| Summary: | Methane activation over molybdenum trioxide and gold catalysts supported on magnesium oxide were studied employing state of the art computational modelling. The localisation of electrons at a terminal oxygen point defect on MoO3(010) was well represented by DFT+U whereas conventional DFT failed to localise the two unpaired electrons on Mo. The hybrid DFT functional PBEO was employed within a cluster model and a spin population of 1.96 was calculated for Mo and used to fit the Dudarev's parameter U-J at a value of 6.3 eV. The asymmetric oxygen centre was found to be the preferred adsorption site for both hydrogen and methyl radical adsorption, losing its asymmetric character. Mo exhibited the weakest interaction with the adsorbates and methyl was found to be weakly adsorbed between Mo and its closest asymmetric oxygen, suggesting mobility on the surface. A transition state for methane activation was located with hydrogen and carbon respectively interacting with terminal oxygen and Mo. A Bader charge analysis of this transition state coupled with a spin population analysis indicated a heterolytic cleavage. An activation energy of 129.7 kJ mol"1 was calculated using the NEB method. On MgO(001), atomic gold was found to preferentially adsorb at five-coordinated oxygen sites and the Au-O bond was partially covalent in nature. F centres represented anchoring points with increased adsorption energy and charge transfer to Au. A three-dimensional growth was favoured over two-dimensional growth and the adsorbed Auio clusters were negatively charged at the interface. Atomic gold on the oxygen kink site of MgO(1310) presented an adsorption energy 0.30 eV higher than that on a five-coordinated oxygen centre of MgO(001). Auio was also found to strongly interact with the oxygen kink site of MgO(1312). Hydroxyl groups were employed to tailor the gold oxidation state and CO was used as a probe molecule. For both atomic gold and Auio, the CO stretching vibrational frequency increased with the oxidation state. CO had an unfavourable binding energy on Au(III) but was strongly adsorbed on Au(I). An increase in the number of adsorbates reduced the negative charge of Auio, also increasing the CO stretching vibrational frequency and bringing it closer to the experiment. It is thus expected that CO will bind to Au8+ particles in the working catalyst.
|
|---|
