AmericanChemicalSociety.com

Magnetite is the active component of high temperature water-gas shift catalysts, and copper is sometimes added to commercial catalysts as a promoter. The manner in which the copper functions has not been definitively established. One hypothesis is that copper cations substitute for iron cations in the oxide, and the associated changes in the electronic structure of the catalyst affect its catalytic behavior. An alternative hypothesis is that under water-gas shift conditions, copper migrates to the surface of the catalyst and forms small metallic copper aggregates that are responsible for the modified catalytic properties. The mechanism of water-gas shift, and consequently its kinetics, might be affected quite differently in these two cases, and a comparison of the kinetics of water-gas shift over an unpromoted catalyst to the kinetics over a copper-promoted catalyst might help discriminate between them.

Cluster models of an active site on the {111} surface of magnetite were created and used to study the geometry and energetics of adsorbed water-gas shift intermediates. In addition to an unpromoted iron oxide cluster, two copper-promoted clusters were used. In one copper-substituted cluster, the copper cation was located in an octahedral site just below the surface, directly below the three surface iron cations representing the active site. This copper cation is fully coordinated to six nearest neighbor oxygen anions. In the other copper-substituted cluster, the copper cation replaced one of the three surface cations corresponding to the active site. Density functional theory calculations were performed using B3LYP correlation and exchange functionals and a TZV** basis set.

Two types mechanisms have been advocated for water-gas shift over iron oxide catalysts. In this work, five relevant surface species (in addition to the vacant site) were studied: molecularly adsorbed water, dissociatively adsorbed water, an oxidized site, a carbonate and a formate. DFT was used to identify minimum-energy geometries for each and to calculate the energy associated with its formation.

All attempts to minimize the energy of molecularly adsorbed water were unsuccessful; it appears that water adsorbs dissociatively on {111} sites. Two geometries of nearly equal energy were found for dissociative adsorption of water. In both cases a hydroxyl group formed from the adsorbing water is located in the three-fold site formed by the surface cations. The difference between the two structures is in the location of the other proton. The structure did not change appreciably when copper was substituted either sub-surface or on the surface. Similarly, adsorbed oxygen, located in the three-fold site formed by the iron cations, is essentially the same for the sub-surface and surface clusters. The surface carbonate is unidentate, and it is located in a bridging position between two of the three surface cations, and again, the structure is essentially the same for all three clusters studied.

The formate that forms on the unpromoted and sub-surface copper promoted site is bidentate with one oxygen atom bridging two iron cations and the other oxygen atom over the remaining cation. The other proton forms a hydroxyl group with a surface oxygen atom. A very different structure was found when copper was present on the surface. The formate remains bidentate, but each of the oxygen atoms binds to a single iron cation and the copper cation does not participate. In addition, the second proton still forms a hydroxyl group with a surface oxygen anion, but it becomes positioned slightly above the plane of the oxygen anions instead of being slightly below that plane.

The energy change associated with forming each surface intermediate from a corresponding gas phase species was computed for each species and cluster. In all cases, the substitution of copper causes a weakening of the bond between the surface and the adsorbed species. When the copper substitution is sub-surface, the energy decreases for every species, by amounts ranging from 1.0 (for the carbonate) to 15.9 kJ/mol (for dissociatively adsorbed water). These changes are of the order of the expected uncertainty in energies calculated using DFT.

The energy changed by a much greater amount (ranging from 57.4, for the carbonate, to 78.6 kJ/mol, for the oxygen adatom) when copper substituted in the surface. This is consistent with the geometric findings. The structural differences in the species between the all-iron cluster and the surface cluster were much greater than those between the all-iron cluster and the sub-surface cluster. In some cases the degree of coordination to the surface decreased and in the other cases the surface bond lengths increased. Both of these effects are consistent with the result that the surface bonds were weaker for the species on the surface-promoted cluster.

The goal of this study was to estimate the magnitude of the changes in the water-gas shift energetics due to copper substitution in order to determine whether detailed kinetic modeling might be useful in discriminating among the models proposed for copper promotion. The present results do suggest that comparing the mechanistic kinetics of a promoted catalyst to an unpromoted catalyst might be useful in this regard. Specifically, if the mechanistic kinetics of water-gas shift are only slightly different on the two catalysts, then the present results would support a promotional model involving substitution of copper within the bulk of the iron oxide. However, if the mechanistic kinetics for promoted and unpromoted catalysts were markedly different, the role of the copper would be less clear. The present results suggest that large differences could be due to copper substitution within the surface of the iron oxide (one model for promotion), but they do not rule out the alternative possibility that large changes could also be due to the formation of metallic copper aggregates. In light of these results, mechanistic kinetic studies are presently being performed.