Lowering the barrier to access information-rich transient kinetic data for machine learning methods

Abstract

Transient kinetic data contain a wealth of information about intrinsic features of a catalyst as well as the reaction mechanism. Currently, high volume transient data is underutilized, and data science methods could increase the value of information that can be extracted from this data, integrate experimental with theoretical data sources, and accelerate the pace of catalyst technology advancement. Transient kinetic characterizations with simple probe molecules exhibiting reversible adsorption, irreversible adsorption and bulk-surface diffusion are presented as training components for similar experiments with more complex surface reactions. By increasing the availability and accessibility of transient kinetic data through details of its structure and acquisition, we aim to decrease the barrier for data scientists to apply machine learning methods to this valuable data source.