Evaluating vehicle emissions controls

A major stumbling block to putting 80-mile-per-gallon cars on the road within this decade is the lack of effective emissions controls for lean-burn engines. Lean-burn engines, whether diesel- or gasoline-fueled, are designed to carry out combustion with an excess of air. Such combustion achieves increased energy efficiency and reduced emissions of greenhouse gases, such as carbon dioxide. If lean-burn engines could be used for passenger cars, the Department of Energy estimates that fuel economy increases of over 30% could be readily achieved. Such improved efficiencies would be quite a step forward in the move to reduce U.S. dependency on foreign oil. However, in spite of higher efficiency, lean-burn emissions continue to be a problem because cleanup technologies are not available for lean exhaust. Conventional catalytic converters are unable to simultaneously reduce the nitrogen oxides, carbon monoxide, hydrocarbons, and particulates from lean-burn engines to required levels. Thus, the development of new exhaust cleanup technologies is critical.

ORNL's Bill Shelton and Stuart Daw are developing supercomputer software for simulating the physics and chemistry of lean exhaust cleanup devices. Most of the promising cleanup technologies involve complex chemical reactions between the gaseous exhaust species and special solid-phase catalytic materials coated on ceramic substrates. Up to now, the level of complexity involved has restricted the development of new cleanup systems to the construction and testing of experimental prototypes. This empirical approach proved adequate in previous decades for developing the automotive catalysts used today, but it is simply too slow and costly for current needs.

In focusing on detailed simulations of the underlying physical processes of cleanup devices, Shelton and Daw are joining a new generation of researchers who are trying to apply the power of high-performance computing to go beyond empiricism. Specific goals of the ORNL researchers include the simultaneous description of atomic-scale interactions on the surface with models for heat and mass transport between the surface and gas. By accurately modeling the dynamics of cleanup devices at multiple scales, it is expected that development of new technology can be greatly accelerated.

"Simulations based on fundamental physics and chemistry can reveal previously unanticipated approaches for formulating the catalytic materials, and better ways to link them with the engine exhaust can be identified and exploited," Shelton says. "Realistically, some degree of empiricism will always be necessary, but even then accurate simulations can be used to more effectively plan and interpret experiments."

ORNL is emerging as a leader in this field because of its experience in the experimental evaluation of emissions control devices, its considerable expertise in fundamental surface physics and chemistry, and its world-class facilities for high-performance computing. "Through workshops and direct collaborations, we have been bringing together a broad range of experts from national labs, universities, auto makers, emissions control manufacturers, and engine companies who are extremely interested in addressing the problems of lean-exhaust simulation," says Daw. "We hope to help DOE's Office of Transportation Technologies set priorities for research and construct a coordinated approach to overcoming this hurdle in developing a clean, efficient car."

"By combining models for the dominant physical processes at multiple scales, we are obtaining previously unavailable insights into the coupling of local surface transformations and chemical reactions with global heat and mass flow through the devices," Shelton adds. "This approach could lead to more innovative solutions for improving performance of emissions controls. For example, the multi-scale approach is crucial to understanding the durability of the catalytic material—that is, how long the catalyst will function before it must be replaced. Our goal is to understand how global heat flow in the gas and ceramic substrate produces coarsening of catalyst nanoparticles. This information would be used to determine options for reducing coarsening to delay degradation in catalyst performance."

Shelton and Daw emphasize that their long-term objective is to produce simulations of lean-exhaust cleanup that are directly relevant to realistic driving conditions. The availability of simulations at this level will allow industry and DOE to calculate cost-benefit ratios for different lean-exhaust-cleanup technologies so they can make more informed decisions.