Global kinetic model for the regeneration of NO x storage catalyst with CO, H 2 and C 3H 6 in the presence of CO 2 and H 2O

Abstract

Heterogeneous 1D model with global kinetics is proposed for industrial NO x storage and reduction catalyst (NSRC) on the basis of lab experiments in a mini-reactor in the temperature range 100–500 ° C. The NO x reduction dynamics and selectivity towards N 2 or NH 3 are modelled for three main reducing components present in the rich exhaust gas: CO, H 2, and unburned hydrocarbons (HCs). The following reactions are considered: CO, H 2 and HC oxidation, NO x reduction, NO/NO 2 transformation, NO and NO 2 storage, oxygen storage effects, water gas shift and steam reforming, and reduction of the stored NO x by H 2, CO and HC. Ammonia is formed mainly by the reaction of H 2 with NO x , but also by the water-assisted reaction of CO with NO x (formation and consequent hydrolysis of isocyanates). Inclusion of the latter route is necessary to explain the NH 3 formation in CO-rich mixtures without H 2 at lower temperatures. At higher temperatures, water gas shift and steam reforming reactions enable in situ H 2 production from CO, HC and H 2O. The formed ammonia subsequently reacts with oxygen and NO x deposited on the catalyst surface downstream, which results in NH 3 wave travelling along the catalytic monolith. The highest NH 3 yield is obtained around 200 ° C, when the NH 3 formation from the accumulated NO x is already ignited, while the ammonia consumption reactions are still relatively slow. At lower temperatures CO inhibits the NO x reduction by H 2 (i.e., the NH 3 production). At higher temperatures the ammonia oxidation reactions become fast enough to eliminate most of the NH 3 produced locally from the stored NO x . However, ammonia can be still observed after the completion of the regeneration, when it is formed steadily from the rich inlet gas containing NO x . Model results are confronted with the lab data, and calculated evolution of concentration profiles inside the monolith is discussed. The developed model is validated by engine test driving cycle data.