Electrochemical reduction of CO2, if coupled with hydrogen (H2) from (photo-)electrocatalytic water splitting, provides a promising solution for the utilization of the abundant greenhouse gas in the Earth’s atmosphere and the intermittent electrical energy from solar panels and wind turbines for fuels and chemicals that are traditionally derived from petroleum. Unfortunately, there are currently no catalysts capable of carrying out those reactions with both high activity and selectivity. Tailoring the local chemical reactivity of catalyst sites by alloying with other metal atoms in the transition metals and/or their oxides could potentially reduce the energy loss for CO2 reduction and H2O oxidation, and enhance the Faraday efficiency to desired products. While the traditional trial-and-error method for catalyst development usually relies on chemical intuition is time consuming, recent emergence of a descriptor-based materials design approach has significantly improved this and allows the discovery of more efficient catalysts from first-principles. In this talk, I will discuss our mechanistic studies of electrochemical CO2 reduction on metals, and develop a machine-learning enabled screening approach that explores the wide geometrical and chemical phase space for electrocatalysts discovery.

Machine learning approach to electrocatalyst discovery

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.