ARPA-E: Three Answers and Three Questions

On Monday, the Department of Energy announced $151 million in grants for ARPA-E, its pie-in-the-sky, high-risk energy research program. Thirty-seven grantees got funded, and Energy Secretary Steve Chu said that the program would serve as a "bold, transformational" step and would "spur the next Industrial Revolution" of energy research. Three things about ARPA-E we've learned, and three questions that remain:

Energy Secretary Steve Chu is serious about shaping ARPA-E in the Bell Labs model. It's pretty rare that federal research managers will say publicly that they expect some of the work they are funding to fail. Usually, they support basic research that has value because it adds to the scientific literature, or they support commercial work that generally leads to products. ARPA-E is different: It supports potentially commercial work that may well crash and burn. As Chu sees it, the way to go is the Bell Labs model: Support a number of risky research projects run by highly talented scientists, and if one or two pan out, the whole effort is worth it.

There's a lot of pent-up interest in energy research, and it's coming from everywhere. 

Among the awardees were experts on jet engines who are making new, small-scale wind turbines and scientists making transgenic plants who want to preprogram enzymes to break down plant tissues directly into the cells of the biomass crop. The flexibility of ARPA-E will strengthen DOE's ability to support areas traditionally outside of the energy sphere. The grantees, whose awards averaged $4 million each, were culled from 3700 original applicants, each of whom simply wrote short descriptions of their proposed research, to 300 applicants who wrote full proposals, to the 37 winners. The large number of applicants allowed DOE to pair some together to do joint projects. That interdisciplinary approach—which has been one of DOE's long-standing weaknesses—is another trademark of the Bell Labs model. DOE official Matt Rogers, who is not a scientist but is knowledgeable about energy, said before the ARPA-E process that he considered battery science "fairly prosaic." But the promise of big money brought in lots of interesting battery projects—and, for what it's worth, DOE funded several battery projects but little on advanced photovoltaics.

ARPA-E will fund ideas that haven't even been published yet. DOE is trying for truly "transformational" things that will totally revolutionize the way we use energy. To do that, the agency is investing in some technologies that have little or no track record at all. One of the projects that DOE has celebrated most is a $7 million grant for MIT battery scientist Don Sadoway to use molten metal to store energy as part of the energy grid. This could allow windmills and solar photovoltaic facilities, which are intermittent by nature, to store energy when they are running at peak power regardless of demand. DOE has funded next to nothing in this area before. More interestingly, Sadoway has no academic papers published on the idea. He went to ARPA-E first, and given his strong reputation in battery research, they took a $7 million chance on his novel idea. "If successful, this battery technology could revolutionize the way electricity is used and produced on the grid," DOE said in a press release. If not, Sadoway will probably return to work on more conventional battery strategies. And DOE will take a mulligan.

Three questions:

What can we learn from this first round that will give clues about ARPA-E's future? That remains to be seen, but probably little. To give out this money as soon as possible, DOE threw together a team of experts—some on the federal payroll, some of them outsiders—to first wade through the volume of proposals and then decide on the final winners. But the process didn't really set up ARPA-E as a solid agency going forward. Director Arun Majumdar was only approved by the Senate for his job last week, and he hasn't hired a full complement of staff yet. (DOE said earlier this year they wanted 12 program managers.) Those managers, once they're in place, have $251 million to spend, have until October of next year to do so, and have to scramble to figure out how. (They're holding public workshops to get ideas.) After that, the budget could be much, much smaller for the agency.

How tough will ARPA-E be on grantees? So far, none of the ARPA-E awards has been negotiated in to final contracts with the universities, businesses, and national labs that have won money. DARPA, the Pentagon blue-sky research outfit after which ARPA-E was modeled, is well known for signing contracts that allow it to cut off scientists if they don't perform, and very quickly. Will ARPA-E be as ruthless?

What will Congress do on ARPA E in the future? ARPA-E got very lucky with $400 million as part of the stimulus package—but that's because House Speaker Nancy Pelosi supported the concept as the stimulus package was being put together. Earlier this year, Congress rejected giving an appropriation for the agency as part of the 2010 budget, arguing, of course, that the new agency didn't need the money.

But the budget deficit is skyrocketing, and one can easily imagine appropriation committees on Capitol Hill giving, say, $20 million to the program in the future as the budget tightens. DOE’s Matt Rogers said that Chu specifically wanted grants to be of medium size. In the past, he said, DOE tended to spend hundreds of millions of dollars on big projects—such as fusion experiments or demonstration projects—or spread the money around into very small grants to get lots of scientists involved. Instead, ARPA-E went for medium-sized grants for big impact on single ideas. But ARPA-E can only support a handful of those medium-sized projects if it gets in the tens of millions of dollars each year from Congress. ARPA-E lacks the natural political constituency that areas like high-energy physics or the national labs have, so maintaining the funding could be hard as the winds shift on Capitol Hill.