Machine learning via artificial neural networks coupled with density functional theory and experiments for thermodynamic optimization of high-entropy alloys for hydrogen storage at room temperature

Shivam Dangwal; Pranav Kumar; Yuji Ikeda; Blazej Grabowski; Kaveh Edalati

doi:10.1016/j.ijhydene.2026.155840

Home/Archive/arXiv

arXiv · 2606.04803

Machine learning via artificial neural networks coupled with density functional theory and experiments for thermodynamic optimization of high-entropy alloys for hydrogen storage at room temperature

Shivam Dangwal·Pranav Kumar·Yuji Ikeda·Blazej Grabowski·Kaveh Edalati

View original source ↗Read source document ↗Export metadata ↓

Abstract

High-entropy alloys (HEAs) have received considerable attention for hydrogen storage because of their compositional flexibility; however, designing HEAs with optimal thermodynamics is critical. This study employs machine learning via artificial neural networks (ANN) and density functional theory (DFT) to design a novel AB-type TixNb2-xVCrMnFe (x = 0.5-2.0) high-entropy system for hydrogen storage at ambient temperature (A: Ti, V and Nb, and B: Cr, Mn and Fe). Both ANN and DFT predict that the hydride formation enthalpy decreases to negative values with increasing the titanium content. Two alloys with x > 1.5 are predicted to achieve enthalpies within the -25 to -39 kJ/mol range, making them appropriate for room-temperature hydrogen storage. Experiments demonstrate good agreement with the enthalpy predictions, with the Ti-rich alloys showing reversible hydrogen storage with fast kinetics at room temperature. These results provide a framework for reliable use of data analysis and ab initio calculations to explore high-entropy hydrides as hydrogen storage materials.

Explore related subjects

cond-mat.mtrl-sci

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Shivam Dangwal, Pranav Kumar, Yuji Ikeda, Blazej Grabowski, Kaveh Edalati. 2026-06-03. Machine learning via artificial neural networks coupled with density functional theory and experiments for thermodynamic optimization of high-entropy alloys for hydrogen storage at room temperature. https://doi.org/10.1016/j.ijhydene.2026.155840

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

A hierarchy of thermodynamics learning frameworks for inelastic constitutive modeling

Recent advances in physics-augmented neural networks have enabled thermodynamically consistent data-driven constitutive modeling of complex inelastic materials. Most existing approaches, however, implicitly adopt a specific thermodynamic framework and embed structural assumptions such as normality, dual dissipation potentials, or other structure from manually constructed models directly into the learning architecture. Consequently, differences in predictive performance may arise not only from data or network design, but also from the underlying theoretical assumptions. In this work, we present a unified comparison of several thermodynamically consistent inelastic modeling frameworks from a machine learning perspective. We consider internal-variable formulations with dissipation potential, generalized standard materials, and metriplectic structures, and we analyze their structural assumptions, admissible dependencies, convexity requirements, and implications for dissipation and evolution. Each framework is implemented within a common neural potential architecture based on invariant representations and neural ordinary differential equations. This unified setting ensures that performance differences can be attributed to thermodynamic structure rather than architectural variation. The models are trained and evaluated on three representative inelastic datasets generated from high-fidelity representative volume element simulations: an elastoplastic alloy, a viscoelastic composite, and a rate-dependent crystal plasticity polycrystal. By isolating the role of thermodynamic structure, we assess how restrictions such as duality, normality, operator-based evolution, and convexity influence learnability, expressiveness, stability, and generalization.

cond-mat.mtrl-sci↗

arXiv2026

Berry Curvature Driven Transport in Silicon-Compatible Altermagnetic $α$-MnTe Thin Films

Integrating spin-dependent functionality with mainstream semiconductor technology is a central goal of modern spintronics, yet most candidate materials remain incompatible with silicon-based platforms. Here, we report the direct epitaxial integration of $α$-MnTe thin films on Si(111) via molecular beam epitaxy and demonstrate a robust anomalous Hall effect (AHE) in this silicon-compatible altermagnetic system. Despite the absence of net bulk magnetization, the films exhibit a pronounced hysteretic Hall response, providing transport evidence consistent with finite Berry curvature generated by symmetry breaking in the thin-film geometry. High-resolution structural and spectroscopic characterization confirms phase-pure, epitaxial growth with hexagonal NiAs-type symmetry, while magnetotransport measurements reveal correlated hysteresis in both transverse and longitudinal channels with systematic temperature evolution. First-principles calculations reveal substantial uncompensated Berry curvature arising from the spin-split band structure, consistent with altermagnetic symmetry and the origin of the observed Hall response. These results establish MnTe/Si(111) as a silicon-compatible altermagnetic platform and chart a concrete pathway for embedding Berry-phase-driven functionalities into scalable semiconductor device architectures.

cond-mat.mtrl-sci↗

arXiv2026

All-Optical Control of Interfacial Polarization in MoS$_2$/WSe$_2$ Heterobilayers

All-optical tuning of van der Waals heterostructures with coherent radiation offers a promising path toward ultrafast memory and optoelectronic devices. In the first-principles framework of real-time time-dependent density functional theory, we predict the induction of a persistent, long-lived out-of-plane polarization in MoS$_2$/WSe$_2$ heterobilayers, resonantly driven by intense ultrafast pulses. While weak fields preserve the intrinsic type-II band alignment, intermediate intensities trigger a four-fold enhancement of interlayer charge transfer. By analyzing the high-harmonic generation spectrum, we identify a transition from the perturbative to the strong-field regime inducing photoinduced interfacial polarity. We additionally show that lattice strain, ubiquitously present in heterobilayers, can be used as additional knob to adjust the resonant condition without compromising the permanent dipole induction. Our findings provide a theoretical blueprint for the all-optical manipulation of polar phases in low-dimensional heterostructures at the femtosecond scale.

cond-mat.mtrl-sci↗