Abstract
We report a simple method for depositing Au nanoparticles on the surfaces of Cu2O cubes, octahedra, and rhombic dodecahedra with preservation of sharp facets. The purpose of this study is to test whether the conventional model of a metal cocatalyst anchored on a semiconductor particle can always enhance photocatalytic activity of the semiconductor. If so, it is also necessary to determine which facet gives the best activity. This is especially interesting for Cu2O cubes, which have been shown to be inactive toward photocatalysis of methyl orange. The Au-decorated Cu2O crystals have been characterized by various techniques. Au-decorated rhombic dodecahedra and octahedra outperformed their pristine counterparts, while Au-deposited cubes remained inactive. Au-decorated small cubes were also inactive. Furthermore, Au‒Cu2O core‒shell rhombic dodecahedra exhibited intermediate photocatalytic activity between those of rhombic dodecahedra and Au-decorated rhombic dodecahedra. Reaction rate constants for the various Cu2O crystals have been determined. EPR measurements showed highest production of hydroxyl radicals for rhombic dodecahedra and a lower level of radical generation for octahedra. Radicals were not produced from Cu2O cubes upon visible light irradiation. The EPR results match perfectly with photocatalysis observations. A modified band diagram has been constructed to account for the observed facet-dependent photocatalytic activity. The inactivity of the {100} face can be understood by considering its surface layer as constituted of a high barrier blocking photoexcited electrons from reaching the cube surface, so no radicals are produced to cause photodegradation. This work firmly shows that the conventional model needs to be modified taking into account of the surface facets of semiconductors, because sometimes this model can fail completely.