Date of Award


Document Type

Open Access Dissertation


Chemical Engineering

First Advisor

Michael D. Amiridis


Lean-burn gasoline engines are approximately 10% more fuel efficient than conventional, stoichiometric-burn ones. Although relatively modest, if implemented across the entire U.S. automotive fleet, this improvement in fuel economy could have far-reaching implications on the amount of gasoline imported in the country on an annual basis. However, the development of a cost-effective catalytic converter catalyst capable of meeting emission regulations for lean-burn vehicles still represents a major technical challenge. Currently, lean NOX trap (LNT) and selective catalytic reduction (SCR) catalysts are used for this purpose, but both systems suffer from significant drawbacks. For example, LNT catalysts generally require high platinum group metal (PGM) loadings and are highly susceptible to sulfur poisoning. SCR catalysts require a costly urea-dosing system for delivery of urea as the reducing agent into the exhaust stream, as well as a secondary "fuel" tank for on-board storage of urea.

LNT catalysts are typically favored for smaller gasoline engines and are designed for periodic operation in lean and rich environments. NOX is stored on the LNT system during a longer (e.g., 60 - 120 s) lean period and rapidly reduced during a much shorter (e.g., 1- 5 s) rich period. The mechanism for NOX storage is fairly well understood, but the NOX reduction mechanism is still the subject of considerable debate. Lean/rich cycling monitored by in situ Fourier Transform infrared spectroscopy (FTIR) confirmed the presence of surface isocyanate (NCO) species during reduction. Quantification of the FTIR results confirmed that surface NCO species could account for as much as 30% of the N2 formed during the rich period. Hydrolysis of the NCO species to NH3 in the presence of water vapor could also play a significant role. The effect of the lean/rich cycle timing on NH3 formation over a commercial LNT catalyst was also considered. At low temperatures, both the release and reduction of stored NOX was kinetically limited and longer rich periods favored increased NOX conversion and NH3 formation. At elevated temperatures, the opposite was true and shorter rich periods favored increased NOX conversion and NH3 formation. The effects of cycle timing were most pronounced in the 250 - 400 °C temperature range, where optimization of the cycle timing could potentially decrease the PGM requirements of the LNT, especially in a coupled LNT-SCR system.

SCR catalysts are typically favored for heavy-duty applications, but General Motors (GM) recently developed a urea-less, passive-NH3, three way catalyst SCR approach (TWC-SCR) for lean-gasoline vehicles. This TWC-SCR approach also relies on lean/rich cycling, but in this case NH3 is intentionally formed over a TWC during rich periods and stored on a downstream SCR catalyst. The stored NH3 is then used during a subsequent lean period to reduce lean-NOX. NH3 generation over TWCs under steady and cycling conditions was investigated. The temperature, catalytic formulation and reductant concentration all affected NH3 formation. Storage of NH3 on the downstream SCR catalyst was also considered. At low temperatures, the selective reduction of NOX by stored NH3 was favored over a Cu-zeolite SCR catalyst. Above 350 °C, NH3 oxidation was favored over NOX reduction. Recent bench reactor screening using a two-reactor, bench-core reactor configuration demonstrated the viability of the TWC-SCR configuration and NOX conversions exceeding 98% were measured.