Direct quantitative assessment of single-atom metal sites supported on powder catalysts

Rational design of effective catalyst demands a profound understanding of active-site structures. Single-atom supported powder catalysts, depicting unique features like enhanced metal dispersion, hold promise in different applications. Here, we present an approach to directly quantify the detailed structural nature of metal sites in single-atom, high surface area, powder catalysts. By combining advanced high-resolution scanning-transmission electron microscopy, deep learning and density functional theory calculations, we determine, with statistical significance, the exact location and coordination environment of Pd single-atoms supported on MgO nanoplates. Our findings reveal a preferential interaction of Pd single-atoms with cationic vacancies (V-centers), followed by occupation of anionic defects on the {001} MgO surface. The former interaction results in stabilization of PdO species within V-centers, while partially embedded Pd states are found in F-defects. This methodology opens a route to the ultimate structural analysis of metal-support interaction effects, key in the design of advanced nanocatalysts for sustainable and energy-efficient processes. Metal active sites are important in catalysts to function in various applications. Here, the detailed structural nature of single-atom metal sites is directly quantified using a combination of high-resolution microscopy, deep learning, and theoretical calculation methods.