The Advanced Research Projects Agency-Energy (ARPA-E), a controversial federal agency created to fund research into radical new energy technologies, has run into some inevitable problems in its first months of operation. But the agency's director, Arun Majumdar, says the agency is learning from the first round of funding and improving the selection process. And the agency has already helped advance technologies that may otherwise have fallen through the cracks.

The agency was first proposed in a 2007 report from the National Academies as a way to maintain America's competitiveness in science and technology. It received funding this year as part of the American Recovery and Reinvestment Act of 2009. The agency is modeled on the Defense Advanced Research Projects Agency (DARPA), which pursues high-risk research with potential military uses.

Some critics have complained that ARPA-E covers ground already accounted for by other federal agencies. They also allege that it puts the government in the inappropriate position of "picking winners" among potential new technologies. Supporters counter that ARPA-E offers a way to fund breakthrough approaches to energy that the private sector is too conservative to fund, and that other agencies typically ignore.

The key difference with ARPA-E, Majumdar says, is that unlike most other government agencies, it has permission to take big risks. "It's in our mandate. We can take risks, and we can fail and that's okay." For example, other groups within the U.S. Department of Energy are funding improvements to lithium-ion batteries, which are already used widely. "They cannot digress too much from that because there are certain metrics to meet," Majumdar says. "ARPA-E gives you a chance to look a little beyond." If even a few of the risky projects succeed, he says, "we will leapfrog over other," more conventional approaches.

Selecting which projects to fund, however, has proved challenging. The agency was swamped with about 4,000 initial proposals, and could only back about 1 percent in its first round of funding. Some researchers have also complained about the lack of qualified reviewers--many of the most qualified potential reviewers in academia and industry were disqualified because they also submitted applications. What's more, in the first round, researchers did not have a chance to respond to reviewer's criticisms, which made it impossible to correct misunderstandings.