If all goes as planned, sometime in the next decade an American robotic lander will arrive at a burgeoning moon base toting a small nuclear reactor. Inside the reactor a boron control rod will slide into a pile of uranium and start a nuclear chain reaction, splitting uranium atoms apart and releasing heat. Next that warmth will be piped to a generator. Then the lights will come on—and stay on, even through long, cold lunar nights.

After a half-century struggle to develop a nuclear power plant for use in space, NASA just completed a successful test of a brand-new design. The next milestone for the new reactor, called Kilopower, could be an inaugural spaceflight sometime in the 2020s. Developed with the Department of Energy (DoE), Kilopower marks the first new nuclear reactor of any kind in the U.S. in 40 years. It could transform energy production for space exploration, especially for permanent human outposts elsewhere in the solar system.

Current space missions use fuel cells, nuclear batteries or solar power. But a night on the moon lasts two weeks, and the strength of sunlight on Mars is only about 40 percent that of Earth. “When we go to the moon and eventually on to Mars, we are likely going to need large power sources not dependent on the sun, especially if we want to live off the land,” says Jim Reuter, NASA’s acting associate administrator for space technology.

Kilopower is a small, lightweight fission reactor that can provide up to 10 kilowatts of electricity. NASA says four 10-kilowatt Kilopower units would provide enough electricity to power a human outpost on Mars or the moon. For comparison, 40 kW is enough to power three to eight typical American houses, says Claudio Bruno, an engineering professor at the University of Connecticut who studies nuclear energy. He adds that 40 kilowatts would be roughly equivalent to 60 horsepower. “You get a sense that it is very modest power. In truth, if you want to do anything useful, especially if missions to the moon or Mars are crewed missions, you need lots more than that,” he says. “But this said, every time this research was done in the past, there were hues and cries about the mortal danger of nuclear power. This is the first time [in decades] they are talking about powering an electricity generator with a nuclear reactor, so it is a first, positive signal.”

In space exploration nuclear energy can be used in two main ways: for generating electricity or for propulsion. Kilopower would be used to produce electricity, much like a power plant on Earth. It would likely produce more than a single spacecraft would need, making it a better fit for larger surface outposts. Kilopower could also be used to drive a spacecraft, primarily by powering an ion engine, although NASA presently has no definitive plans to use it in this way.

Kilopower has been in development since 2012, but its legacy reaches much farther back, to NASA’s Systems for Nuclear Auxiliary Power (SNAP) program of the 1960s.

The SNAP project developed two types of nuclear power systems: The first was radioisotope thermoelectric generators, or RTGs, which capture heat from radioactive decay to provide warmth and electricity. Dozens of deep-space spacecraft have used RTGs, including the Curiosity rover on Mars and the New Horizons Pluto probe now exploring dwarf planets of the outer solar system. The second SNAP project was for a fission reactor, which splits atoms to produce energy. This is the same type of technology that powers nuclear submarines. NASA launched one nuclear power plant, called SNAP-10A, in April 1965. It worked for 43 days and produced 500 watts of electricity before a part failed; it remains in Earth orbit today and is considered space junk.

During the 1960s and ’70s NASA also researched nuclear power for rocket propulsion under the Nuclear Engine for Rocket Vehicle Application (NERVA) program. This would have used nuclear reactors to heat hydrogen and expel it through a nozzle, much like a chemical rocket burning fuel to push a rocket forward. But this program ended in 1973.

Russia has flown more than 30 fission reactors in space, according to the World Nuclear Association. But after Pres. Richard Nixon canceled NASA’s nuclear propulsion research in 1973, Russia backed down from its program, too, Bruno says. “Everything went basically on the back burner or was frozen by 1973,” he adds. “By 2018, most of the people working on this are either retired or passing away. We don’t have firsthand knowledge of what they did. We have reports, sure, but reports don’t speak to you. Humans speak to you.”

The thaw began in 2012, when NASA and the DoE performed a preliminary test of Kilopower’s progenitor, the Demonstration Using Flattop Fissions (DUFF) experiment, producing 24 watts of electricity. DUFF used a heat pipe to cool its reactor and demonstrated the first use of a Stirling engine to convert reactor heat into electricity. (Stirling engines use external heat to drive a piston, which turns a crankshaft to produce power.) Following the DUFF test, NASA’s Game Changing Development program endorsed Kilopower and the project received its first funding in 2014.

The latest NASA and DoE Kilopower tests occurred from November 2017 through March of this year. Dubbed Kilowatt Reactor Using Stirling Technology—KRUSTY, which like DUFF pays homage to The Simpsons—the tests put the Kilopower reactor through its paces, culminating in a 28-hour run where the reactor was turned on, operated at full power, then cooled and shut down. Operating at temperatures of 800 degrees Celsius the reactor produced more than 4 kilowatts, says Marc Gibson, Kilopower lead engineer at NASA Glenn Research Center where KRUSTY took place.

NASA and DoE officials say the reactor is safer than previous generations because of how it works. The fission chain reaction is passively controlled and can even be stopped, using boron control rods and beryllium reflectors. Atom-splitting would not begin until after the reactor is far from Earth. If a reactor or its rocket exploded on the launch pad, the uranium 235 in the core would expose people one kilometer away to radiation levels no more than natural background levels, according to Patrick McClure, Kilopower project lead at the DoE’s Los Alamos National Laboratory. “Under all worst-case situations, we don’t believe there is any chance the reactor would come on accidentally during a launch accident,” he says.

David Poston, chief reactor designer at Los Alamos, says a similar reactor could provide electricity to power ion thrusters, which could in turn propel a spacecraft. But the amount of material required to start the chain reaction would likely mean a reactor too large and too heavy for practical use, according to Bruno. NASA is separately developing a new uranium-based nuclear thermal engine concept, which would work much like current chemical rockets to accelerate fuel out the back end of a thruster. But the Nuclear Thermal Propulsion project started in August 2017 is not as far along as Kilopower.

Most nuclear-powered spacecraft use RTGs, which simply harness heat from plutonium decay to make electricity. But RTG efficiency is extremely low—and what’s more, plutonium dioxide fuel is in short supply. The DoE resumed plutonium 238 fuel production in 2015 after a 30-year gap, but currently there is only enough in the nation’s stockpile to power NASA’s 2020 Mars rover and maybe one or two other potential missions to the outer solar system.

Kilopower could serve as an alternative—but that’s a big maybe, officials and experts warn. “We kind of start where RTG stops, from a power standpoint. We are kind of taking [up] where they are leaving off and trying to give us a power range for things like human exploration, where you need tens to hundreds of kilowatts,” Gibson says. In other words, human activities on the moon or Mars would require 10 to 100 times more power than what a single Kilopower reactor or even a handful of reactors is projected to produce. But Poston says the reactor's modular design can easily be scaled up to meet those needs.

Nevertheless, Kilopower is an important step toward a workable nuclear power plant for use in space, Bruno adds. The next step would likely be a test of the reactor in space. NASA has not yet approved such a mission, but at a press conference earlier this month Reuter says the next 18 months will be devoted to figuring out how such a test flight would work. One possibility is to fly a small Kilopower reactor on a lunar lander, which may be developed under NASA’s newly moon-focused exploration mission.

Poston says the successful ground tests are an important step in the next phase of human space exploration. “We demonstrated a concept NASA can use right now. To me, the most exciting thing is the potential. This really is the first step in using fission power in space,” he says.