Advancing electrocatalyst discovery through the lens of data science: State of the art and perspectives

Abstract

The integration of data science into electrocatalysis has revolutionized the discovery of high-performance catalysts for sustainable energy applications. To emphasize the role of data science and guide future research in electrocatalyst design, this mini-review traces the evolution from low-dimensional data science—rooted in density functional theory (DFT) descriptors such as d-band center and binding/adsorption energies—to high-dimensional analytics powered by large-scale computational datasets and machine learning (ML). First, DFT-derived parameters establish predictive volcano models for various electrochemical reactions, linking atomic-scale descriptors to macroscopic performance within the framework of low-dimensional data science. Meanwhile, with the development of large-scale datasets, ML deciphers complex structure–property relationships, accelerating the design of promising electrocatalysts. Additionally, machine learning potentials (MLPs) bridge quantum precision and scalability, not only accelerating thermodynamic adsorption energy calculations but also enabling simulations of dynamic catalytic mechanisms more efficiently. Finally, we discuss emerging opportunities to deepen data science’s impact. This mini-review highlights the transformative role of data science in bridging theoretical insights, computational efficiency, and experimental validation, ultimately accelerating the design of next-generation electrocatalysts for a sustainable energy future.