Training dependency of neural network interatomic potential for molecular dynamics simulation of Ru-Si-O mixed system

Abstract

We investigated training dependency of neural network interatomic potentials for molecular dynamics simulation of Ru-Si-O mixed system. Our neural network interatomic potential was improved by data augmentation technique for training dataset, including data points of reference energies and forces related to reference structures. We demonstrated that data augmentation technique, focusing on lattice expansion coefficient of bulk structures in the training dataset, requires moderation to ensure optimal training of the neural network interatomic potential. We found that Ru/SiO2 interfaces were accurately represented using the neural network interatomic potential trained with Ru and SiO2 surfaces in addition to Ru/SiO2 interfaces. In case of modeling Ru/SiO2 interfaces include unbonded atoms, training the surfaces with unbonded atoms is effective to generalize the neural network interatomic potential. Our demonstration and finding shed light on a pivotal role of training dataset on development of the neural network interatomic potential for Ru-Si-O mixed system.