Establishing and Analysis of Perovskite Photocatalyst Database

Abstract

Hydrogen is one of the ideal alternatives to current energy sources, most of which are fossil fuels. And one of the eco-friendly ways to produce hydrogen is to splitting water using a photocatalyst. Photocatalysts vary in performance depending on materials, synthesis methods, and experimental conditions, and many papers have been published on them. To use this vast amount of data, a database was built and analyzed using machine learning. A database was constructed for 609 ABOx-type perovskite materials, and the data were analyzed using association rule mining and t-SNE techniques. Through this, the tendency of the conditions that generate high hydrogen production was confirmed and the direction of perovskite photocatalyst research was suggested.