Machine Learning-Aided Materials Design Platform for Predicting the Mechanical Properties of Na-Ion Solid-State Electrolytes

Na-ion solid-state electrolytes (Na-SSEs) exhibit high potential for electrical energy storage owing to their densities and low manufacturing cost. However, mechanical properties that are critical maintaining structural stability at the interface still insufficiently understood. In this study, a machine learning-based regression model was developed predicting of Na-SSEs. As training set, 12,361 materials were obtained from well-known database (Materials Project) represented with respective chemical descriptors. The surrogate exhibited remarkable accuracies (R2 score) 0.72 0.87, mean absolute errors 11.8 15.3 GPa shear bulk modulus, respectively. This then applied predict 2432 Na-SSEs, which have been validated first-principles calculations. Finally, optimization process performed develop an ideal screening platform by adding minimized data wherein prediction uncertainty is reduced. We believe proposed in study can accelerate search Na-SSEs minimum