Distributed computing search for next-gen photovoltaics

Usually, when the word "distribute" and it’s derivatives are used in relation to solar, it’s in terms of distributed generation. But the Clean Energy Project led by a team of researchers at Harvard University, Universidad Nacional Autonoma de Mexico and Haverford College are using distributed computing to identify the most likely molecules to foment the next generation of organic photovoltaics, and fuel cells and they need your—well your computer’s—help!

The distributed computing model has been used to search the stars for aliens, to research new drugs for cancer and Parkinson’s and other uses. Now it’s being put to use to help develop organic photovoltaic molecules to produce alternatives to photovoltaics that use inorganic compounds or hazardous materials.

Basically, under the distributed computing model, when a computer with an Internet connection is not being used fully—or at all, but is still on, some of its processing power is diverted to working on the issue. This is done through IBM’s World Community Grid.

In this case, participating computers are used to develop theoretical molecules that could create organic photovoltaic compounds that are around 15 percent efficient at converting sunlight into electricity, according to the project Web site. That’s close to the average efficiencies of photovoltaics now available.

The distributed computing model can help significantly reduce the cost of the computer research.

“The more people we can get, the better. These kinds of chemistry calculations are expensive,” said Roberto Olivares-Amaya, a researcher with the project and co-author of “The Harvard Clean Energy Project: Large-Scale Computational Screening and Design of Organic Photovoltaics on the World Community Grid,” which was published in the August 2011 issue of The Journal of Physical Chemistry Letters.

The study has thus far identified 2 million to 2.5 million potential molecules for study, according to Olivares-Amaya.

“Each with different geometries within every molecule,” he said. “These are normally polymers. We’re hoping there’s not a difficult road to synthesizing them.”

Of the 2 million or more compounds, only a handful will show potential for use in low-cost photovoltaics. Already, the search has found roughly 30,000 compounds that show promise, Olivares-Amaya said.

“This 30,00 represents about 1 percent of the search space,” he said.

Once the research is farther along, the group will make the results public.

“We are planing to [make] the results library open by next year,” Olivares-Amaya said.

That’s one of the requirements of collaboration with IBM’s platform.

“We’re working on details to make it publicly available,” he said.

But it may take until 2013 to work out all the kinks, he said.

While making it public lends the research to the open-source software model, the group would prefer to see its collaborators synthesize the materials first, according to Olivares-Amaya. 

Photo courtesy of Clean Energy Project