LMNO cathode demonstrate an initial discharge capacity of ∼206 mAh g-1 at 0.05 C and exhibit good cycling stability with 85.7% capacity retention after 100 cycles at 0.5 C under a high cut-off voltage of 4.6 V. This work not only provides a promising strategy to enhance the compatibility of PVDF-HFP-based electrolytes with high-voltage cathodes through the facile in-situ solidification of plastic interlayers but also promotes the application of LMNO cathode material in high-energy ASSLBs.

Atomistic simulations generate large volumes of noisy structural data, yet extracting phase labels and continuous order parameters (OPs) in a robust and general manner remains challenging. Existing tools are often specialized to a limited set of prototypes and split thermal-noise removal, phase classification, and OP construction into separate steps. Here we present a unified probabilistic framework for analyzing noisy atomic configurations with respect to known crystal prototypes. The model predicts per-atom, per-prototype logits and aggregates them into a scalar log-probability (logP) landscape over atomic coordinates. Its gradient defines a conservative denoising field, while the logits provide local phase labels, prototype-resolved OPs, and ambiguity measures through logit margins. We train on AFLOW-mapped crystalline structures from the Materials Project with synthetic positional and elastic perturbations, then test extrapolation to stronger noise, finite-temperature disorder, point defects, water–ice coexistence, binary polymorphs, and shock-compressed Ti. A single differentiable scalar model recovers prototype identity after denoising, tracks smooth transformations such as Bain and Burgers paths, and exposes low-confidence regions near defects and phase boundaries. This provides an integrated and extensible tool for analyzing complex atomistic simulations.

Model ensembles are effective tools for estimating prediction uncertainty in deep learning atomistic force fields. However, their widespread adoption is hindered by high computational costs and overconfident error estimates. In this work, we address these challenges by leveraging distributions of per-sample errors obtained during training and employing a distance-based similarity search in the model latent space. Our method, which we call LTAU (Loss Trajectory Analysis for Uncertainty), efficiently estimates the full probability distribution function of errors for any test point using the logged training errors, achieving speeds that are 2-3 orders of magnitudes faster than typical ensemble methods and allowing it to be used for tasks where training or evaluating multiple models would be infeasible. We apply LTAU towards estimating parametric uncertainty in atomistic force fields (LTAU-FF), demonstrating that it produces well-calibrated confidence intervals and predicts errors that correlate strongly with the true errors for data near the training domain. Furthermore, we show that the errors predicted by LTAU-FF can be used in practical applications for detecting out-of-domain data, tuning model performance, and predicting failure during simulations. We believe that LTAU will be a valuable tool for uncertainty quantification in atomistic force fields and is a promising method that should be further explored in other domains of machine learning.

Magnesium (Mg) alloys, especially high-modulus Mg alloys, are catching more and more attention for their excellent performance. Doping rare earth elements (RE) and then generating hard intermetallic is one of the popular ways to strengthen the Mg alloys. As an important intermetallic, Mg24RE5 is rarely investigated. In this work, we comprehensively investigate the structural, mechanical, and thermodynamic properties of Mg24RE5 by first-principles calculations. The accuracy of our results is confirmed by comparing with limited previous results. All Mg24RE5 are energy-stable, mechanically stable, and dynamically stable. The lattice constant is almost linear with the ionic radii of the corresponding RE elements. Mg24Y5 shows the largest Young's modulus (54.6 GPa), followed by Mg24Lu5 and Mg24Tm5, while Mg24Sm5 and Mg24Gd5 show the lowest anisotropy. Moreover, we calculated the thermodynamic properties for all Mg24RE5 and obtained the critical temperature for heat capacity. All in all, this work provides plentiful fundamental data for designing RE-containing Mg alloys.

Due to their exceptional high-temperature strength compared to pure metals and dilute metal alloys, refractory complex concentrated alloys (RCCA) have attracted significant attention in recent times. However, predictive theory capable of guiding design and compositional optimization of RCCA for mechanical strength is yet to emerge. Here we present a series of large-scale Molecular Dynamics (MD) simulations of crystal plasticity devised to explore the space of alloy compositions in order to maximize their mechanical strength. As test-bed materials we focus on the Fe-Ta-W and the Nb-Ta-Mo-W alloy families, modeled here using Embedded Atom Model (EAM) and Spectral Neighbor Analysis Potentials (SNAP) interatomic potentials. To efficiently navigate our search towards mechanically strong alloy compositions, we employ iterative optimization using Gaussian process regression. Many of the simulated RCCA compositions exhibit well pronounced cocktail strengthening displaying strengths markedly exceeding that of their strongest constituent metal (tungsten in this case). At variance with common expectations, our computational experiments suggest that the highest strength may be achievable on the binary edges of the RCCA composition space. Taking advantage of the fully resolved atomistic trajectories, our simulations indicate that, much like in pure BCC metals, screw dislocations play a primary role in the plastic response of RCCA. However at large strains well past the yield point, dislocation multiplication and dislocation interactions (Taylor hardening) take over as the dominant contribution to RCCA strength.

We propose score dynamics (SD), a general framework for learning accelerated evolution operators with large timesteps from molecular dynamics (MD) simulations. SD is centered around scores or derivatives of the transition log-probability with respect to the dynamical degrees of freedom. The latter play the same role as force fields in MD but are used in denoising diffusion probability models to generate discrete transitions of the dynamical variables in an SD time step, which can be orders of magnitude larger than a typical MD time step. In this work, we construct graph neural network-based SD models of realistic molecular systems that are evolved with 10 ps timesteps. We demonstrate the efficacy of SD with case studies of the alanine dipeptide and short alkanes in aqueous solution. Both equilibrium predictions derived from the stationary distributions of the conditional probability and kinetic predictions for the transition rates and transition paths are in good agreement with MD. Our current SD implementation is about 2 orders of magnitude faster than the MD counterpart for the systems studied in this work. Open challenges and possible future remedies to improve SD are also discussed.

The ability to rapidly develop materials with desired properties has a transformative impact on a broad range of emerging technologies. In this work, we introduce a new framework based on the diffusion model, a recent generative machine learning method to predict 3D structures of disordered materials from a target property. For demonstration, we apply the model to identify the atomic structures of amorphous carbons ($a$-C) as a representative material system from the target X-ray absorption near edge structure (XANES) spectra--a common experimental technique to probe atomic structures of materials. We show that conditional generation guided by XANES spectra reproduces key features of the target structures. Furthermore, we show that our model can steer the generative process to tailor atomic arrangements for a specific XANES spectrum. Finally, our generative model exhibits a remarkable scale-agnostic property, thereby enabling generation of realistic, large-scale structures through learning from a small-scale dataset (i.e., with small unit cells). Our work represents a significant stride in bridging the gap between materials characterization and atomic structure determination; in addition, it can be leveraged for materials discovery in exploring various material properties as targeted.

The diffusion model has emerged as a powerful tool for generating atomic structures for materials science. This work calls attention to the deficiency of current particle-based diffusion models, which represent atoms as a point cloud, in generating even the simplest ordered crystalline structures. The problem is attributed to particles being trapped in local minima during the score-driven simulated annealing of the diffusion process, similar to the physical process of force-driven simulated annealing. We develop a solution, the grand canonical diffusion model, which adopts an alternative voxel-based representation with continuous rather than fixed number of particles. The method is applied towards generation of several common crystalline phases as well as the technologically important and challenging problem of grain boundary structures.

Due to the low economic benefits and environmental pollution of traditional recycling methods, the disposal of spent LiFePO4 (SLFP) presents a significant challenge. The capacity fade of SLFP cathode is primarily caused by lithium loss and formation of a Fe (III) phase. Herein, a synergistic repair effect is proposed to achieve defect repair and multi-functional interface construction for the direct regeneration of SLFP. Tannic acid (TA) forms a compact coating precursor for a carbon layer on SLFP with abundant functional groups and creates a mildly acidic environment to enhance the reducibility of thiourea (TU). Therefore, TU reduces Fe (III) to Fe (II) and repairs Li-Fe anti-site defects of SLFP, while at the same time acting as a source of N/S-doping elements for the carbon layer at a lower temperature (140 °C). The multi-functional carbon layer improves the properties of the regenerated LiFePO4 (RLFP) due to the enhanced conductivity, structure maintenance and protection, and the improved kinetics of Li+ transport. Furthermore, the Fe─O and P─O bonds are strengthened, further enhancing the structural stability of the RLFP. Consequently, the RLFP demonstrates outstanding performance with a discharge capacity of 141.3 mAh g-1 and capacity retention of 72% after 1000 cycles at 1 C.