Computational Study of Single-Atom Catalysts Supported on Transition Metal Nitrides

Debit Subedi, Bipin Lamichhane, and Shyam Kattel

Florida A&M University, Florida A&M University, Florida A&M University

Single-atom catalysts (SACs), which possess the combined advantages of homogeneous and heterogeneous catalysts, are an emerging class of materials that reach near 100% metal utilization and offer opportunities to manipulate the binding energies of reaction intermediates, thus providing a window of opportunity to tune the activity and selectivity of catalysts. First-principles density functional theory (DFT) calculations were performed to explore the formation of Pt, Pd, Rh, Ir, Ru, and Au SACs on the low index (111) and (100) facets of 29 transition metal (TM = 3d-5d) nitrides. Our results show that the formation of many of these SACs is thermodynamically favorable on the polar TM terminated (111)  and  non-polar (100) surfaces of TMNs, suggesting the feasibility of their experimental synthesis. The DFT calculated surface energies and formation energies of SACs were utilized to develop simple regression machine learning (ML) models using the readily available elemental properties of transition metals and nitrogen as features. Our preliminary results show that ML methods are capable of predicting the SAC formation energies with reasonable accuracy. Our future work will focus on improving the ML models to accurately predict the SAC formation energy with accuracy comparable to DFT calculations.