https://www.cas.cn/syky/202411/t20241122_5040280.shtml
https://www.science.org/doi/10.1126/science.adp6034
The research group of Professor Li Weixue of the University of Science and Technology of China (USTC) summarizef experimental interface interaction data from publications covering 25 metals and 27 oxides. Using an explainable AI algorithm which takes material properties as basic features, they established a governing equation between metal-support interaction and material properties and derived a general theory of metal-oxide interactions. It shows that for late-transition metal catalysts, metal-metal interactions dominate oxide support effects and suboxide encapsulation over metal nanoparticles. Wang et al. performed machine learning on experimental studies of metal nanoparticle adhesion to metal oxides, as well theoretical energetics calculations and molecular dynamics simulations. The key descriptors were the oxophilicity of the supported metal element and the affinity of the supported metal element to the metal element of the oxide support. The authors also showed that strong metal-metal interactions can predict the encapsulation of metal nanoparticles by the oxide support. (Phil Szuromi)
