In each case, scientists have devised materials that attract and suck up the target molecules. Known as adsorbents, they are usually porous -- meaning they soak up materials like a sponge wipes up water -- and they are specific for certain molecules. They also tend to be less expensive than other separation techniques, such as distillation, used in industry. Although adsorbents are widely used, scientists have much to learn about how they work and how to control the size and shape of the microscopic pores.
"Synthetic chemists now are able to control the manufacture of these porous materials much better than before," says Keith Gubbins, professor of chemical engineering at Cornell University, "but this is still largely an art, and so far most of the materials are imperfect. They are chosen empirically from what is available and what seems to work best in the laboratory, but they're not perfectly crystalline and regular. Usually, there are large pores or spaces, in addition to the small ones of interest, which reduces the selectivity of separation."
Gubbins and his colleagues -- at Cornell and other universities -- are tackling different aspects of the adsorption problem at the Pittsburgh Supercomputing Center, using both the Connection Machine and the CRAY C90. "There are zillions of combinations of possible materials, pore sizes and pore shapes, which would be impossible to test in the laboratory," Gubbins says, "so it is extremely unlikely that these existing materials are anywhere near optimal. Our calculations show that if you tune the pore size and shape, you can achieve much better separations, often by factors of 1,000 or more. If this is possible, the goal of the experimentalists will be to make the material and test it in the laboratory."
Gubbins and Michael Maddox are simulating adsorption in buckytubes, hollow tubes of carbon atoms related structurally to buckyballs. In a related project, Gubbins and Shaoyi Jiang simulate methane adsorption in parallel layers of carbon atoms. This approach shows promise as a way to store natural gas at low pressure, with greater safety and less cost than pressurized gas cylinders.
Researcher: Keith E. Gubbins, Cornell University.
Hardware: CRAY C90; Connection Machine CM2
Software: User developed code
Keywords: adsorbents, adsorption, pore, pore size, pore shape, buckyball, buckytube, separations, methane, graphite, carbon, structure, interaction, new materials, micorporous material, synthetic chemistry, gas pressure, surface, storage.
Related Material on the Web:
Projects in Scientific Computing, PSC's annual research report.
References, Acknowledgements & Credits