▎ 摘　　要

We propose via first-principles calculations that graphene tailored with specific structural or foreign-atom defects can be used as catalysts for O-3 decomposition. Compared to the pristine graphene, we show that introduction of thermodynamically stable surface defects, that is, Stone-Wales, 555-777 divacancy or nitrogen atoms to graphene, can drastically lower the chemisorption energy barrier of O-3. The ozonides once formed can evolve into epoxide or ketone-like intermediate structures on graphene, which assist in the sustainable conversion of O-3 to O-2. The minimal energy needed to complete the O3 decomposition cycle on different graphene substrates has the order of E555-777 (0.22 eV) < EN-doping (0.33 eV) < EStone-Wales (0.61 eV) < E-pristine graphene (1.08 eV). As two most promising catalysts, 555-777 divacant and N-doped graphene shows a clear adsorption selectivity to O-3 under the ambient conditions. N-doped graphene outperforms 555-777 divacant graphene at the stage of physisorption in that the former catalyst separates O-3 from the ambient gases by a much larger adsorption energy. These results and conclusions are a pivotal and necessary step that should stimulate both the proof-of-principle experiments and the computational search of metal-free catalysts for O-3 decomposition.