High-Accuracy Neural Network Interatomic Potential for Silicon Nitride

In the field of machine learning (ML) and data science, it is meaningful to use the advantages of ML to create reliable interatomic potentials. Deep potential molecular dynamics (DEEPMD) are one of the most useful methods to create interatomic potentials. Among ceramic materials, amorphous silicon nitride (SiN(x)) features good electrical insulation, abrasion resistance, and mechanical strength, which is widely applied in industries. In our work, a neural network potential (NNP) for SiN(x) was created based on DEEPMD, and the NNP is confirmed to be applicable to the SiN(x) model. The tensile tests were simulated to compare the mechanical properties of SiN(x) with different compositions based on the molecular dynamic method coupled with NNP. Among these SiN(x), Si(3)N(4) has the largest elastic modulus (E) and yield stress (σ(s)), showing the desired mechanical strength owing to the largest coordination numbers (CN) and radial distribution function (RDF). The RDFs and CNs decrease with the increase of x; meanwhile, E and σ(s) of SiN(x) decrease when the proportion of Si increases. It can be concluded that the ratio of nitrogen to silicon can reflect the RDFs and CNs in micro level and macro mechanical properties of SiN(x) to a large extent.