Wk Willy Offermans

Catalytic ammonia oxidation on platinum: mechanism and catalyst restructuring at high and low pressure

Catalytic ammonia oxidation over platinum has been studied experimentally from UHV up to atmospheric pressure with polycrystalline Pt and with the Pt single crystal orientations (533), (443), (865), and (100). Density functional theory (DFT) calculations explored the reaction pathways on Pt(111) and Pt(211). It was shown, both in theory and experimentally, that ammonia is activated by adsorbed oxygen, i.e. by O(ad) or by OH(ad). In situ XPS up to 1 mbar showed the existence of NH(x)(x= 0,1,2,3) intermediates on Pt(533). Based on a mechanism of ammonia activation via the interaction with O(ad)/OH(ad) a detailed and a simplified mathematical model were formulated which reproduced the experimental data semiquantitatively. From transient experiments in vacuum performed in a transient analysis of products (TAP) reactor it was concluded that N(2)O is formed by recombination of two NO(ad) species and by a reaction between NO(ad) and NH(x,ad)(x= 0,1,2) fragments. Reaction-induced morphological changes were studied with polycrystalline Pt in the mbar range and with stepped Pt single crystals as model systems in the range 10(-5)-10(-1) mbar.

Computational modeling of catalytic reactivity

Basic molecular concepts in heterogeneous catalysis are illustrated by recent results from periodic and cluster density functional theory (DFT) calculations. For reactions on platinum surfaces, differences in the activation barriers found for ammonia activation are analysed in terms of bond-order conservation principles. The notion of early and late transition-states is introduced and used to understand the difference between step edges versus terrace activation. Another important concept relates to the stereoselectivity of a catalytic reaction. For immobilized organo-metallic complexes, the ligand–reactant interaction dominates the selectivity. We analyse activation by immobilized salen complexes and show that attachment of the catalytically active complex to the wall of a microporous system can dramatically affect the conformation of the ligand and hence selectivity. In zeolites we find that the match of shape and size of pre-transition-state structures with the zeolite cavity dominates the stereochemistry of the reaction.

Structure dependence of Pt surface activated ammonia oxidation

Computational advances that enable the prediction of the structures and the energies of surface reaction intermediates are providing essential information to the formulation of theories of surface chemical reactivity. In this contribution this is illustrated for the activation of ammonia by coadsorbed oxygen and hydroxyl on the Pt(111), Pt(100), and Pt(211) surfaces.

Lateral interactions in o/pt(111): density-functional theory and kinetic Monte Carlo

It has become clear over the last couple of years that interactions between adsorbates are very important for the kinetics of processes in heterogeneous catalysis. We show that we can calculate the nearest-neighbor, next-nearest-neighbor, and linear 3-particle interaction between oxygen atoms on a Pt(111) surface using density-functional theory. Trying to compute more interactions leads to overfitting. Kinetic Monte Carlo simulations show that the calculated interactions give a good description of the desorption kinetics.

Determining interactions between adsorbates on transition metals

The last couple of years it has become more and more clear that lateral interactions (interactions between adsorbates) play an important role in the kinetics of surface reactions. We work on the determination of these interactions, and on methods to predict the kinetics includes these interactions. Our main method is Dynamic Monte Carlo. With this method we can simulate all reactions on a surface with between 104 adn 108 reactives sites for about 102-103 seconds. The simulations yield predictions of the kinetics that are exact for a given model of the surface reactions. As input we need rate constants (prefactors and activation energies) and lateral interactions. These are either obtained by fitting to experimental results or by calculating them using Density-Functional Theory. We use Evolutionary Computation methods to vary and optimize the kinetic parameters when fitting to experiments.