A novel approach to designing supported metal oxide catalysts

Supported metal oxides are used as catalysts in a variety of applications including the partial oxidation of alcohols, selective catalytic reduction (SCR) of NO _x, oxidative dehyrogenation of alkanes and hydrodesulfurization. The role of the support in influencing the reactivity of the overlayer, beyond just the provision of mechanical stability and high surface area, has been recognized. The specific nature of the support and its properties, such as the electronegativity of the support cation, has been proposed to affect the catalytic activity of the overlayer. It has also been shown that modification of the support of the overlayer, by doping, alters the catalytic performance. Considering that several transition metal oxides that are typically used as catalysts, including V ₂O ₅ and TiO ₂, are (wide band-gap) semiconductors, we present a novel approach to understanding the performance of and designing supported catalysts - from the perspective of semiconductor heterojunctions. This approach also allows us to assess the effect of doping the catalyst layers with altervalent ions on the reactivity of the supported catalyst. The reactivity of the surface is known to depend on the surface Fermi level position. The supported catalyst is considered as a heterostructure with a thin overlayer. Heterojunction physics is used to estimate the Fermi level position at the surface of a supported catalyst for various doping levels of the substrate and for various thicknesses of the overlayer. The Fermi level position is then linked to the reactivity through empirical correlations involving surface acidity values. V ₂O ₅/TiO ₂ and partial oxidation of methanol to formaldehyde were chosen as the test catalytic system and test reaction respectively. Our calculations indicate that for the test system, higher n-type doping of the TiO ₂ substrate and lower thickness of the V ₂O ₅ layer lead to higher methanol oxidation rates, as can be expected. Fitting of experimental data to the model gives the value of a parameter that is indicative of the ionic/covalent nature of the reaction.