Pd–Mn Silica-Supported Catalysts: 2. Description of the Catalytic Sites and Surface Properties for CO and NO Chemisorption

To explain the beneficial effect of Mn addition to Pd catalysts in CO–NO conversion, a detailed study of the surface structure and composition of the catalytic sites was undertaken. The combination of analytical microscopy and EXAFS confirms the presence of bimetallic particles in which manganese is partly alloyed to palladium, as described in Part 1 (A. J. Renouprez, J. F. Trillat, B. Moraweck, J. Massardier, and G. Bergeret, J. Catal.179, 390 (1998)). Monte-Carlo modeling coupled with a simple energetic model and low-energy ion scattering (LEIS) shows that Mn has a larger concentration at the surface of the particles than the mean value measured by chemical analysis. This is attributed to a segregation of this element at the surface. To establish a correlation between the structure of these particles and their reactivity, infrared experiments of CO, NO adsorptions and coadsorption were carried out at both 300 and 573 K. The measured frequencies were interpreted on the basis of density functional theory (DFT) quantum chemical calculations. At 300 K, on pure Pd, CO and NO are located on the same sites and are displaced by each other. On the alloys, both molecules remain partly on the surface on adsorption of the other. DFT calculations show that the Mn atom at the surface induces a significant shift of the NO stretch frequency to lower values, especially if a Mn atom is present in the surface site. On mixed Pd–Mn sites, the chemisorbed NO molecule is found in a geometry quasi-parallel to the surface, with a long N–O bond and a low stretch frequency. This structure is stable on Pd3Mn, but not on pure Pd. This lying down geometry is a precursor for the dissociation. A dual-functional mechanism for the NO reduction, involving Mn oxides with oxygen vacancies and this form of adsorbed NO, is proposed.