Low temperature water–gas shift: Applications of a modified SSITKA–DRIFTS method under conditions of H2 co-feeding over metal/ceria and related oxides Original Research Article

Three applications of a modified SSITKA–DRIFTS technique demonstrate the importance of designing catalysts for fuel processor applications in terms of a formate mechanism. Switching between 12CO and 13CO was carried out under steady state water–gas shift conditions with a feed containing co-fed H2. The dynamic responses of surface species and CO2 product were evaluated in terms of their times required to achieve 50% fractional isotopic exchange. In all cases studied, the formate ν(CH) band was observed to switch at a rate similar to the gas phase asymmetric ν(CO2) band. In the first application, the method was employed to monitor and corroborate the normal kinetic isotope effect that has been suggested to link the rate limiting step of water–gas shift to formate decomposition via C–H bond cleaving. Also observed was that when the ν(CO) band for adsorbed CO on Pt was completely exchanged, the ν(CO2) band had only achieved 50% exchange, casting doubt on the proposed mechanism involving the direct reaction of Pt–CO with O adatoms on ceria to produce gas phase CO2. In a second study, the impact of metal loading on the formate switching rate led to a remarkable decrease in the switching times of both formates and CO2 product, emphasizing the need to express the reaction pathway in terms of a bifunctional mechanism involving both the oxide, where formates are formed at Type II bridging OH groups, and the metal, where dehydrogenation of formate proceeds rapidly. The results suggest that when the loading of Pt was greater, a higher fraction of rapidly reacting formate close to the metal–ceria interface was present. In contrast, when the Pt loading was low, a considerable fraction of formate was located further from the metal and required time to diffuse to the metal–oxide interface prior to decomposition. Finally, a demonstration of the usefulness of the technique in determining whether an associative mechanism holds true for other partially reducible oxide components is provided. Oxides included ceria, thoria, zirconia, and a mixed ceria–zirconia oxide sample. Limitations of the modified SSITKA technique regarding experimentation and data interpretation are also considered.