GOVERNMENT

by Karla Harby for NCSA -- Computational methods are helping chemists unravel the complexities of inorganic catalysts with possible implications for ozone depletion and the automotive industry. CHAMPAIGN, IL--Chabazite is a rare mineral with a beauty that makes it a rock collectors' favorite. Creamy colored or pinkish with a vitreous sheen, it is pressed into nearly cubic crystals when volcanic rocks are subjected to the heat and pressures of metamorphosis. Although chabazite looks impressively solid sitting on a coffee table, on the atomic level it has properties possessed by all zeolites -- it's full of holes. Zeolites, a class of natural and synthetic minerals possessing large, vacant spaces in their crystalline structure, provide tunnels and hideouts for such atoms as potassium, sodium, and calcium and even for molecules like water and ammonia. This property makes zeolites useful for things such as water purification and softening, removing radioactivity from spent nuclear fuel, and controlling kitty litter odor. The cavities in a zeolite's structure also help explain its effectiveness as a catalyst. A catalyst is a substance, in this case an acidic solid, that facilitates chemical reactions without itself being changed. Not much is currently known about these zeolite caves on the atomic level, including how they interact with the atoms that enter their passages. Successful spelunking here could have ramifications far beyond the contentment of the household cat. "Zeolites are rather large, complex systems. They're analogous to enzymes," says David White, a chemistry professor at the University of Pennsylvania. "The big question is: What kind of chemistry goes on in small cavities -- just like what you have in cells?" Because cells also contain complex networks of cavities, a better understanding of the chemistry of zeolites could reveal much about the chemistry of biological systems. Bernhardt Trout, professor of chemical engineering at the Massachusetts Institute of Technology, White, and other collaborators are using the Alliance's SGI Origin2000 supercomputer at NCSA to simulate catalysis computationally. They chose chabazite as their archetype zeolite because it's used in industry to convert methanol to olefins. Despite having rather complex zeolitic systems, chabazite is small enough (only 36 atoms) to be successfully modeled by supercomputers, yet big enough to yield important information. Trout is also using the Origin2000 to study both catalytic behavior relevant to the automotive catalysts used in catalytic converters and the chemistry of ice crystals. Understanding how ice in the stratosphere reacts with other compounds could offer insights into ozone depletion and how to prevent it. Trout came to Alliance resources with specific questions in mind. In less than a year he had already obtained a key -- albeit still preliminary -- finding. Chabazite has four different oxygen sites and, consequently, four different acid sites available for absorbing other molecules. Scientists previously thought these various acid sites had substantially different energetics -- in other words, that the energy required for chemical reactions varied. This assumption suggested that the four sites could differ greatly in their chemical activity. But simulations on the Origin2000 have so far revealed quite the opposite. Among the four sites, the energetics for catalytic activity is the same, Trout says. "[Zeolite chemistry] can be very fascinating," says White, whose research group designed experiments to check the validity of Trout's computer simulations with chabazite. "The big thing today is the study of the dynamical properties of large macromolecules. If the supercomputer is important to us, it's because it can be predictive." He adds: "Much of the calculation in chemistry is not really predictive, so experiments are done to get a model. But physicists can do that -- they predict when the earth is going to end, and we sit here, waiting. Chemists have not been able to do this, because chemical systems are, in a way, much more complex." Although Trout's work has focused on the chemistry of zeolites, he's quatitatively studying other important catalysts as well. In a joint project with the Ford Motor Company, Trout has been using computational catalysis to see whether sulfur resistant catalysts in cars' catalytic converters can be developed for lean-running engines. A lean engine runs with a high air-to-fuel ratio, or an excess of oxygen (O2). "Lean in general is a good thing because you get more efficiency," Trout says. "If you can get more efficiency, you decrease emissions in addition to saving money on fuel." A new and potentially promising, catalyst for lean-running engines is composed of platinum and barium oxide (BaO). Unfortunately, a BaO catalyst cannot be used today because, under lean conditions, the sulfur present in gasoline oxidizes on the catalyst's platinum to produce sulfur trioxide (SO3). This byproduct then poisons the BaO, rendering the whole catalytic material useless for absorbing the nitrogen oxide (NOX) emissions in air pollution. But, there might be a way to design catalysts that resist damage from sulfur. "We're on our way to computing about 100 configurations" for automobile catalysts, Trout says. "We have the energetics, and we're working on kinetics of the SO2 oxidation reaction." In collaboration with Mario J. Molina, an Institute Professor also at MIT, Trout is using Alliance computational power to explore how chemical reactions occur on the ice crystals of stratospheric clouds. That the subject is under study at all is noteworthy. Until the 1980s, chemists thought that ice was chemically inert -- that when dilute solutions froze, pure ice crystals grew and the impurities remained largely in the liquid phase. But work by Molina and others suggested that ice has a layer of disorder on its surface that engages in chemical reactions. "I stay away from the word 'melting,'" Trout comments about this layer because melting implies the conversion from a solid to a liquid. "This is just the interfacial region. In an ice cube in water, you see the ice, you see the water. There's a region between the two that looks rigid to the eye, but it isn't so rigid." While everyone now agrees that ice -- including the ice crystals in clouds -- participates in chemical reactions that contribute to ozone depletion, the mechanisms involved remain controversial. Trout speculates that one particularly active site on ice may explain activity without resorting to the disordered-region hypothesis. In other words, focusing on disorder on the surface of ice may not be needed; a perfectly ordered region may explain this chemistry. If confirmed, that would be a fascinating new finding. "From what we've been working on, we don't have something to solve the ozone [depletion] problem, but then no one else does either, so we try not to feel too bad about that," Trout allows. But, he adds, someday computational catalysis may help us "know better the consequences of what humans produce and make informed on decisions about what we are sending up there [into the ozone layer]." In the meantime, computational catalysis has already proven its power as a new instrument for helping scientists understand a wide variety of important chemical behaviors at the molecular level. "We [theorists] don't work on a thing that can't be validated," Trout explains. But "the level of detail we can get with computers is not possible in the laboratory." Relevant URLs: --Access story: http://access.ncsa.uiuc.edu/Stories/zeolites/ --Trout research group: http://troutgroup.mit.edu/