Silicon still rules in solar cells, but Harvard has now ranked 2.3 million possible replacements

Harvard gave a big gift to the solar community today: a database of 2.3 million materials that highlights those with the most potential to be used in solar cells. All of the materials are carbon based, which researchers believe could someday replace silicon cells due to its low cost and flexibility.
The database includes more than 35,000 materials that are at least 10 percent efficient at converting solar energy into electricity. The best are 11.13 percent efficient, which is in line with the current record for organic materials. But all of them degrade to 4 to 5 percent when actually incorporated into a solar cell. Most silicon solar cells hit at least 15 percent.
Much more research will need to be done before organics are in line with silicon. Scientists will need to increase their efficiency and prevent them from degrading when they come in contact with the environment, ideally increasing their lifetime to more than 10 years. Because the database is open access, any researcher can search it to pinpoint the most ideal material for their line of research. Materials are searchable by features like their efficiency, mass and energy level, eliminating the need for labs to identify and catalogue them on their own.
Harvard researchers built the database with IBM’s World Community Grid, a program that invites volunteers to donate their unused computing time to scientific projects. The equivalent of more than 17,000 years of computing time resulted in 400 TB of data and 20,000 material characterizations a day.
The database is part of the federal Materials Genome Initiative, which is meant to double how fast advanced materials are developed. Officials made a slew of announcements today to mark the initiative’s two-year anniversary.
The Georgia Institute of Technology and University of Wisconsin-Madison announced they will create material innovation institutes with $15 million in funding. They will launch an accelerator network with the University of Michigan. The National Institute of Standards and Technology will also fund an accelerator with $25 million.
Following Harvard’s open-access example, Lawrence Berkeley National Laboratory, MIT and a private research and development company will develop software tools that predict material behavior. Digital modeling could make it 10 times faster to develop a new material.