Machine learning and DFT database for C-H dissociation on single-atom alloy surfaces in methane decomposition

Methane decomposition using single-atom alloy (SAA) catalysts, known for uniform active sites and high selectivity, significantly enhances hydrogen production efficiency without CO₂ emissions. This study introduces a comprehensive database of C-H dissociation energy barriers on SAA surfaces, generated through machine learning (ML) and density functional theory (DFT). First-principles DFT calculations were utilized to determine dissociation energy barriers for various SAA surfaces, and ML models were trained on these results to predict energy barriers for a wide range of SAA surface compositions. The resulting dataset, comprising 10,950 entries with descriptors and energy barriers, as main predictive outcomes, has been validated against existing DFT calculations confirming the reliability of the ML predictions. This dataset provides valuable insights into the catalytic mechanisms of SAAs and supports the development of efficient, low-emission hydrogen production technologies. All data and computational tools are publicly accessible for further advancements in catalysis and sustainable energy solutions.