A design strategy applied to sulfur resistant lean NOx̳ automotive catalysts

Abstract

Catalyst poisoning due to sulfur compounds derived from fuel sulfur presents a major challenge, intractable thus far, to development of many advanced technologies for automotive catalysts such as the lean NOx, trap. Under lean conditions, sulfur will be oxidized to S0₃ and then form sulfate on the trap. The sulfate on the trap is thermodynamically very stable and, thus, difficult to purge. The NOx trap will then be deactivated over time. Our objective has been to build up a framework for the design of selective, sulfur resistant, oxidation automotive catalysts, which are active for the oxidation of NO to NO₂ but relatively inactive for the oxidation of SO₂ to S0₃. It is well known that the catalytic properties of alloys are often superior to those of pure metals, because of either the electronic effect or the ensemble effect or both. The ensemble effect is due to a change in distribution and availability of surface reaction sites, while the electronic effect is due to a change in electronic structure, leading to a change in rate constants of elementary steps. However, a very large number of possible compositions of alloys exist for any particular application.

(cont.) Therefore, a fundamental understanding of the relationship between the electronic structure, the composition, and the activity of alloys, which could aid in catalyst design, is first developed. This is accomplished by constructing a generalized weighted d-band center model for the prediction of the binding strength of chemisorbed molecules, in which the various atoms in the molecules bind unequally to multiple types of surface atoms. This model is then applied to estimate the adsorption energies of SO₂ and NO at both initial states and transition states on various surfaces. Both energetic data and electronic structure data are obtained from first principles density functional theory calculations. Our model is found to predict well the relative stability of adsorbates on surfaces and can be used to predict the effects of different compositions on the energy of adsorption. A strong linear correlation is found between our new weighting of the positions of the d band of the surface and the molecular adsorption energy. These linear relationships are then used together with energy decomposition scheme for a coadsorbed system on surfaces, to predict the reactivity of SO₂ and NO oxidation on different surfaces.

(cont.) A catalyst which is selective for the oxidation of NO over SO₂ is then developed. This study should aid in the development of more effective catalysts for an extremely important environmental application.