Materials genome project: Mining the ionic conductivity in oxide perovskites

We present in this work a strategy for predicting new oxide perovskites with the potential for achieving high ionic conductivity for applications as a solid oxide fuel cell (SOFC). We employ a throughout multivariate technique based the principal component analysis (PCA) and the partial least square methods (PLS) in order to identify empirical design rules (activation and migration energies) for stable ionic conductors. Among the 892 tested hypothetical oxide perovskites, which yet to be experimentally synthetized, similar to 150 compounds may exhibit an interesting potential as ionic conductors. These results may serve as a map for the design of perovskite-related solid electrolytes.