Tuning the properties of copper-based catalysts based on molecular in situ studies of model systems.

Studying catalytic processes at the molecular level is extremely challenging, due to the structural and chemical complexity of the materials used as catalysts and the presence of reactants and products in the reactor's environment. The most common materials used on catalysts are transition metals and their oxides. The importance of multifunctional active sites at metal/oxide interfaces has been long recognized, but a molecular picture of them based on experimental observations is only recently emerging. The initial approach to interrogate the surface chemistry of catalysts at the molecular level consisted of studying metal single crystals as models for reactive metal centers, moving later to single crystal or well-defined thin film oxides. The natural next iteration consisted in the deposition of metal nanoparticles on well-defined oxide substrates. Metal nanoparticles contain undercoordinated sites, which are more reactive. It is also possible to create architectures where oxide nanoparticles are deposited on top of metal single crystals, denominated inverse catalysts, leading in this case to a high concentration of reactive cationic sites in direct contact with the underlying fully coordinated metal atoms. Using a second oxide as a support (host), a multifunctional configuration can be built in which both metal and oxide nanoparticles are located in close proximity. Our recent studies on copper-based catalysts are presented here as an example of the application of these complementary model systems, starting from the creation of undercoordinated sites on Cu(111) and Cu2O(111) surfaces, continuing with the formation of mixed-metal copper oxides, the synthesis of ceria nanoparticles on Cu(111) and the codeposition of Cu and ceria nanoparticles on TiO2(110). Catalysts have traditionally been characterized before or after reactions and analyzed based on static representations of surface structures. It is shown here how dynamic changes on a catalyst's chemical state and morphology can be followed during a reaction by a combination of in situ microscopy and spectroscopy. In addition to determining the active phase of a catalyst by in situ methods, the presence of weakly adsorbed surface species or intermediates generated only in the presence of reactants can be detected, allowing in turn the comparison of experimental results with first principle modeling of specific reaction mechanisms. Three reactions are used to exemplify the approach: CO oxidation (CO + 1/2O2 → CO2), water gas shift reaction (WGSR) (CO + H2O → CO2 + H2), and methanol synthesis (CO2 + 3H2 → CH3OH + H2O). During CO oxidation, the full conversion of Cu(0) to Cu(2+) deactivates an initially outstanding catalyst. This can be remedied by the formation of a TiCuOx mixed-oxide that protects the presence of active partially oxidized Cu(+) cations. It is also shown that for the WGSR a switch occurs in the reaction mechanism, going from a redox process on Cu(111) to a more efficient associative pathway at the interface of ceria nanoparticles deposited on Cu(111). Similarly, the activation of CO2 at the ceria/Cu(111) interface allows its facile hydrogenation to methanol. Our combined studies emphasize the need of searching for optimal metal/oxide interfaces, where multifunctional sites can lead to new efficient catalytic reaction pathways.