Kinetic measurements on Lean NOx Traps

Shadab Sharif Mulla, Purdue University

Abstract

The Lean NOx Trap (LNT) process is one of the technologies in development to abate the oxides of nitrogen (NOx, primarily NO and NO2) in the exhaust emitted from lean-burn combustion engines (e.g., diesel engines). The LNT catalyst consists of a noble metal (usually Pt) and a NOx storage material (such as Ba or K), on an alumina support. The process involves a cyclic operation of the engine between long fuel-lean phase ( e.g., 60 s) and relatively short rich pulse (e.g., 5 s). The entire cycle can be divided into numerous steps with the following three being the most important: (1) oxidation of NO to NO2 on Pt and (2) NOx storage on Ba under lean conditions; and (3) regeneration of the catalyst reducing the NOx to N2 under rich conditions. This work will focus on the kinetic aspects of the NO oxidation and the NOx reduction steps (steps 1 and 3) of the cycle. The first step, NO oxidation, was studied on various Pt catalysts with varying average Pt particle size (2-9 nm) and support (Al2O 3, SBA-15, Carbon, Nb2O5). The forward reaction was found to be close to first order with respect to the reactants NO and O2, while it was close to negative first order with respect to the product NO2, indicating that NO2 inhibits NO oxidation. These apparent reaction orders were found to be independent of the size of the Pt clusters or the support material. A reaction mechanism consistent with these experimentally observed reaction orders was proposed, and consisted of the following steps: (1) Quasi-equilibriated adsorption of NO on Pt; (2) Quasi-equilibriated dissociative adsorption of NO2 on Pt giving surface NO and O; (3) irreversible molecular adsorption of O2 on Pt. Surface O was proposed to be the most abundant reaction intermediate. The rate per unit of surface Pt atom was found to be a function of both, the Pt particle size and the support. The rate was found to increase by about 20-50 times, depending on the support, when the average Pt cluster size increased from 1.9 to 9.1 nm. Based on in-situ XAS, XPS and TEM evidences, it is hypothesized that only those Pt particles capable of resisting Pt oxidation, which is the cause of the catalyst deactivation under the oxidizing reaction conditions, are responsible for the reaction. Pt oxidation, in turn, depends on the Pt particle size and the support material, and results in a higher NO oxidation rate on those catalyst with lower extent of Pt oxidation during the reaction. The final step, which is the regeneration of the trap catalyst under fuel rich conditions, is proposed to involve a localized reaction front of the reductant H2, which travels through the catalyst bed and gets completely consumed while regenerating the trapping sites. For typical 60 s lean-6 s rich cycles, the N2 selectivity was found to be 96%. The process involves the release of NOx from the trapping material, followed by its reduction over Pt. The regeneration reactions were found to be fast and insensitive to the changes in the temperature or to the presence of CO 2 and H2O, and were limited only by the supply of the H-atoms. The catalyst does not differentiate between H2 or NH3 as the source of H-atoms, and NH3 is equivalent to and as effective as H2 in regenerating the trap catalyst, giving similar concentration profiles and product selectivity. The results also indicate that the catalyst regeneration using H2 as the reductant occurs via intermediate NH3 formation, which serves as a carrier of the hydrogen atoms. This NH3 eventually gets further oxidized by the stored NOx to N2, thereby restricting the NH3 slip to only the end of the cycle where the stored NOx are depleted, while maintaining the high selectivity of the process towards N2.

Degree

Ph.D.

Advisors

Delgass, Purdue University.

Subject Area

Chemical engineering

COinS