Surface structure of catalytically-active ceria nanoparticles

Catalytic mechanisms, and therefore activity, depend on the structure of catalyst surfaces. In turn, surfaces may reconstruct and/or exhibit local configurations that vary from bulk composition and structure. CeO2 (ceria) is a redox catalyst of interest in numerous automotive, energy and, increasingly, biomedical applications. Previous studies aimed at understanding catalytic mechanisms on ceria have limited consideration to systems with bulk-like stoichiometric or sub-stoichiometric surfaces. Here we summarize previous computational studies on ceria surfaces, nanoclusters, and nanoparticles, and highlight challenges in constructing physically-representative ceria nanoparticle (CNP) structures. Setting aside assumptions of bulk-like stoichiometric or sub-stoichiometric ceria surface terminations, we report results of DFT + U calculations and show that sufficiently small CNPs are not bulk-terminated, but rather are stabilized by the formation of Oxq groups (−2 ⩽ q ⩽ 0, x ⩽ 3) at corners, edges, and {1 0 0} facets. These surface structures, not the annihilation and regeneration of O-vacancies, may directly control reduction/oxidation catalysis at CNPs below a critical size. As anion groups other than Oxq groups could be incorporated in stable CNP surfaces, this suggests the possibility of tailoring small CNP structures and mechanisms for particular catalytic reactions.