Sintering and reactivity of model oxide-supported catalysts: platinum/zinc oxide(000-1)-oxygen and palladium/aluminum oxide(0001)

The reactivity and selectivity of oxide-supported metal nanoparticle catalysts are highly dependent on the size of the active metal clusters. This dissertation focuses on the reactivity and sintering of Pt and Pd particles on the ZnO(000-1) and alpha-Al2O3(0001) surfaces, respectively.; Low-energy ion scattering (LEIS) shows that Pt grows on the oxygen-terminated ZnO(000-1) surface in the critical coverage growth mode, with onset of 3D island growth, thetac, occurring at 0.5 ML coverage. Temperature-programmed LEIS (TP-LEIS) shows real time sintering of the metal film by following the Pt LEIS intensity as the sample is heated. TP-LEIS showed that the Pt islands on the ZnO(000-1) surface start sintering at 600 K. Perdeuterated cyclohexane, dosed onto Pt islands of known thickness, shows higher dehydrogenation probability on the 2D islands, suggesting that thinner, smaller islands are more reactive than the larger ones. Perdeuterated benzene dosed on this model catalyst surface confirms that the cyclohexane dehydrogenation goes through a benzene intermediate, as on Pt(111). There is also evidence that oxygen vacancies on the ZnO(000-1) surface are active for dehydrogenation of cyclohexane and benzene.; Pd on the alpha-Al2O3(0001) surface also shows critical coverage growth mode, with thetac = 0.25 ML. Unlike Pt/ZnO, Pd on alumina starts sintering below 400 K, possibly even at room temperature. After annealing to 900 K, the Pd islands cover ∼50% less of the oxide surface than they did at room temperature. Exposure of the Pd islands to O2 does not change their sintering behavior. Non-contact atomic force microscopy (NC-AFM) of the Pd-covered alumina surface showed nearly constant particle density at room temperature for Pd coverages from 0.2 to 1.0 ML, ∼2 x 1012/cm2. Annealing to 1000 K results in larger clusters and a decrease in particle density to ∼50% of the room temperature density. Because of the insulating nature of alumina and the long-range van der Waals forces utilized in NC-AFM, the smallest clusters, especially 2D islands, are often difficult, if not impossible, to image. When the NC-AFM images clusters, it tends to show them as wider and shorter than they really are.