Predicting Fast Oxygen Ion Conductors Via Descriptors for Migration Barrier Based on Electronic Structure and Lattice Dynamics

Abstract

Solid-state oxygen ion conductors are of paramount importance to solid oxide fuel cells and electrolytic cells. However, with the oxygen ion conductors proposed to date, these technologies require too high (close to 800 oC) temperatures to operate, leading to technological challenges and high costs. Therefore, it is of primary importance to design faster oxygen ion conductors at intermediate and low temperatures (below 400 oC) that can enable sustainable energy technologies to mitigate climate change. Descriptors for oxygen ion conduction are highly desirable to identify promising materials, which can be done by leveraging fast-growing statistical and machine learning approaches for screening of material databases [1]. Recently, computational tools combined with experiments have allowed researchers to gain atomistic insights and relate ion transport to features of the lattice dynamics [2] and electronic structure [3] of a given material. In this work, building upon the previous literature, we leverage both the lattice dynamics and electronic structure to predict the migration barrier for oxygen migration in pure and mixed electronic ionic conductors and search for promising oxygen ion conductors.Using density functional theory simulations, we model oxygen migration in perovskite, rutile, fluorite, and Ruddlesden-Popperoxides. We analyze oxygen ion transport in pure and mixed electronic ionic conductors and oxygen vacancies with and without charge. We study changes in the electronic structure during oxygen migration within the solid host lattice in different oxides. By analyzing trends of migration barrier and electronic density of states, we relate the migration barrier to the electronic structure of the transition state. Our results show that the migration barrier depends on the capability that the host lattice has of screening charge associated with the mobile ion in the transition state, which if increased would lower the localization of charge around the mobile ion and in turn the oxygen migration barrier. In pure ionic conductors – where no charge is associated with the mobile ion - the migration barrier is largely dependent on the M-O strength. Based on the mechanistic understanding derived by these initial analyses, we further relate oxygen ion conductivity to the lattice dynamics. We show that in pure ionic conductors, increasing migration barrier follows trends of increasing oxygen vibrational frequency based on phonon-electron coupling, which also relates the prefactor of Arrhenius law with the migration barrier. However, when trying to predict trends of the migration barrier across different crystal structures or mixed ionic conductors, no single descriptor can capture the complex physiochemistry behind oxygen migration. Thus, we leverage statistical learning tools to identify multivariate descriptors and analyze specific features of lattice moieties that improve oxygen ion conduction. The results of this work provide mechanistic understanding of oxygen ion migration and strategies to accelerate the designs of fast oxygen ion conductors.