Researchers at DOE's National Energy Technology Laboratory and Carnegie Mellon University (CMU) of Pittsburgh, Pa., have longstanding experience in conducting research and applying computational modeling to energy-related problems. They are collaborating to combine computational and experimental methods to test a variety of hydrogen separation membranes that seem to have the potential to remove hydrogen from mixed gas streams and simultaneously remove impurities from the hydrogen product. Although hydrogen is very abundant, producing it from fossil fuels, water, or renewable energy sources is very costly. With a goal of reducing the cost of hydrogen production, researchers at NETL and CMU are studying hydrogen gas separation.

Selecting the most cost-effective membranes to separate hydrogen from mixed gas streams through laboratory experimentation is both time-consuming and expensive. NETL and CMU researchers have been collaborating to develop better methods to determine the optimum composition of membranes. Coupling the predictive capability of computers with laboratory experiments, the researchers were able to develop a computational method for screening alloys using theoretical models.

By screening candidate alloys with special computer programs to determine the ones with the best separation potential, researchers save time and money by avoiding the need to create a multitude of alloys and then test their properties in a laboratory setting. Using computational chemistry, only the most promising alloys would be subjected to experimentation.

NETL researchers have demonstrated recent successes with a new hydrogen membrane material made of palladium and copper (Pd-Cu) alloy. Their results to date have shown the ability of the membrane—resembling a flat disc about three-fourths inch in diameter*to allow pure hydrogen to pass through without contamination of the membrane by other gas impurities (such as hydrogen sulfide) during the separation process.
The team's research, supported by DOE's Office of Fossil Energy, appeared in the January 28, 2005, issue of the prestigious journal Science published by the American Association for the Advancement of Science.
 

~