Crystallite-scale model to predict the reactor-scale performance of spillover-exhibiting automotive catalysts
15:00 hrs on Wednesday, 28th October 2020.
Abstract:-
A model incorporating the geometry features, transport, and reactions at the crystallite scale is developed to predict the oxygen storage capacity (OSC) of a Pd-containing three-way catalyst (TWC) during cyclic lean and rich conditions. The reactions on the metal crystallites, metal-ceria interface, and the ceria surface are considered and the coupling between these reactions and the spillover processes is incorporated. The crystallite-scale model is integrated with a monolith reactor model to predict the experimental data of OSC at various temperatures. The dependence of oxygen urface diffusivity on the temperature is used to predict the reported variation of OSC with temperature. It is quantitatively shown that in addition to the active surface area, the interfacial erimeter plays a key role in the spillover and reverse spillover processes. In agreement with the reported trends, the model predicts that the incomplete regeneration of the TWC at low temperatures is responsible for the low values of OSC. The model predicts the effect of Pd crystallite size, Pd-ceria interfacial contact, and Pd dispersion on the spillover and reverse-spillover of oxygen between Pd and ceria, with CO, H2, and CH4 as reductants. Consistent with the experimental findings, the model predicts that H2 results in a higher OSC than CO and CH4. The effective surface diffusivity with H2 is estimated to be higher than that with CO and CH4, which is attributed to the possible spillover of hydrogen in addition to the reverse-spillover of oxygen.
The extent of hydrogen spillover on two different Pd-based catalysts supported on mesoporous SBA-15 and MCM-41 is quantitatively predicted using a crystallite-scale model, which considers the effect of size and location of crystallites, metal-support interfacial contact, and diffusivity of the adsorbed species on the support surface. The developed model, in combination with a fixed-bed reactor model predicts that in the absence of any reductant in the feed, the NO conversion is higher (=37%) on the SBA-15 supported catalyst as compared to 25% on MCM-41, which is consistent with the reported data in the literature. The difference in the catalytic activity is ascribed to the different spillover rates of hydrogen on SBA-15 and MCM-41, which is further correlated to the fraction of Pd crystallites present in the support channels. The model is also used to study the effect of cycle time, H2 concentration, Pd dispersion, and Pd loading on NO reduction under cyclic conditions. Consistent with the literature findings, the model predicts that the Pd surface area and the metal-support interfacial contact increase with an increase in the Pd loading and dispersion, which result in higher spillover and hence increase the NO conversion. The fundamental insights obtained in the present work are expected to be applicable to other catalytic systems which involve the spillover effects. One such catalyst is a lean NOx trap which involves spillover of NOx and/or hydrogen whereas other applications include TWCs for gasoline as the fuel.