Mishra, Adesh Rohan | Search NASA

DOE OSTI2026

Mechanistic origin of solvent-dependent thermal stability in sodiated Sn anodes for sodium-ion batteries

Understanding the thermal stability of high-energy density alloy anodes is critical for the safe deployment of sodium-ion batteries (SIBs). Here, accelerating rate calorimetry (ARC), post-mortem characterizations, and density functional theory (DFT) calculations are combined to understand the thermal reactivity of fully sodiated Sn, Sn-hard carbon (HC) blends, and HC anodes in carbonate- and ether-based electrolytes. ARC measurements show that propylene carbonate (PC) causes earlier self-heating rate (SHR) onset and higher reactivity than tetraethylene glycol dimethyl ether (TEGDME), indicating inferior thermal stability. Sodiated Sn exhibits better thermal stability than sodiated HC, while Sn-HC blends show intermediate behavior that improves with increasing Sn content. Post-ARC analyses reveal desodiation of Na15Sn4 to metallic Sn with particle coalescence, whereas Sn-HC blends and HC retain finer morphologies. PC promotes Sn oxidation to SnO, while TEGDME suppresses oxide formation; NaPF6-containing electrolytes additionally form NaF. DFT calculations show that PC adsorption lowers Na extraction energy and enhances interfacial electronic interactions, facilitating Na release and reductive decomposition. These results establish a direct correlation between solvent-dependent reaction pathways and thermal stability in SIB alloy anodes.

Accelerating rate calorimetry↗

DOE OSTI2025

Materials Graph Library (MatGL), an open-source graph deep learning library for materials science and chemistry

Graph deep learning models, which incorporate a natural inductive bias for atomic structures, are of immense interest in materials science and chemistry. Here, we introduce the Materials Graph Library (MatGL), an open-source graph deep learning library for materials science and chemistry. Built on top of the popular Deep Graph Library (DGL) and Python Materials Genomics (Pymatgen) packages, MatGL is designed to be an extensible “batteries-included” library for developing advanced model architectures for materials property predictions and interatomic potentials. At present, MatGL has efficient implementations for both invariant and equivariant graph deep learning models, including the Materials 3-body Graph Network (M3GNet), MatErials Graph Network (MEGNet), Crystal Hamiltonian Graph Network (CHGNet), TensorNet and SO3Net architectures. MatGL also provides several pre-trained foundation potentials (FPs) with coverage of the entire periodic table, and property prediction models for out-of-box usage, benchmarking and fine-tuning. Finally, MatGL integrates with PyTorch Lightning to enable efficient model training.

chemistry↗