NASA Goes Nuclear

When your batteries are dead and solar power is only a distant memory, you’re going to need something else in your power pack.

OFFICIALS AT NASA HEADQUARTERS IN WASHINGTON, D.C., WERE BATTING AROUND names for their soon-to-be-announced nuclear program last fall when Administrator Sean O’Keefe offered a suggestion. Why not call it Project Prometheus, after the giant in Greek mythology? It was Prometheus, after all, who brought fire, and with it civilization, to humankind. So far so good. But the story has a disturbing end. For Prometheus’ effrontery, Zeus had him chained to a rock, helpless to defend himself as an eagle pecked out his liver, over and over, day after day.

“We had several rounds of e-mails with the Administrator over whose liver was going to be pecked out,” jokes Al Newhouse, a former Navy nuclear engineer who heads Project Prometheus at NASA headquarters. “That was funny, except it was always mine.”

After decades of false starts and dashed hopes, Prometheus marks the return of nuclear reactor development for U.S. spaceflight. In June 1993, after 10 years of work, Congress voted to end the last space reactor project, called SP-100, before it even got to the point of ground testing. Newhouse remembers the day well. He had moved over from the Navy to join the Department of Energy as SP-100 director. “I was put in the position of shutting down the program that I was brought in to nurture and support,” he says.

Now O’Keefe, a former Secretary of the Navy who is familiar with nuclear-powered submarines and aircraft carriers—and whose father was a nuclear sub engineer—has recruited Newhouse and other Navy veterans with nuclear expertise to run Prometheus, which will produce advanced systems for both power and propulsion.

This time Newhouse hopes things will be different. The trouble with SP-100, he says, was that it never had a guaranteed customer. While the reactor was in development, both of its intended users—Ronald Reagan’s “Star Wars” space weapons program and the (short-lived) proposal of his successor, George Bush, to send astronauts to the moon and Mars—fell by the wayside. This time, though, NASA has asked the energy department to build a reactor for a specific purpose—to power a robotic mission to explore three moons of Jupiter as early as 2011. After that, it will provide electricity for future planetary spacecraft far more capable than past Vikings and Voyagers, which used either solar power or small plutonium batteries.

Prometheus is a gamble, both technically and politically—because launching radioactive material is likely to generate protests and create a public relations problem for NASA. But many people, O’Keefe foremost among them, believe nuclear power is the only way for NASA to take the next step in space exploration. Agency science chief Edward Weiler recently told a committee of the National Research Council that O’Keefe “not only calls it the future of planetary exploration, he calls it the future of NASA.”

If you compare the proposed Jupiter Icy Moons Orbiter (JIMO) to the Cassini spacecraft now headed for a July 2004 rendezvous with Saturn, it’s easy to see what he means. Cassini’s electricity comes from radioisotope thermoelectric generators, which have been standard equipment on spacecraft venturing too far from the sun to rely on solar power. The RTGs on Cassini produce power from the decay of plutonium and generate about 900 watts, enough electricity to power nine standard light bulbs. JIMO’s nuclear reactor will produce 100 kilowatts, or several times the average daily household use.

For planetary exploration, that kind of power output is revolutionary. It means data will come back to Earth in unprecedented volumes—120 CD-ROMs’ worth for the entire mission, compared to a couple of floppy disks for Cassini. Instead of observing Jupiter’s moon Europa for a few hours at close range, which had been the plan for a non-nuclear mission NASA was considering as recently as last year, JIMO will study three Jovian moons for a total of 180 days. And it can carry a much more powerful sounding radar to probe for an ocean suspected to lie beneath Europa’s icy crust.

First, though, Prometheus has to deliver the fire. On a conceptual level, a space nuclear reactor would work much like a reactor on the ground. Neutrons given off by a radioactive fuel, in this case uranium, would strike other uranium atoms, which would then split to create more neutrons, perpetuating the reaction and generating heat, which would be absorbed by a coolant and converted to electricity.

But most nuclear reactors aren’t launched on rockets from the densely populated Florida coast. So Department of Energy engineers will have to assure critics that in the event of a launch accident, a space reactor won’t suddenly start splitting atoms. “Safety becomes the driver in the reactor design,” says Earl Wahlquist, head of DOE’s Office of Space and Defense Power Systems in Maryland.

In a space reactor, the fuel elements would be surrounded by neutron-reflecting materials. Without the reflectors in place, there aren’t enough neutrons bouncing around to cause a chain reaction. So engineers would devise a system to move the reflectors into place to start the reactor, then back them out to stop it. With the reflectors in safe mode during launch, the uranium fuel would be no more than “marginally radioactive,” Wahlquist says. If the rocket exploded on the launch pad or suffered some other catastrophic failure on its way to orbit, “the rocket fuel would be more toxic than the uranium,” he says.

JIMO and its reactor would be launched on a conventional rocket to an altitude of just over 600 miles; only then would the reactor be turned on. At that altitude, say NASA officials, if something went wrong after controllers start the reactor, it wouldn’t pose a threat to people on the ground. SNAP 10A, a reactor-powered spacecraft launched by the U.S. military in 1965, has been circling Earth ever since malfunctioning on its 43rd day of operation. About 1,000 years from now, its orbit will have decayed to the point where the spacecraft will reenter the atmosphere. By then, its radiation will have dissipated, and “we think it will be [just] a hunk of metal,” Newhouse says.

DOE expects to build the JIMO reactor at one of its facilities, most likely the National Environmental and Engineering Laboratory in Idaho or the Argonne National Laboratory in Illinois. Wahlquist says it won’t be a simple matter of resuming work on SP-100; other designs will also be considered. For example, SP-100 used liquid lithium metal for its coolant, but Prometheus may use a light gas like helium, or vapor transported through heat pipes. Whatever the choice, the reactor will have to be as light as possible, a requirement for any hardware that is space-bound.

Once the JIMO reactor is turned on, the heat it produces will be converted to electricity to drive a new type of thruster that propels the craft with a glowing stream of ions. “This is not a nuclear rocket,” Newhouse says, still chafing from an article in the Los Angeles Times last year that failed to distinguish Prometheus’ nuclear electric engines from more advanced—and controversial—nuclear thermal rockets, which would circulate hydrogen through a reactor and spew the exhaust out a nozzle. JIMO’s reactor is only a power source, not part of the engine itself.

Stanley Borowski, a 15-year veteran of the agency’s Glenn Research Center in Ohio and unofficial keeper of the nuclear flame at NASA, has another nit to pick. The word “is pronounced ‘nu-clee-ar,’ ” he says. “It’s not ‘nook-u-ler,’ which still a lot of people say.” After SP-100 was scrapped and all talk of nuclear-powered Mars missions ended in the early 1990s, Borowski, a nuclear engineer with a Ph.D., retreated into the bowels of the Glenn center, where he continued working on low-level internal studies. Now his field is hot again. The budget plan that NASA sent last summer to the White House Office of Management and Budget, where O’Keefe used to work, included a stepped-up nuclear program. Having seen official excitement rise before, only to fade away quickly, some nuclear proponents were skeptical that the new plan would go anywhere. “Quite frankly, I didn’t think we had a ghost of a chance,” Newhouse says, “but it was approved.” Prometheus was born.

Engineers Dave Manzella and Rob Jankovsky bend down to look through a porthole at the base of a white schoolbus-size vacuum chamber at the Glenn center. Inside the tank, a circular rocket engine about the size of a large pizza gives off a steady, pale blue glow, like a TV in a darkened room. The only sound is the hum of the chamber itself. No need to hide in a blockhouse from the thundering rocket blast. In fact, the thrust from this engine is imperceptible to all but the sensitive disk it’s mounted on.

It will be just such an engine, powered by a nuclear reactor, that pushes JIMO toward Jupiter and lets it maneuver through the Jovian system with a new kind of nimbleness. The thrust produced by ion propulsion—the slow, steady expulsion of ions to accelerate a spacecraft—ranges from a fraction of an ounce to three pounds, minuscule by the standards of most liquid-fuel rockets. Cassini’s twin thrusters, for example, produce 100 pounds each. But, explains Jankovsky, chief of Glenn’s onboard propulsion branch, the ion engine would be so fuel-efficient that it could run continuously, building up speed and eventually outpacing chemical rockets.

When the Galileo probe approached Jupiter in 1995, it fired its liquid-fuel main engine for 49 minutes to slow down so Jupiter’s gravity could pull it into orbit. After that big burst, Galileo had only enough fuel left for minor tweaks to its trajectory. Most of the subsequent course correction came from carefully timed gravity-assisted swing-bys of the Jovian moons. “If you look at a chart of the Galileo orbits, it looks like a line drawing of a flower where each orbit represents a petal of the flower,” explains Ron Greeley, a planetary geologist at Arizona State University.

Mission designers measure such maneuvers in terms of delta-V, or change in velocity—basically, how much energy is needed to change a spacecraft’s speed and direction. Cassini carries enough fuel to provide a total of 6,500 feet per second of delta-V over the lifetime of the mission. JIMO may have 30 or 40 times that. “With JIMO, we’ll orbit Callisto, then slip over and orbit Ganymede, and finally over to orbit Europa,” says Greeley. Such dramatic, energy-demanding orbit shifts were well beyond the capability of earlier planetary spacecraft, because the trajectories would have required many times more liquid fuel than they could affordably carry.

The efficiency of a rocket is generally given in terms of specific impulse, measured in seconds. A typical planetary spacecraft thruster might have a specific impulse of 300 seconds. “People who build chemical rockets would kill for a couple extra seconds of specific impulse,” says Jankovsky. With JIMO’s ion drive, NASA engineers hope to achieve 4,000 seconds.

As he enters the nearby Electric Propulsion Research Building, Jankovsky points to a huge circular engine in a corner. It measures five feet across and looks like the housing of a large industrial fan. Engineers tested this 200-kilowatt ion engine in a vacuum tank here in 1967, back when it was assumed nuclear reactors would be generating millions of watts of electricity for future missions to Mars.

The engines being developed in this building are far less ambitious, but still an advance over the ion engines that have flown in space so far. At one end is a cathode tube that spits out electrons. They collide with a neutral gas, in this case xenon, knocking off more electrons and creating positively charged xenon ions. Other fuels could be used—krypton gives off a greenish glow, neon glows red. Xenon is popular because its electrons orbit farther from their nuclei, and that makes them easier to bump.

In 1998, a NASA technology demonstration mission called Deep Space 1 used a xenon engine; solar panels instead of a nuclear reactor supplied electricity. Although ion propulsion had already flown on U.S. commercial satellites and dozens of Russian military satellites, its use on DS1 was the first time it was included on a spacecraft dispatched beyond Earth orbit—in this case, to a comet and asteroid. The ion drive worked like a champ. The test proved that the engine could be throttled up or down, that its exhaust would not corrupt scientific readings, and that the ions wouldn’t short out electronics or block radio signals.

But DS1’s engine was not a powerful one, even by ion thruster standards. And although it ran for 678 days and was still going when NASA ended the mission in 2001, that wasn’t long enough to demonstrate the years-long operation required for the JIMO mission.

So, says Mike Patterson, who co-built the DS1 engine with fellow Glenn engineer Bob Roman, “when somebody asks, ‘Why are you still working on ion thrusters—I thought you flew on DS1?’ I find that laughable. It’s like saying, ‘You’re still working on chemical rockets? I thought Robert Goddard did that in 1927.’ We’re in the infancy.”

Patterson is working on the Next Generation Ion Propulsion System, a larger, more powerful, and more fuel-efficient version of the DS1 engine. That thruster was about 12 inches wide and operated at 2.3 kilowatts. The new engine will be about 16 inches wide, consume up to seven kilowatts, and perform with 28 percent greater efficiency. Patterson’s goal is to build on the DS1 work without leaping too far, too fast. One major reason is that NASA doesn’t have the budget it did in the glory days of its youth. “Back in the ’60s, when the guys were working on the [five-foot-wide engine], I suspect they thought we were probably going to go to Mars by 1975 or something like that,” he says. “It just didn’t happen.”

The greatest challenge facing Patterson’s team is proving that an ion engine can operate for 10 years. “That’s 88,000 hours of operation,” he says. “If you look at your standard automobile engine, your car only lasts about 2,000 hours. And you’re constantly maintaining it. These we can’t maintain.”

No one knows how long an ion engine can last. At NASA’s Jet Propulsion Laboratory in California, engineers continue to run an identical flight spare of the DS1 engine in a test chamber. As of January, it had operated for 27,000 hours and consumed over 430 pounds of xenon, says DS1 program manager Marc Rayman. NASA engineers were debating how long to keep the test going. “If you run it to failure, you may destroy evidence to say ‘This is the rate at which it erodes,’ ” Al Newhouse says.

Designers of traditional chemical-fuel rockets have the luxury of firing test engines for one and a half times the duration they will eventually operate in space. That isn’t possible for an ion engine intended to run for 10 years. So finding a cost-effective and accurate way to predict lifetime without firing thrusters for their full duty cycle is critical. “If we make a mistake we won’t see for six or eight years, then we’ll be six or eight years behind,” Patterson says. And if it happened during the actual mission, JIMO could be lost in space. The team is working on a system that uses lasers to measure concentrations of particles as they sputter off the electrodes during short tests. From that data, engineers would extrapolate the expected engine lifetime.

Patterson’s team also hopes to boost the power of its engine beyond that of the DS1 thrusters. In the smaller engine, ions were shot through a circular molybdenum grid that looked a little like the screen filter in a kitchen faucet. To make a more powerful engine, Patterson can’t simply shoot more ions through a grid of the same size and material—the molybdenum would erode too quickly. So he is experimenting with carbon graphite grids that are more resistant to so-called sputter erosion. A DS1-type engine built with graphite grids might erode seven times slower, says Patterson. Engineers would then have the choice of running the engine longer at the same power, or working it harder.

In all likelihood, a JIMO-type mission would be powered by a cluster of ion engines that would take full advantage of the 100 kilowatts of power produced by the nuclear reactor. One option would be to line them up in an array along a boom. “Is it going to be three thrusters, five thrusters, or 10 thrusters?” Jankovsky asks. At the moment no one knows, because no one knows for sure how much thrust a single ion engine will be able to produce.

But there’s little doubt that in terms of mission design, the combination of nuclear power and ion propulsion offers a sharp departure from the past. Prometheus and JIMO could open up whole new types of missions for exploring the solar system. “For example, if one were to go to Saturn with a follow-on to Cassini, you could envision the possibility of [placing] landers down on the solid parts of [Saturn’s moon] Titan,” Greeley says. Multiple atmospheric probes might plunge into Saturn or Neptune to assess the planet’s chemical makeup. All this would be possible because nuclear electric propulsion lets mission designers devote more of their budgeted mass to scientific instruments and less to fuel.

Designers could also forgo gravity-assist fly-bys to gain additional delta-V. The way it works now, scientists have to wait for the planets to get in a certain alignment before launching to the outer solar system via Jupiter. A delay on the ground can result in the narrow launch window being missed, and thus a wait for months or sometimes even years for the next alignment.

Beyond the maneuverability of the spacecraft, the extra power from nuclear reactors offers other advantages. “We want to have long-lived landers on the polar areas of Mars,” Greeley says. Scientists suspect that if Mars has traces of life, past or present, they might be found at high latitudes. Exploring these areas hasn’t been possible because they are in shadow much of the time, which rules out solar power. Scientists also would like to drill into the Martian ice caps, but that too requires extra power, says Greeley.

At Jupiter, JIMO will take the study of Europa and the other icy moons to a new level. Galileo’s instruments were passive: They soaked up whatever feeble light and other radiation was reflected from the moons and converted them into images. With JIMO, scientists will be able to beam powerful radar signals at the moons, and the returning signals will be used to generate pictures, measure the altitude of varous features, and even see beneath the ice. Another idea is to melt the ice with a laser, making it possible to determine its chemical makeup. “That is the breakthrough—to get to active sensing of the outer planets,” says Ray Taylor, NASA’s overall system engineer for Project Prometheus and another Navy nuclear propulsion veteran.

In February, JIMO managers at JPL briefed U.S. spacecraft manufacturers on the conceptual design for the mission. NASA then awarded contracts to Lockheed Martin, Boeing, and Northrop Grumman to have them investigate designs for the Jupiter spacecraft. Newhouse wants the contractors to feel free to brainstorm. He warns against assuming that thrusters based on the DS1 technology will be the only answer. Another proposed ion thruster developed in Russia, for example, uses as its fuel the metallic element bismuth, which can be stored as a solid. At this point almost any propulsion design is still on the table.

For that reason, Newhouse says it would be foolhardy now to attempt a cost prediction for JIMO. “Come back in two years and I’ll tell you the cost,” he says. Everyone agrees that the technology will cost billions, however. “This could be the biggest procurement since the space station in terms of dollar value, so we have to do it right,” Newhouse says. It’s even possible, he concedes, that the contractors will tell him he’s asking for the impossible. Then again, that kind of advice would not have deterred Prometheus.

The glow of success: NASA has already flown 12-inch ion engines. Ions shoot out the holes in a circular grid, producing a small but steady thrust. NASA Glenn Research Center

Get the latest stories in your inbox every weekday.