Active site models for high temperature water gas shift reaction over iron oxide catalysts
VanNatter, Rainee M.
MetadataShow full item record
Quantum chemical calculations and microkinetic modeling were carried out in order to study the high temperature water gas shift reaction over iron oxide catalysts. Cluster models corresponding to potential active sites on (100), (110), and (111) surfaces of an Fe 3 O 4 crystal have been generated, and the mechanistic thermochemistry of the water gas shift reaction was studied using DFT computations on these model surfaces. At the same time, microkinetic modeling has been carried out to study various proposed mechanisms for the water gas shift reaction. These 2 sets of data were then compared to discriminate between potential reaction mechanisms. The fitting of experimental data to a simple 2-step redox mechanism yielded a heat of localization for an oxygen adatom of −605 kJ/mol, and predicted virtually full oxygen coverage on the catalyst surface. The model provided a good fit to experimental data, and the predicted entropy loss for the adsorbing species were within the expected range. The oxygen localization enthalpy fell within the range of SCF energy values obtained from the DFT calculations on the (111) clusters. Cluster DFT models of the sizes used were shown to have significant limitations, and thus only qualitative comparisons could be made. More detailed redox mechanisms provided a better fit to the kinetic data. They also predicted a mainly O* covered surface and an oxygen atom adsorption energy on the order of −600 kJ/mol. Fitted localization enthalpies for the redox mechanism in which surface water dissociates using a cluster oxygen were in qualitative agreement with a (110)-(111) mixed surface. Fitted localization enthalpies for the redox mechanism in which surface water dissociates using a neighboring adsorbed oxygen were not consistent with the current surfaces studied. The formic acid & formate kinetic models both predicted a nearly full catalyst coverage of CO 2 *, and a CO 2 * bonded too strongly to the surface compared to the DFT cluster models. They were a poorer fit to the kinetic data than all of the redox mechanisms studied. Preliminary studies with copper promoted catalyst clusters in which the Cu 2+ was substituted below, rather than in, the active site showed qualitative agreement with available literature.