Through it's ability to be compiled into the ChemShell QM/MM modelling environment, GAMESS-UK has been used to model the Quantum Mechanical region in the research described below.

Understanding the Chemistry of Oxide Surfaces[1]

For improvements to be made in long standing industrial catalytic processes an understanding of the atomistic mechanism of the reactions is required. A variety of experimental techniques can be used to study sorption and reaction processes, but when both catalyst and reactant mixtures are multicomponent, mechanisms are particularly complex and often controversial. Computational techniques can be used to gain valuable insight and interpret experimental evidence.

Industrial Methanol Synthesis

A large quantity of methanol, in excess of 22 million tonnes worldwide, is produced annually using the multicomponent Cu/ZnO/Al2O3 catalyst and feed gas, CO2/CO/H2. Many experimental studies of this process have been performed but without any definite reaction mechanism being established. The rate-determining step is thought to be the hydrogenation of adsorbed intermediates, for example the formate ion, at the active sites.

Active Site

Proposed mechanisms for methanol synthesis require the chemisorption of CO2 before hydrogenation via formate to methanol. The nature of the active site for sorption/catalysis of CO2 remains unclear; as a test system (model catalyst) it has been proposed to use clean oxygen terminated surfaces of zincite. We have concentrated on the polar (000-1) oxygen terminated surface that is stabilised by vacant oxygen interstitial surface sites. The presence of such sites has been confirmed by experiment, they have been suggested as the active catalytic sites for methanol synthesis. Using QM/MM embedding techniques, we have undertaken calculations on the reactants, intermediates and products of methanol synthesis.

Catalytic Cycle

The proposed catalytic cycle is summarised in Figure 1. We start with the adsorption of CO2 and H2. CO2 upon adsorption retains its linear structure. Upon adding an electron the molecule bends and the extra electron populates an antibonding level. The interaction with the surface stabilises the radical CO2 - species. The reaction then proceeds via the hydrogenation of the adsorbed CO2 -, by surface hydrogen, to the formate ion. Further hydrogenation can proceed either through the formation of H2CO2 - or HCOOH- (formic acid) as shown in the figure. Experiment does not detect these species, which suggests a short lifetime and therefore high reactivity. Computational techniques allow us to differentiate and investigate these different scenarios. Further hydrogenation and interactions of the resulting species with the surface and possible surface defects lead to a large variety of possible intermediates. We show examples of a methoxy ion (CH3O-) chemisorbed to the surface and physisorbed methanol. To complete the catalytic cycle, methanol is removed from the surface and the active site is recycled by desorption of carbon dioxide and water.


figure 1
Figure 1: Proposed catalytic cycle for Methanol synthesis.

Adsorbed Copper

We have investigated the interaction of copper clusters with the surface of zincite, as the active site of the industrial catalyst is known to involve copper. Although such catalysts have been used for many years the role of copper is still largely unknown. The following unresolved questions are of interest:

  1. the nature of the interaction of Cu with the ZnO surface
  2. the electronic state of the active copper sites (0, +1, +2)
  3. the surface morphology of the catalytically active Cu clusters (111), (110) or (110)-like.

We have attempted to address these issues by investigating the adsorption of a single Cu ion in different charge states. Our calculations show that Cu+ and Cu2+ ions are ideal anchor sites. Larger clusters have then been built up using the single Cu ion as a seed. As a promising example we show the embedding of a Cu4 2+ cluster (Figure 2) anchored on the (0001)-Zn terminated surface of zincite as compared to the gas phase geometry shown in Figure 3.


figure 2figure 3
Figure 2:Cu4 2+ cluster embedded on the (0001)-Zn surface. Figure 3: Cu4 2+ gas phase cluster.

Summary

The application of computational methods to such complex materials and their catalyic activity clearly demonstrates the power of the solidstate embedding scheme available in ChemShell. A paper giving further details of the catalytic cycle has recently been published in Angew. Chem. Int. Ed., 2001, 40, p 4437-4440.

1   The project combines software development work of three academic groups active in the area, CLRC Daresbury (UK), the group of Prof. Walter Thiel at the Max Planck Institute für Kohlenforschung, Mülheim (DE), and the group of Prof C.R.A. Catlow at the Royal Institution (UK) with demonstration and applications work from modelling teams within three major European chemical industries, Norsk Hydro(NO), BASF (DE) and ICI (UK).

Applications - Main Page