Machine Learning for Phonons in Inorganic Crystals

The advent of artificial intelligence (AI) and machine learning (ML) has revolutionized many aspects of modern science and technology and has sparked significant interest in the material science community in recent years. Despite some early deployment of AI/ML in thermal science area, the power of AI has not been maximized. Existing ML methods for predicting phonon properties of crystals are limited to either small amount of training data or a material-to-material basis, primarily due to the exponential scaling of model parameters with the number of atomic species or elements. This renders high-throughput infeasible when facing large-scale new materials. In the first part of this talk, I will introduce some state-of-the-art AI/ML approaches for predicting phonon transport in crystals. Particular focus will be our recently developed Elemental Spatial Density Neural Network Force Field (Elemental-SDNNFF) with abundant atomic level environments as training data. Benefiting from the innovative architecture of the algorithm, sub-trillion atomic data can be integrated to train a single deep neural network for predicting complete phonon properties of >100,000 inorganic crystals spanning 63 elements in the periodic table. In the second part of the talk, I will illustrate a new look of ion (mass) transport from phonon perspective, focusing on a quantitative and strong positive correlation between materials’ lattice dynamics feature and ion mobility. The newly identified lattice dynamics features are expected to advance the discovery and design of novel superionic conductors that can be used as electrolytes in all-solid-state energy storage applications, such as by incorporating phonon features into AI/ML algorithms and workflows.