AI and ML for selecting viable electrocatalysts: progress and perspectives

Abstract

The urgent need to address the current energy crisis and combat excess carbon dioxide (CO₂) in the atmosphere has emphasised the importance of transitioning to sustainable energy sources. Fossil fuel reliance has significantly contributed to global warming, underscoring the need for solutions that can mitigate CO₂ emissions effectively. One promising technology is electrolysis, which not only removes CO₂ from the atmosphere but also generates clean and renewable fuels with net-zero CO₂ emissions. By utilising a direct electric current to drive non-spontaneous oxidation and reduction reactions, electrolysis has the potential to convert CO₂ into more useful chemical species. Another electrochemical technique, water splitting focuses on producing molecular oxygen and hydrogen, offering an environmentally friendly alternative to traditional fossil fuel-dependent methods. However, electrolysis is often plagued by sluggish kinetics; thus, electrocatalysts are used to enhance reaction rates. Nanoscale electrocatalysts with complex structures have demonstrated excellent catalytic activity and selectivity for the desired products under certain reaction conditions; however, determining the most suitable shape and their corresponding reaction parameters can be an extremely tedious and time-consuming process. Recent studies have shown that artificial intelligence (AI) and machine learning (ML) can greatly simplify this process, saving both time and resources. This review highlights the role of AI and ML in optimising catalyst shape and reaction conditions for water splitting and CO₂ reduction, fostering the development of clean, efficient energy technologies for a sustainable future.