The Advanced Research Projects Agency for Energy (ARPA–E) works on a three-year cycle: Funded projects have three years to prove worthy—or not. Program directors who help fund projects such as Plants Engineered to Replace Petroleum (PETRO) or Batteries for Electrical Energy Storage in Transportation (BEEST) have three years to steer the research. And, after three years at the helm as the founding director of ARPA–E, mechanical engineer Arun Majumdar has announced that he will be stepping down in June.

"Under Arun's leadership, we have seen ARPA–E grow from a fledgling program to become a leading agency for innovation and energy research," Secretary of Energy Steven Chu wrote in an e-mail announcing the surprise change. "Arun has been an invaluable resource to me, to the department, to the administration—and we will miss his leadership."

Majumdar has led the agency's quest for what Secretary Chu has called "a second Industrial Revolution, a revolution that gives [the] developed and developing world the energy it wants and needs, but that can be clean energy," overseeing more than $500 million in funding for roughly 180 projects in 12 program areas. He even helped devise program acronyms like PETRO.

Modeled on the Pentagon's successful Defense Advanced Research Projects Agency (DARPA) but uniquely focused on developing energy technologies that can be cheap and have a big impact, ARPA–E has already seen some of its most successful projects garner an additional $200 million in private investment. "Pound for pound, dollar for dollar, it's hard to find a more effective thing the government has done than ARPA–E," argued FedEx chairman Fred Smith at the agency's most recent conference.

Scientific American spoke with Majumdar about ARPA–E and its future at the agency's third annual conference in March.

[An edited transcript of the conversation follows.]

ARPA–E has recently announced some successes like a new, more energy-dense lithium ion battery. As these technologies flower or fail, how do you reassess your programs?
There are the original milestones in the cooperative agreements that we have with projects. Before we start any project, before we even have a contract signed, we negotiate technological milestones, which in [the case of the Envia battery] included both energy density and cycle life. For the successes like Envia, they will probably have to raise some private sector money, which they are doing. General Motors is investing in the technology, for example. They will need more to put up a manufacturing plant. Ultimately, it's a business proposition. I don't know exactly what they're doing. They don't have to share it with us. I would love to see them off and running.

How does it work with other parts of the Department of Energy (DoE)? Do you sometimes hand projects off?
The DoE has applied-energy programs. Some technologies take longer for the private sector to come in. That's where applied energy programs are helpful in general—to carry on the baton to the point where the private sector comes in.

And what about the failures so far?
We negotiate technology milestones, and if projects cannot meet their milestones by a big margin—if the idea just doesn't work out—we first roll up our sleeves as scientists and engineers and try to help them. The program directors are there to say, "Hey, do you have any ideas that could help these guys?" And I read papers and try to help. This is a whole process that lasts for five or six months.

But some ideas just don't work out. It sounds like a great idea but unless you try it you'll never know. And when those ideas don't work out and we know it is a blind alley, we discontinue the project. Do not keep putting money where you know it is not going to work out. Discontinue the project and put whatever money you save, put it in something that is actually working. That is the best stewardship for taxpayer dollars.

Does the U.S. have a fear of failure at present? Do we need to take more risks?
Here's the thing: as a federal government agency our role is [to invest] where things are too risky for the private sector—to do research that may not pan out. If 100 percent of these projects worked out, we're not doing our job. The private sector should be doing that. Our job is to do research. What we do is translate science and apply it. That is something everyone should understand. If the federal government does not support research then we have a problem.

When you're doing research, then by definition some things don't work out. These are home-run ideas, when you try to hit a home run. If you hit every ball for a home run, I'd like to meet that baseball player. You do not. Some of them you are going to miss but that is what research is about.

I don't call them failures, I call them opportunities to learn. So these people can come back again with a better idea. That is how Americans have always done things. You think Edison got everything right the first time? I doubt it. You think the Wright brothers got it right the first time or even first 10 times? No. Eventually they got it. But they missed a few times.

How do you balance between this goal of hitting home runs and the reality of hitting singles?
We are certainly not going for singles, we are not going for incremental progress. We made it very clear if you look at [our calls for proposals], you will see the definition of what is incremental and what is not. And we are clearly going for home runs. Now, can I claim I've hit a home run yet? That's going to take a few more years to find out, but the trajectory of that ball that I hit or that we are hitting seems like a home run. That's all I can say right now.

But, for example, you have a new program in natural gas and that's certainly not a risky fuel. Why do we need breakthroughs in natural gas?
I can tell you why. For example, if you were to use natural gas for light-duty vehicles—passenger cars—it is really expensive to put the infrastructure in place for compressed natural gas that is similar to the gasoline stations on every street corner. It will cost you several hundred billion dollars. There may be a better way to do it.

The better way to do it is to see if you could do home refueling, because natural gas reaches your home. For that to pay back in four to five years, that technology doesn't exist today. That needs a breakthrough. If you could bring down the cost of natural gas transportation for light-duty vehicles so you could refuel at home, then the adoption of that in the mass market will certainly be there. If it pays for itself in four years or five years, I think a lot of people would be thinking of buying natural gas vehicles. So that is what we are trying to do.

But natural gas vehicles already exist.
Sure, [for] many of the long-haul trucking companies. Navistar announced it's going to be making [liquefied natural gas, or LNG] trucks. I was at the Port of Long Beach and we saw the LNG trucks there. Clean Energy Fuels is putting in the infrastructure for LNG and putting it every 200 miles along major trucking routes. That is fantastic. Their cost is going to pay for itself in a few years, and the truckers' cost is going to pay for itself in a few years because they are running 100,000 miles a year.

But that is not true for light-duty vehicles—passenger cars—so that's where the gap is. It is very expensive to put that infrastructure like that all over. So home refueling is something of a long shot, and it's risky. If someone can come up with a technology that can do that, I think it could have a major impact on our transportation sector. It could reduce oil imports significantly if you got mass adoption of natural gas vehicles.

The same kinds of suggestions were made about hydrogen fuel cell vehicles a few years back and, of course, hydrogen can be made from natural gas. So why not hydrogen?
It's the same thing. People always talk about storing hydrogen. The best way to store hydrogen is in a hydrocarbon. In natural gas, for every atom of carbon, you've got four atoms of hydrogen. That's a hydrogen fuel. Even if you do a hydrogen-based thing, which I think is worth considering, the question is: Where does the hydrogen come from? It comes from natural gas. Well, why not just use directly the natural gas?

Now natural gas is all the rage. Five years ago we were enamored of the hydrogen economy, and in five years maybe we'll go back to nuclear. How do we implement a sustained and consistent approach to the energy problem?
We have to be clear by what we mean by a hydrogen economy. If you talk about where the hydrogen comes from, it can come from natural gas, it can come from nuclear. It can be used directly in engines, it could be used in fuel cells. One of the advantages of hydrogen is that if you were to use fuel cells, the overall system efficiency could be quite high. That's a great advantage. You can use proton exchange membrane fuel cells. You don't have to go to high temperatures.

The challenge with natural gas and fuel cells is you have to use solid oxide fuel cells. [For] solid oxide fuel cells you've got to go to high temperatures, and there are other issues for the transportation sector. I think a fuel cell is a viable option for transportation. The DoE is doing [research and development] in [its Energy Efficiency and Renewable Energy division], and ARPA–E is very supportive of that. We need to do that R&D, bring down the cost and provide that option. Our job is to provide options for the nation. Our job is not to decide what the business should do, that's business's role. The government's role is to provide options and create competition. And that's what we are trying to do.

What about algae? The federal government has a long and tortured history of research in algae but is, at present, back into that research.
In biofuels you have to have a portfolio approach. It started off with corn to ethanol—fine. Now there is talk about cellulose. Before that even, there was sugarcane to ethanol—great. Then there was cellulose to sugar, okay, and then to ethanol, and now to drop-in fuel. Fine, that's another approach. It's all plant-based. Algae directly to oil, that's another approach.

We are taking a completely new route of engineering plants like loblolly pine trees to produce oil. Our other approach is "electrofuels," which is completely new. It doesn't even use plants.

The thing is we don't know. This is what the president talked about: an all-of-the-above approach. We look at biofuels, we're taking a portfolio approach because we don't know which one is going to scale in cost and volume and be competitive with petroleum-based fuels. At the end of the day, I think everyone's goal is the same: it's to reduce our oil imports.

According to Bill Gates, the nuclear industry hasn't had any innovation since the 1970s. Is there room for ARPA–E-style breakthroughs in nuclear power?
We are open to everything. We are looking at everything. We have considered nuclear before. Nuclear is always considered base load [the type of power plant that consistently pumps out electricity], but if you can store the nuclear heat, you can use it as peaking, [or only when the electricity is needed]. This is the reverse of solar and wind because they are intermittent, and people want solar and wind to be base load. We want nuclear to be peaking because you get more revenue that way. The price of electricity goes up when demand is high. Those are the kinds of things we are considering.

We have a nuclear energy program in DoE and we work very closely with them to identify where ARPA–E can provide value in a sector where if ARPA–E puts in $30 [million] to 40 million, the typical program size, it can really move the needle significantly. That's what we'd like to see. We are exploring that. So nuclear certainly is on the table.

One of the big issues in the energy sector is water—whether it's water used to cool a nuclear power plant or water used to produce biofuels. Does ARPA–E plan to do any work in the area of the energy–water nexus?
We had projects on water and desalinization. It didn't quite work out. We had a workshop on water. It didn't quite create a program. We're actively looking at people who might be interested in water to recruit them. We have an internal ARPA–E fellow who has looked into water as to what the opportunities are. So yes, we are very interested in water, especially where ARPA–E can go in and really move the needle on something.

The way we work is that we hire really smart people, technically savvy, who have one foot in science and engineering and the other foot in business, [and] who can translate science to technology to business. And we empower them: Go figure out what's the best opportunity. If you're so smart, go figure it out. And they'll tell us and have to explain it to us. That's how we are going to work.

What's next for ARPA–E?
We are putting a great team together. At the end of the day it's the people. And we are recruiting some of the top-notch people in the technical community to come to ARPA–E and serve the nation.

And you, you presently have two jobs. How's that going?
Let me use the words that Senator [Lisa] Murkowski used in my confirmation hearing: Everyone's wondering when do you sleep? I'll just leave it at that. She may be right.

I'm going to guess that you sleep on the plane back to Berkeley, Calif., where your family is, and that's about it.
I do. I am asleep before the plane takes off. When the plane hits the runway I'm gone. And I try not to answer phone calls or e-mails when I'm home.

It's a three-year cycle and then out at ARPA–E, whether for program leaders or projects. What does that mean for the ARPA–E team in the long term?
There's some flexibility in the statute [that set up ARPA–E]. It's a director's discretion to have some flexibility. You don't want to make that rule hard and fast because of discontinuity of programs and all that. That discretion is there in the statute and I'll use it when I need to. But I think there's an understanding in ARPA–E that once you've done three years, you ought to move on.