Deep Neural Network for Accurate and Efficient Atomistic Modeling of Phase Change Memory

This letter presents a general-purpose fully-atomistic method to simulate phase change memory (PCM), by combining density functional theory (DFT) and deep neural network (DNN). Its maximum calculation error of atomic forces is about 10(-1) eV/angstrom, which is 1-2 orders of magnitude more accurate than state-of-art artificial neural network (ANN) in PCM literature (over 10(1) eV/angstrom). Its simulation time, t(s), scales linearlywith the number of atoms n(a) (t(s) proportional to n(a)), which is more efficient than DFT (t(s) proportional to n(a)(3)) widely used to model PCM, leading to approximately 2, 4, 6 orders of magnitude reduction of modeling time when n(a) approximate to 10(1), 10(2), 10(3), for instance. Its efficiency and accuracy may be useful to develop next-generation atomistic modeling tools to enable in-depth optimization of PCM.