(575az) Experimental and Modeling Studies of NOx Storage and Reduction for a Diesel Exhaust In the Presence of H2O and CO2

Storage and regeneration kinetics on a Pt/BaO/Al2O3 catalyst for a lean NOx trap are determined through a combination of extensive experiments and detailed modeling studies. The experimental conditions are maintained to simulate a diesel engine exhaust with O2 as high as 3% in the exhaust during rich conditions. H2O and CO2 are used in the reactor inlet since they have been found to alter the storage as well as the regeneration behavior of the catalyst. Also, their addition makes the experimental conditions closer to that of actual diesel engine exhaust. Various reductants including H2, CO, NH3 and hydrocarbons are compared for their effectiveness in NOx reduction during the regeneration phase. It is found that the relative activity of the reductants depends on the temperature and inlet concentrations. Regeneration experiments are done at various temperatures and concentrations to evaluate their effect on NOx conversion and product selectivities. The product selectivities are found to depend on the relative concentration of the reductant and the NOx in the exhaust and that stored on the catalyst.

We use a low dimensional reactor model to find reaction kinetics for NOx storage on barium and the reduction of stored NOx by various reductants. Low dimensional modeling is used since it includes the effect of diffusional limitations in the washcoat, which is neglected by most of the two-phase integral reactor models in the literature. Also, it has the advantage of a lower computation cost and much shorter computation time. Given the extremely complex nature of the NOx storage and reduction process, its modeling requires a detailed understanding of the surface reactions. But, this would involve a large number of microkinetic reactions, the kinetic parameters of which may be difficult to estimate. On the other hand, global reaction kinetics may be too simple to explain the experimental results. Hence, a combination of the necessary surface reactions and the relevant global reactions are used to model the experimental data. The obtained kinetics are used to perform transient simulations for the NOx trap and the results are compared to the experiments. Simulations have been used to study the effect of parameters such as lean cycle time, rich cycle time, length of the monolith, thickness of the washcoat etc. on the performance of the cycle averaged reactant conversions and product selectivities. Future work involves experimental and modeling studies of a DOC and NOx trap connected in series, wherein the outlet from DOC acts as a feed to the NOx trap.