The sun is constantly raining free energy down on Earth, but we have thus far been unable to make proper use of it. The efficiency of solar cells is still low, and the cost is too high to make this clean power source a viable option in most areas. A team of researchers out of Harvard is seeking to change that by harnessing the power of distributed computing and adding a little dash of quantum mechanics. Harvard’s Clean Energy Project has now released a public database of potentially useful organic molecules containing thousands of candidates for further testing.
This new database is only the latest product of the so-called Materials Genome Initiative, a federally-supported program to speed up the rate of innovation in materials science. Developing completely new substances with useful properties has traditionally been an extremely tedious process. Scientists would have to synthesize numerous compounds before finding one that behaved as needed, and better materials might never be evaluated. The Materials Genome Initiative takes a different approach — it’s about prediction.
Instead of making every new material, the goal now is to simulate organic molecules with massive amounts of computational might, and figure out how they will work in real life. Beginning with 26 promising molecular fragments, millions of different permutations of those components were analyzed. Harvard didn’t even need a supercomputer — it just needed a bunch of regular ones.
IBM lent a helping hand with its World Community Grid, a distributed computing platform of over 600,000 volunteers donating spare processor cycles. Using this system, each molecule could be modeled and analyzed in 12 hours using complex quantum chemistry models. The process could take place in parallel because there were so many individual computers involved. The World Community Grid contributed over 17,000 hours of computing time, and generated a whopping 400 terabytes of data. In short, it saved researchers a lot of time, and the results weren’t bad either.
Most existing solar cells are silicon-based, and boast an efficiency of 15-20%. That means they convert only 15-20% of solar energy into electricity. Organic molecules, like the ones being investigated by the Harvard project, are usually in the single digits. A few have been identified that are over 10% efficient, but the record for organic solar cells is only 11%. By comparison, the database contains over 1,000 molecules with a predicted efficiency of 11% or higher.
None of these compounds actually exist yet — they’re all theoretical, packed inside the database awaiting their moment to shine. The next step is for researchers to synthesize the most promising materials and see how closely the simulations match up with reality. The hope is that one or more organic compounds will show promise as a solar cell substrate. Panels made from such a material would likely be flexible and much lighter than silicon-based ones. The organic panels would also be considerably cheaper to produce, which might be the thing that finally kicks solar power into high gear.
LuckyRobot is brought to you by Frequency Group