(594a) First-Principles Computational and Machine Learning Approach to Innovative Design of Electrocatalysts

plays a key role in realization of various energy devices and efficient production of industrial goods. This presentation introduces frontiers of research works for the objective using the first principles-based multiscale computational framework combined with machine learning technique. As the example, nanoscale electrocatalysts in proton exchange membrane fuel cells (PEMFCs) are introduced. It clearly suggests that the method provide new design principles beyond conventional wisdom through identification of fundamental descriptor and mechanisms on the catalysis. Furthermore, combined with machine learning technique the method even can screen best electrocatalysts among a wide range of candidates, eventually rendering a computational materials genome approach to substantially reduce temporal and material cost necessary for the cycle from discovery to commercialization into markets of a new electrocatalyst.