Gold catalysts for pure hydrogen production in the water–gas shift reaction: activity, structure and reaction mechanism

Abstract

The production of hydrogen containing very low levels of carbon monoxide for use in polymer electrolyte fuel cells requires the development of catalysts that show very high activity at low temperatures where the equilibrium for the removal of carbon monoxide using the water-gas shift reaction is favourable. It has been claimed that oxide-supported gold catalysts have the required high activity but there is considerable uncertainty in the literature about the feasibility of using these catalysts under real conditions. By comparing the activity of gold catalysts with that of platinum catalysts it is shown that well-prepared gold catalysts are significantly more active than the corresponding platinum catalysts. However, the method of preparation and pre-treatment of the gold catalysts is critical and activity variations of several orders of magnitude can be observed depending on the methods chosen. It is shown that an intimate contact between gold and the oxide support is important and any preparative procedure that does not generate such an interaction, or any subsequent treatment that can destroy such an interaction, may result in catalysts with low activity. The oxidation state and structure of active gold catalysts for the water-gas shift reaction is shown to comprise gold primarily in a zerovalent metallic state but in intimate contact with the support. This close contact between small metallic gold particles and the support may result in the "atoms" at the point of contact having a net charge (most probably cationic) but the high activity is associated with the presence of metallic gold. Both in situ XPS and XANES appear unequivocal on this point and this conclusion is consistent with similar measurements on gold catalysts even when used for CO oxidation. In situ EXAFS measurements under water gas shift conditions show that the active form of gold is a small gold cluster in intimate contact with the oxide support. The importance of the gold/oxide interface is indicated but the possible role of special sites (e.g., edge sites) on the gold clusters cannot be excluded. These may be important for CO oxidation but the fact that water has to be activated in the water gas shift reaction may point towards a more dominant role for the interfacial sites. The mechanism of the water gas shift reaction on gold and other low temperature catalysts has been widely investigated but little agreement exists. However, it is shown that a single "universal" model is consistent with much of the experimental literature. In this, it is proposed that the dominant surface intermediate is a function of reaction conditions. For example, as the temperature is increased the dominant species changes from a carbonate or carboxylate species, to a formate species and eventually at high temperatures to a mechanism that is characteristic of a redox process. Similar changes in the dominant intermediate are observed with changes in the gas composition. Overall, it is shown that reported variations in the kinetics, structure and reaction mechanism for the water gas shift reaction on gold catalysts can now be understood and rationalised.