Evolutionary Algorithm Directed Synthesis of Mixed Anion Compounds LaF2X (X = Br, I) and LaFI2.

Mixed-anion compounds have attracted growing attentions, but their synthesis is challenging, making a rational search desirable. Here, we explored LaF3-LaX3 (X = Cl, Br, I) system using ab initio structure searches based on evolutionary algorithms, and predicted LaF2X and LaFX2 (X = Br, I), which are respectively isostructural with LaHBr2 and YH2I, consisting of layered La-F blocks with single and double ordered honeycomb lattices, separated by van der Waals gaps. We successfully synthesized these compounds: LaF2Br and LaFI2 crystallized in the predicted structure, while LaF2I is similar to the predicted one but with different layer stacking. LaF2I exhibits fluoride ion conductivity comparable to that of non-doped LaF3, and has the potential to show better ionic conductivity upon appropriate doping, given the theoretically lower diffusion energy barrier and the presence of soft iodine anions. This study shows the structure prediction using evolutionary algorithms will accelerate the discovery of mixed-anion compounds in future, in particular those with an ordered anion arrangement.