Angelos Michaelides receives one of four Honorable Mention awards associated with the IUPAC Prize for Young Chemists, for his Ph.D. thesis work entitled "Towards an understanding of simple reactions in Heterogeneous Catalysis"

Current address (at the time of application)

The Queen's University of Belfast
Department of Chemistry
Belfast BT95AG, Northern Ireland

Tel: + 028 28273631 ext . 2270
Fax: +028 28273719
E-mail: [email protected]

Academic degrees

Ph.D. Thesis

Title Towards an understanding of simple reactions in Heterogeneous Catalysis
Adviser Dr Peijun Hu
Thesis Committee Robbie Burch, School of Chemistry, The Queen's University of Belfast; Richard Catlow, Royal Institution of Great Britain.

Essay

Few fields of scientific endeavour can have contributed so dramatically to human existence over the last forty years as heterogeneous catalysis. Over 80 % of industrial chemical production proceeds with the aid of a solid catalyst, and heterogeneous catalysis represents 20-30 % of global GNP annually. The key to future advances in heterogeneous catalysis must be to understand surface processes at an atomic level, and the surface catalysed reaction is of paramount importance. Aside from the certain fundamental significance of such an understanding, it would facilitate the development of designer catalysts with enhanced reactivities and selectivities.

Acquiring this knowledge is not, however, an easy task. To determine microscopic reaction pathways high temporal (picosecond) as well as high spatial (Angstrom) resolution is simultaneously required, making experimental investigations in this area problematic. Computer simulations - chemistry in a "virtual laboratory" - are, therefore, an attractive alternative. Thanks to recent methodological and computational developments it is now possible to apply highly accurate quantum mechanical techniques to the study of catalytic reactions. The accuracy of one particular technique, density functional theory (DFT), has improved dramatically over recent years with the effect that DFT calculations now yield precise structural and energetic information. This makes DFT particularly amenable to the study of reaction pathways and barriers.

My studies involved the application of state-of-the-art DFT methods to surface catalysed reactions of relevance to heterogeneous catalysis. Our approach was to model reactions in simulation cells. Each cell contains 20-30 atoms, repeated infinitely in three dimensions. By determining the electronic structures of these infinite systems, accurate descriptions of catalytic surfaces, with arrays of reactants adsorbed upon them, can be obtained. Simply by reducing the distance between the reactants and obtaining updated electronic and molecular structures one can determine both low energy reaction pathways and their associated activation energies. A variety of reaction systems have been examined. I will present three examples, each of particular importance, which illustrate well the sorts of issues that can be addressed by theoretical chemistry.

A new mechanism for hydrogenation
Hydrogenation is a process of considerable scientific and technological importance. The conventional view in heterogeneous catalysis had been that H atoms adsorbed on the surface of the catalyst carry out the hydrogenation reactions. Recently, however, a group from M.I.T. revealed an exciting new hydrogenation mechanism. It was demonstrated that H atoms absorbed within a Ni catalyst (subsurface H atoms) were more reactive than H atoms adsorbed on the surface of the catalyst (surface H atoms). It was shown, moreover, that certain hydrogenation reactions are only possible when subsurface H atoms are involved. It was not clear, however, how this new class of reaction proceeds or why subsurface H atoms were so reactive. In the first application of DFT to reactions involving subsurface species, we determined microscopic reaction pathways for reactions involving surface and subsurface H atoms (Figure 1). Quantum mechanics was used to reveal precisely how subsurface H atoms diffuse and react at the selvedge of a Ni catalyst. By careful examination of each reaction pathway and barrier a model was proposed to explain the high reactivity of subsurface H atoms. Essentially the bond strength between a subsurface H and the Ni catalyst is weaker than that between a surface H and the Ni catalyst, due to greater nuclei-nuclei repulsion experienced by subsurface H atoms. As a consequence subsurface H atoms are metastable with respect to surface H and it is this metastability that accounts for the observed reactivity. This model is of some importance since it can be applied generally in this new important class of catalytic reactions. Furthermore, it will assist in the identification of other systems in which subsurface species have increased reactivities, thus facilitating the identification of new catalysts.

Figure 1. Microscopic reaction pathway for a subsurface H (blue) reacting with methyl to produce methane on the surface of a Ni (grey) catalyst.

Catalysis: back to the beginning
On contact with platinum, hydrogen and oxygen react to produce water. The remarkable ease with which this reaction proceeds prompted Berzelius in 1836 to first use the term "catalysis". Despite apparent simplicity, its mechanism has remained the subject of much debate. We have performed an extensive study of this original catalytic process. Two distinct temperature dependent mechanisms of H₂O formation have been identified. At high temperatures H₂O is formed by the simple and intuitively attractive addition of H to O and then to OH:

At low temperatures a proton transfer reaction (disproportionation reaction) from H₂O to O produces hydroxyls and ultimately H₂O much more readily than the reaction of O and H:

This dual path mechanism to H₂O formation is an important finding, one that nicely explains and rationalises many seemingly contradictory experimental observations.

A further significant discovery is the role played by H₂O in catalysis. In the low temperature mechanism one of the H₂O molecules dramatically improves the efficacy of the disproportionation process. Although not directly involved in the reaction this H₂O molecule, through hydrogen-bonding, lowers the transition and final state energies and acts as an extremely effective thermodynamic and kinetic promoter. Because of the ubiquitous presence of H₂O molecules in industrial catalytic systems this effect should be considered in the future design of industrial catalytic systems.

A new framework
Finally, perhaps the most exciting example of my research is a systematic study into the microscopic reaction pathways of catalytic reactions. The basic principles which govern the pathways of most chemical reactions, especially those of relevance to heterogeneous catalysis, are not yet clear. Our objective was the development of a generalised framework. By careful examination of twelve catalytic reaction pathways of our own determination, and others from the literature, we were able to make important initial steps towards this goal. Crucially, we discovered that the valency of the reactants involved is the principal determining factor of a catalytic reaction pathway. Reactants of the same valency access similar transition states and the location of transition states is determined by the valency of the reactants involved. To provide a simple example CH₄, NH₃ and H₂O, all saturated molecules, dissociate preferentially over a single metal atom on a Pt surface. On this basis, we developed rules to facilitate the prediction of catalytic reaction pathways without recourse to experimental or theoretical investigations.

Conclusion
The application of modern quantum mechanical techniques to the study of surface catalysed reactions has been the focus of my work. Studies such as those outlined reveal how quantum mechanics can be used to accurately illustrate important issues in heterogeneous catalysis. Theory facilitates our better understanding of the most intimate details of catalytic reactions and assists in the translation of this data into a conceptual form that fellow chemists can use to design new processes .

> Back to Prize index page