Progress in prediction of photocatalytic CO2 reduction using machine learning approach: A mini review

Abstract

The rapid progression of industrial revolution has contributed to an increase in greenhouse gases (GHGs) emission, particularly carbon dioxide (CO₂), exacerbating global warming and placing unwanted pressure on limited energy resources, ultimately causing energy scarcity. Achieving carbon neutrality has been recognized as one of the most pressing goals in addressing the global greenhouse effect. The photocatalytic conversion of CO₂ into sustainable, high-value chemicals and fuels has emerged as an effective strategy to combat the issues of global warming and energy scarcity on a worldwide scale. However, the traditional trial-and-error approach renders the process costly, time-intensive, and challenging to understand the comprehension of the underlying mechanisms governing reaction pathways, complicating the discovery of groundbreaking photocatalysts with improved performance for photocatalytic CO₂ reduction (PC-CO₂R). Emerging Machine learning (ML) techniques have significantly improved the predictive performance and operational efficiency of photocatalytic systems. These models can handle extensive experimental datasets, optimize operational parameters, and provide insights into CO₂ reduction (CO₂R) mechanisms. ML models also simplify the material selection, reducing experimental validation time and thereby improving the overall efficiency of PC-CO₂R. This comprehensive review provides an insight into the current progress in ML-guided prediction of PC-CO₂R. The fundamentals of PC-CO₂R and the significance of utilizing ML models as a facilitative tool in photocatalysis are critically discussed in this review. Furthermore, the review outlines the commonly employed ML algorithms, presents recent progress, and identifies crucial parameters influencing PC-CO₂R efficiency. The study also highlights the opportunities and challenges associated with the design of ML-assisted photocatalysts, which need to be addressed to fully harness the potential of ML for converting CO₂ into sustainable energy sources.