Harvard Harnesses Solar Power through the Grid 

This June, Harvard's Clean Energy Project will release a list of the top 20,000 organic compounds used for the printing of cheap photovoltaic cells (PVCs) to solar power developers.

The list is the result of a two-year long crowdsourcing project that started with about seven million molecules. If the effort is successful, the cost to cover a one-meter square wall with PVCs could equal the cost to coat it with paint, which would be a significant boon to the industry.

Currently, the most popular PVCs cost about $5 per wafer and are made of silicon, and convert about 12 percent of light that hits them into energy. Current organic photovoltaics offer about four or five percent efficiency, which means that in order for them to be cost effective, they can cost no more than 50 cents.

But the Clean Energy Project is introducing a variable that could change that equation. By taking advantage of IBM's World Community Grid to compute the best molecules for organic PVCs, the project may not only find cheaper, but also more efficient molecules.

To accomplish these computations, the World Community Grid doesn't use an HPC cluster, but instead draws from over 6,000 personal computers around the world. The software can be installed on any computer, and while the machine idles the software taps its spare power to run its calculations.

Normally, computational chemists analyze photovoltaic potential one molecule at a time. This means that over the last few years, only two or three new compounds have been discovered that are capable of delivering 10 percent efficiency. Now 20,000 additional molecules can be added to that list.

What's even more remarkable is that some of the molecules sport 13 percent conversion rates, making them even more efficient than silicone PVCs.

But until the June release, the full implications of this research, including which molecules are most promising in terms of both efficiency and cost, is still unknown.

“We're in the process of wrapping up our first analysis and releasing all the data very soon,” said Alan Aspuru-Guzix, associate professor of chemistry and chemical biology at Harvard.

Full story at Computerworld