462 papers
This paper benchmarks three machine learning frameworks (MODNet, CrabNet, and a Magpie-based Random Forest) for predicting battery electrode properties using the Materials Project dataset, demonstrating that CrabNet consistently outperforms the others across rigorous statistical validation while highlighting both the potential and practical limitations of ML-driven materials discovery.
This paper presents an AI-driven workflow combining a Gaussian mixture variational autoencoder with Pearson correlation coefficients to analyze sparsely sampled X-ray hyperspectral data, enabling the generation of nanometer-resolution multiphase maps that reveal complex phase heterogeneity and transition zones in individual Na-ion cathode particles during electrochemical cycling.
This paper demonstrates that robust defect detection in fluctuating magnetic systems, such as Ni80Fe20, can be achieved by training U-Net models on statistical descriptors like temporal mean, standard deviation, and latent entropy derived from micromagnetic simulations, provided the training data accurately reflects the expected noise statistics.
This perspective evaluates the impact of dataset size, reference quality, and model architectures on training machine learning force fields for solid-state electrolytes, concluding that rigid frameworks prioritize data quality over quantity and that force accuracy alone is insufficient for predicting transport performance.
This study utilizes 3D pFIB-SEM microscopy and '2.5D' reconstruction to demonstrate that high-temperature austenite twin boundaries significantly govern variant selection and grain growth during the solid-state phase transformation in low carbon steels, offering new avenues for engineering microstructures with enhanced mechanical properties.
This study reveals that in 310S TWIP steel, the transition from dislocation slip to deformation twinning and the associated evolution of dual-fibre texture progressively reduce the material's capacity to store deformation energy, thereby facilitating shear-band-mediated deformation.
This paper introduces the Practical Time Averaging (PTA) framework to overcome the timescale limitations of conventional molecular dynamics, enabling efficient atomistic simulations of quasistatic deformation in aluminum nanocrystals that reveal size-dependent strengthening, dislocation dynamics, and yield behavior at experimentally relevant strain rates.
This paper introduces a spin neural network potential (SpinNNP) that integrates spin degrees of freedom and spin-orbit coupling to successfully simulate the antiferromagnetic-paramagnetic phase transition in uranium dioxide, offering a computationally efficient route for predictive modeling of complex magnetic actinide oxides.
This study demonstrates that biaxial strain can dynamically tune the topological order and piezoelectric functionality of Tc-adsorbed penta-hexa silicene monolayers by engineering momentum-space orbital hybridization, thereby transforming static materials into versatile, strain-responsive quantum platforms.
This paper presents a device physics model demonstrating that Fermi-level pinning by donor-like defects at the p-ZnTe/p-CdSeTe hole-selective contact causes downward band bending and fill factor losses in CdSeTe/CdTe solar cells, while suggesting that passivating these interfaces could significantly enhance efficiency, particularly in thinner devices with longer carrier diffusion lengths.
This study investigates the magnetic properties and crystal electric field scheme of the spin-orbit coupled distorted honeycomb magnet BiErGeO, revealing short-range antiferromagnetic correlations, long-range order at 0.4 K, and persistent slow spin fluctuations below the transition temperature through a combination of magnetization, heat capacity, muon spin relaxation, and inelastic neutron scattering experiments.
This paper extends a statistical framework to predict the thermodynamics of self-interstitial dumbbells in concentrated multicomponent alloys, revealing that solute elements like Cr can stabilize high-energy defects and induce symmetry-breaking distortions in the defect free energy surface.
This study introduces a computationally efficient framework that utilizes non-interacting electron density and Bayesian active learning to achieve highly accurate, zero-shot extrapolative predictions of alloy properties across vast compositional landscapes, significantly reducing the data requirements for discovering refractory high-entropy alloys.
This study investigates the correlations between dielectric properties, domain structure morphology, and phase states in Bi1-xSmxFeO3 nanoparticles by combining experimental measurements of temperature-dependent dielectric behavior with theoretical modeling based on the Ginzburg-Landau-Devonshire-Stephenson-Highland approach to elucidate ferro-ionic coupling effects.
This study utilizes molecular beam epitaxy to systematically map the adsorption-controlled growth window of -GaSe on GaAs (111)B substrates, revealing that while higher temperatures improve crystalline quality and surface smoothness, they also induce a transition from singly oriented to twinned domains.
This paper introduces AIMD-L, an automated, high-throughput laboratory featuring custom instruments (HELIX and MAXIMA) and a robotic workflow designed to rapidly characterize the microstructure and properties of structural metals and ceramics for extreme environments, thereby closing the loop between AI-driven experimentation and data analysis.
This study demonstrates that the performance gap between laboratory and industrial roll-to-roll gravure-printed non-fullerene organic solar cells stems from device architecture and optical losses rather than intrinsic material physics, establishing a roadmap for high-efficiency manufacturing using commercially available materials and non-halogenated solvents.
This paper investigates how hierarchically patterned architected thin films in bi-layer composites can localize interfacial failure and enhance fracture toughness by creating a buffer region for diffuse damage dissipation, utilizing a novel network formalism that integrates discrete differential geometry and spectral graph theory to analyze load redistribution and deformation modes.
This study establishes that maintaining structural and electrostatic consistency between interface and bulk reference calculations, specifically through the use of mixed hybrid-semilocal exchange-correlation functionals combined with strained bulk references, is the dominant factor for achieving near-experimental accuracy in predicting Schottky barrier heights at Si/Metal interfaces.
High-resolution neutron scattering and DFT studies on LaSrMnO reveal that despite weak magnetoresistance, Jahn-Teller-active oxygen vibrations collapse above the Curie temperature due to giant electron-phonon coupling driving diffusive oxygen distortions, suggesting that the magnitude of magnetoresistance correlates with the rate of this diffusion rather than the strength of the coupling itself.