Predicting Lattice Phonon Vibrational Frequencies Using Deep Graph Neural Networks
This work addresses the bottleneck of materials screening for researchers by providing a faster alternative to DFT calculations, though it is incremental as it extends existing graph neural network methods to a new property.
The paper tackles the problem of computationally demanding DFT calculations for lattice vibration frequencies by proposing a deep graph neural network algorithm that predicts these frequencies from crystal structures with high accuracy, achieving aggregated R² scores of 0.554 and 0.724 on datasets of 15,000 and 35,552 samples, respectively.
Lattice vibration frequencies are related to many important materials properties such as thermal and electrical conductivity as well as superconductivity. However, computational calculation of vibration frequencies using density functional theory (DFT) methods is too computationally demanding for a large number of samples in materials screening. Here we propose a deep graph neural network-based algorithm for predicting crystal vibration frequencies from crystal structures with high accuracy. Our algorithm addresses the variable dimension of vibration frequency spectrum using the zero padding scheme. Benchmark studies on two data sets with 15,000 and 35,552 samples show that the aggregated $R^2$ scores of the prediction reaches 0.554 and 0.724 respectively. Our work demonstrates the capability of deep graph neural networks to learn to predict phonon spectrum properties of crystal structures in addition to phonon density of states (DOS) and electronic DOS in which the output dimension is constant.