Something in the water

01.06.09 -

Researchers at the University of Puerto Rico use NCSA resources to engineer methods of removing drugs from drinking water supplies—and to educate the next generation of engineers.

For many of us the findings of a March 2008 Associated Press investigation were startling. The article included a litany of pharmaceuticals that were in American's drinking water, the side-effect a nation taking their medicine, absorbing most but not all of those medicines, and then letting nature take its course the next time they hit the restroom. Traces of 56 drugs or byproducts showed up in Philadelphia's water. Anti-epileptic and anti-anxiety medications made their way into the supply used by 18.5 million southern Californians.

Arturo Hernandez-Maldonado isn't among the many of us. An associate professor of chemical engineering at the University of Puerto Rico in Mayaguez, Hernandez-Maldonado has been carefully studying how to remove drugs like these from water supplies for the last few years. He uses NCSA's machines for the computational part of that effort, and NCSA staffers help him parallelize his computing code so that it can run on multiple processors on the cluster.

Pharmaceuticals have traditionally not been targeted in the water treatment process, which cleans wastewater before it is returned to sources that might be used for drinking water.

Drugs appear in vanishingly small quantities—a few parts per every million or even billion parts of water. Further, contemporary methods for removing these drugs are unselective. If implemented they would pull out things that don't necessarily need to be, making them inefficient. And, for the most part, the drugs aren't currently considered a threat. They don't make us sick immediately like common water-treatment targets, such as an E. coli or other bacteria, might.

Experts debate what impact these drugs might have on humans, or whether they have any impact at all. But it seems more and more experts are taking the matter seriously. "We have to figure out how to remove these drugs that are already present at very, very low levels," says Hernandez-Maldonado. "No one knows what the effect on humans will be."

According to the Associated Press, drug companies are acknowledging the issue as well. Mary Buzby, director of environmental technology for Merck, was quoted in the March article, saying: "[T]here is genuine concern that these compounds, in the small concentrations that they're at, could be causing impacts to human health or to aquatic organisms."

Attacking specific compounds

Hernandez-Maldonado and his research team simulate ways of attaching metals to a silica-rich sorbent material that could be used to filter potable water by removing various compounds like pharmaceuticals. "Current water processing arrangements don't work for pharmaceuticals. We're trying to see if we can get a material that can be used in existing processes as an add-on," he says.

Using density functional theory calculations, the team estimates the interaction energy between a typical pharmaceutical drug and materials functionalized with a metal. The higher the energy, the more readily the drugs are adsorbed from the water and onto the material. These simulations also show them which path occurs most readily and thus which of these custom materials can be most efficiently created in the laboratory.

Separate calculations, meanwhile, show specifically how the custom materials will then interact with drug compounds in water—how the materials and drugs swap electrons. The team can then work in the lab or with partners to manufacture the most promising candidates and put them through their paces in the real world.

"Computing allows us to better screen our experimental works. It gives us a north. Points us where to go," Hernandez-Maldonado says.

A 2008 publication in Microporous and Mesoporous Materials by Hernandez-Maldonado and graduate student Sindia Rivera-Jimenez explored nickel attached to sorbent materials called MCM-41 and their interaction with naproxen, a common pain reliever. Results showed that a technique known as grafting for attaching the nickel worked better than a thermal monolayer dispersion technique. Though this method reduced the overall surface area available on the material to adsorb naproxen, the method still proved as efficient as traditional activated carbon methods of removing materials like naproxen. Activated carbon, however, is largely unselective. It pulls out materials that don't need to be removed and is not as efficient, as a result.

"It's fundamental for us to attack specific pharmaceutical compounds selectively in the process," says Hernandez-Maldonado.

'Both states of mind'

Computation acts as a compass for more than just experimentation. It also acts as a compass for education. Students within disciplines like chemical engineering must be prepared to think as computational scientists, too, according to Hernandez-Maldonado. "They have to have both states of mind and get ready for the future. They have to say, 'Computation is a tool I will always use.'"

To that end, he teaches a graduate-level class that includes computational modeling of several nanostructure materials. The class is in its second year and was also offered through videoconference to a second University of Puerto Rico campus. The students perform calculations and small-scale projects to get a sense of how high-performance computing works and its promise and current limits. About 15 projects were run last year, all relying on NCSA resources.

"You need a specific infrastructure for students to work this way," says Hernandez-Maldonado. "NCSA was great for that."

This research is supported by the National Science Foundation (grant number CBET-0546370) and the Puerto Rico Institute for Functional Nanomaterials.

Team members
Arturo Hernandez-Maldonado
Sindia Rivera-Jimenez