Electrocatalytic CO2 Reduction on Amorphous Cu Surfaces: Unveiling Structure-Activity Relationships

We present a computational investigation combining machine learning forcefields (ML-FF) and DFT calculations into the potential of amorphous copper (Cu) surfaces towards electrochemical CO2 reduction (eCO2R) to one-carbon (C1) and two-carbon (C2) products. The “on-the-fly” ML-FF developed for Cu replicate DFT energies and structures, offering a computationally efficient tool for simulating amorphous Cu systems. These ML-FFs were used to generate atomistic amorphous models of bulk and surfaces, and the amorphous bulk Cu exhibited slightly higher stability than crystalline Cu. The amorphous Cu surface provide a wider range of Cu coordination sites (5-9) compared to crystalline Cu, which offered a multitude of active centres for CO2 adsorption. Some of the amorphous surfaces investigated in this study spontaneously activated CO2, evidenced by the stable chemisorption, highlighting their potential for efficient CO2 conversion. The intermediates formed during the eCO2R on amorphous Cu surfaces are stabilized compared to crystalline surfaces, leading to a lower overpotentials, and improved faradaic efficiency. This study demonstrates for the first time theoretically, the potential of amorphous Cu-based catalysts towards sustainable CO2 conversion and paves the way for further research and development in this promising field.