ConspectusAtomically dispersed M-N-C catalysts, owing to their high metal utilization and well-defined local structure, have been extensively applied in oxygen reduction reaction (ORR), oxygen evolution reaction (OER), and hydrogen evolution reaction (HER) for clean energy devices, such as fuel cells, metal-air batteries, and water electrolyzers. However, traditional trial-and-error approaches to synthesize M-N-C catalysts are significantly time-consuming and resource/labor-intensive and have a low success rate. With the burgeoning development of artificial intelligence (AI) and related computational techniques (such as density functional theory (DFT) and machine-learning (ML)) in heterogeneous catalysis, theory-driven experimental paradigm exerts a huge potential to guide experimental synthesis of M-N-C catalysts and simultaneously reveal the corresponding catalytic mechanisms, which greatly save research and development costs and would speed up applications of clean energies.In this Account, our emphasis is placed on theory-driven experimental discovery of highly efficient M-N-C catalysts, especially single- and dual-atom catalysts (SACs/DACs) for ORR and OER. Indeed, it is a common goal to quantitatively reveal the structure-property relationship of catalysts. Although energy descriptors depending on the adsorption free energy have been proposed by Nørskov et al., they cannot yet achieve the direct prediction of catalyst performance from the intrinsic structure without DFT calculations. Therefore, structure descriptors have been proposed by our group to achieve the direct prediction of catalyst performance from its intrinsic structure and have been successfully applied to four kinds of M-N-C catalysts, including SACs, axial coordination ligand decorated SACs (ACL-SACs), defect-SACs and DACs, which are summarized here. In addition, the key effects of the axial pre-adsorption microenvironment reconstruction of M-N-C catalysts under working conditions and the O2 adsorption step without proton-electron transfer on catalytic performance are discussed, and a new high-throughput (HTP) screening method was proposed, and a series of highly efficient DACs were screened. To accurately describe the interaction between the catalytic surface and intermediates, the "BASED" theory was proposed, and its origin and importance in surface catalysis are also summarized. Driven by the above structure descriptors, high throughput screening method, and "BASED" theory, a series of M-N-C SACs/DACs (such as Fe-N-C, Co-N-C, Cu-N-C, Fe-Mn-N-C, and Fe-Cu-N-C, Mn-Co-N-C) with outstanding performance in half-cells and related devices were synthesized. In short, the theory-driven experimental synthesis of M-N-C catalysts play a pivotal role in propelling the development of heterogeneous M-N-C electrocatalysts and also provide a new research paradigm to develop high-performance catalysts for other reactions and applications.
Huang et al. (Fri,) studied this question.