Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 8: Materials for the Storage and Conversion of Energy (joint session MM/KFM)
MM 8.4: Vortrag
Montag, 17. März 2025, 18:00–18:15, H22
Modelling LLZO Grain Boundaries with Amorphous Domains by Adaptively Trained Machine-Learning Interatomic Potentials — •Yuandong Wang, Yute Chan, Hao Wan, Kyeonghyeon Nam, Karsten Reuter, and Christoph Scheurer — Fritz-Haber-Institut der MPG, Berlin
Garnet Li7La3Zr2O12 (LLZO) is a highly promising solid state electrolyte (SSE) for lithium batteries. However, its practical application faces challenges, primarily arising from Li dendrite formation and the impact of grain boundaries (GBs) on Li transport and stability. Amorphous LLZO combines several desirable properties like blocking Li dendrite growth, high Li mobility and high electronic impedance. Controlling amorphous domains between crystalline grains could therefore offer an intriguing approach to tune electrolyte performance. For this, an atomistic understanding of the interplay between composition, structure and the properties of LLZO glass-ceramics is crucial.
This study introduces a Machine Learning Interatomic Potential (MLIP) tailored to accurately represent amorphous and GB structures in LLZO. Developed through an iterative training protocol using simulated annealing, this MLIP includes diverse structures in its training set, ensuring comprehensive modeling of complex LLZO phases. The MLIP enables large-scale molecular dynamics simulations, allowing the construction of realistic amorphous and GB models, and providing a foundation for in-depth analysis of LLZO structural and electrochemical behavior.
Keywords: all-solid-state battery; interface; amorphous structure; machine learning; Li diffusion