Experimental and Modeling Study on the Thermal Aging Impact on the Performance of the Natural Gas Three-Way Catalyst

The prediction accuracy of a three-way catalyst (TWC) model is highly associated with the ability of the model to incorporate the reaction kinetics of the emission process as a lambda function. In this study, we investigated the O₂ and H₂ concentration profiles of TWC reactions and used them as critical inputs for the development of a global TWC model. We presented the experimental data and global kinetic model showing the impact of thermal degradation on the performance of the TWC. The performance metrics investigated in this study included CH₄, NOx, and CO conversions under lean, rich, and dithering light-off conditions to determine the kinetics of oxidation reactions and reduction/reforming/water-gas shift reactions as a function of thermal aging. The O₂ and H₂ concentrations were measured using mass spectrometry to track the change in the oxidation state of the catalyst and to determine the mechanism of the reactions under these light-off conditions. The experimental data indicate that the NO_x and CH₄ conversions were higher under rich lambda conditions, thereby generating more NH₃ than that observed under lean lambda conditions. Conversely, the NH₃ formation was mitigated under the dithering conditions resulting from the recovery of redox properties. The measured O₂ and H₂ concentration profiles indicated that the conversion of CH₄ was attributed to its reaction with O₂ until O₂ was fully consumed. Further, it was converted through a reforming reaction that produced H₂ when O₂ was depleted. Consequently, dithering conditions with a substantial amount of O₂ showed a delayed onset of reforming chemistry and NH₃ formation than those observed under rich conditions. The global kinetic model was developed based on the O₂ and H₂ data obtained under lean and rich conditions as inputs. The model predicted the NOx reduction chemistry in the presence of CO and H₂, steam reforming, and the total oxidation of methane reasonably well. We also discussed an approach to improve the model predictions for the partial oxidation reaction of methane.