Ab-initio Simulation of Adsorption: Small Molecules on Solid Surfaces

Metaloxides and their surfaces are of outstanding technological and scientific interest because of their applicability as catalysts. A theoretical investigation of these materials and the adsorption process itself with modern ab initio methods gives a fundamental insight to the chemical and physical properties and is therefore inevitable to understand and further on improve catalytic processes. However, for modelling a surface realistically, its attachment to the bulk medium has to be considered as well as its boundary towards the vacuum. Due to the limitations of even modern supercomputers, the sizes of the unit cells have to be kept moderately low (~ 100 atoms). Therefore, modern ab initio methods for surface calculations do not take into account a realistic bulk attachment: Surfaces are modelled by slabs, relying on a sufficient screening of the surface towards the center of the film. The long-ranged Coulomb interaction in rather ionic materials, however, does not vanish in the center of the slabs, even for quite thick films, and might substantially influence the results. Therefore, one of the main aims of the project will be to design a model for embedding in a bulk-like medium which is as well feasible for ab initio calculations providing the important ab inito informations (total energies and forces) but is -on the other side- physically reasonable. The "bulk-reservoir" will be simulated by a bulk-potential bulk wavefunctions, which is calculated by FLAPW, and which is kept frozen further on. A rigorous treatment by Green`s function embedding is therefore not possible given these conditions. A first concept would be to simulate the ionic solid by an fixed external potential determined by the lattice distribution of the ions. Another model -which should be useful also for metallic systems- will involve the fixed charge densities of the bulk system. So far no model along the lines of the proposed project exists. First experiences with fixed external potentials have been gained by applying the free-slab version of FLAPW to the simulation of a charged silver-electrode in an electrochemical medium. Having this in mind, it is obvious that in general a reasonable matching of some different medium to a finite slab preserving the quality of ab initio methods would have far-reaching consequences in several technological and scientific fields. The new model will be tested for some relevant substrate/adsorbate systems, like the adsorption of water on the rutile TiO2 (110) surface, which also has been studied in a thin-film geometry by applying the free-slab full potential augmented plane wave (FLAPW) method.

During the last few years, nanotechnologies, materials design and the development of new catalysts have become a growing field, which need fundamental knowledge of the properties of surfaces and surface reactions, such as stability, catalytic activity and corrosion resistence. Ab-initio methods are a powerful tool to investigate these properties, based on the most fundamental formalism by solving the Schrödinger equation. The calculated results often help to reveal trends which are unresolvable by experiment because of the complexity of the problems. Thus, theory and experiment have to be combined to design catalysts (e.g. for cleaning combustion gases or for producing alternative, decarbonized fuels), to optimize surface proprties (e.g. of biomaterials) or to analyze and possibly improve surface reactions of interest for manufacturing. The calculated results helped to understand some technologically important items more profoundly. Rutile TiO2 is a transition metal oxide known for its wide applications as biomaterial. It forms on the surface of Ti-alloy based bone implants and the surrounding tissue. Furthermore, it is a catalyst for the photodissociation of water, which is a promising technology to produce alternative fuels. Converting solar energy into chemical energy could help to reduce the greenhouse effect caused by the combustion of fossil fuels. Fundamental knowledge of the mechanisms and energetics of water adsorption and possible poisoning of the catalyst`s surface is required to optimize the performance of such a material. Our calculations show that dissociative adsorption of water - which is the first step towards hydrolysis - is favoured by oxygen defects. The existence of such defects was confirmed by experimentally measured and calculated STM images and photoelectron spectra. However, the calculated adsorption energies of Chlorine, one of the most common pollutants in aqueous solutions, are even much higher. This means that the reactive surface sites would immediately be blocked by Cl. The results show that the purity of water is extremely important for the photocatalytic performance of TiO2 . One of the important questions of alternative fuel technology is the storage of hydrogen. Group Vb transition metals like V show an excellent bulk-storage capacity of atomic hydrogen. One of the most important unsolved questions was to find the detailed mechanism of hydrogen dissociation at and diffusion into the bulk material. Because clean vanadium sufaces cannot be prepared experimentally, the influence of the most common surface contaminants, oxygen and carbon, on the adsorption and diffusion process has to be understood. The calculated results show that the maximum hydrogen uptake of the surface, the adsorption energies as well as the diffusion barrier heights significantly depend on the local environment, concerning the kind and coverage density of the pre- adsorbed coadsorbate. Hydrogen adsorption and diffusion into the metal is favoured on surface islands covered by 25-50% of carbon, whereas the presence of oxygen does not change the adsorption energies but increases the energy barrier for diffusion into the bulk. The reverse process (diffusion of hydrogen from the metal bulk through the surface) turned out to be almost independent of the presence of surface contaminants.