Electrocatalyst design for NO3- and CO2 reduction

Abstract

Electrochemical reduction processes have gained prominence as versatile and sustainable approaches for converting source chemical compounds into valuable products. This process provides distinct advantages, including mild operating conditions, scalability, and compatibility with renewable energy sources. To design advanced electrocatalysts, surface decoration and alloying have been adopted to introduce new active sites to the catalysts and modulate their chemical properties. However, the extensive range of possible element combinations and the associated costs of synthesizing and characterizing these complex materials present challenges for the active discovery of new catalysts. In this work, machine learning and density functional theory (DFT) calculation approaches were employed to design effective electrocatalysts. For example, DFT calculations demonstrated that the surface decoration of Cu2O by the Pd cluster enhanced the activity and selectivity of the catalysts for nitrate reduction reaction (NO3RR) to produce NH3. Additionally, the effect of alloying was investigated by alloying five noble metal elements. Using a machine learning model trained on DFT-calculated data to predict adsorption energy, three high entropy alloys (HEA), IrNiPdPtRh, AgIrPdPtRh, and IrPdPtRhRu, were identified to have carbon dioxide reduction reaction (CO2RR) activity and selectivity better than Cu catalyst. These results provide a proof-of-concept for the efficacy of the two design strategies, highlighting the potential for accelerated catalyst design with machine learning and DFT calculation approach.