Identification of oxygen diffusion mechanisms in Nd1−xAExBaInO4−x/2 (AE = Ca, Sr, Ba) compounds through molecular dynamics

Abstract

Molecular dynamics simulations have been widely adopted to study oxygen ion diffusion mechanisms in materials for application in solid oxide fuel cells. Indeed, understanding the fundamental aspects of oxygen diffusion is important to develop new materials for this application. In this work, Nd_1−xAE_xBaInO_4−x/2 (AE = Ca, Sr, Ba) compounds have been studied by MD simulations focusing on oxygen diffusion mechanisms. Two general clustering methods were used, namely a convex hull classification method and a DBSCAN machine learning algorithm, to identify oxygen ion diffusion pathways. Here, relevant details are provided for an efficient use of these two approaches during MD analysis of ion conductors. The calculations show that Ca is the most favorable dopant for substituting Nd in NdBaInO₄, while Ba is the least desired. Indeed, the substitution of Nd by Ca hardly changes the pristine lattice parameters of NdBaInO₄ and leads to the highest oxygen diffusion coefficient compared to other dopants. The oxygen vacancies induced by doping mainly locate on two specific oxygen sites over four oxygen sites available. Concerning the diffusion process, jumps involving these two sites play the main role and are associated with smaller migration enthalpies. For the main diffusion path, ions migrate along the b (2 routes) and c (4 routes) directions. Some other oxygen sites can be considered as barriers for the diffusion process inducing a strong anisotropy in the diffusion process. Additionally, the residence time analysis of oxygen ions confirms that ions at different sites have different motion abilities. As a whole, the approach presented here can be extrapolated to other ion conductors for gaining detailed information about the diffusion process.