Corp.) in 1953 and quickly found application in Antarctica to power scientific research stations (Jordan and Birden, 1954; Morse, 1963). SNAP-1 (an RTG) was built at the Mound Laboratory under AEC supervision in 1954 (Anderson and Featherstone, 1960). This was followed by the use of nuclear power systems on spacecraft in the early 1960s.

The possibilities of space nuclear power first entered the public sphere in January 1959 when President Dwight D. Eisenhower posed with a SNAP-3 RTG in the Oval Office of the White House. Ultimately, the Transit 4A and 4B navigation satellites were provided with SNAP-3B power sources from the AEC. They were the first satellites to operate in space with RPSs. Both satellites were also equipped with solar panels that supplied 35 W of power (Dassoulas and McNutt, 2007). These and subsequent missions proved the feasibility of using RPSs for space missions.


Space nuclear power reactors are another potential option for missions where solar power is not practical. However, the United States has launched only one space nuclear power reactor (SNAP-10A), and that took place in 1965. That early system was designed to produce 40 kW of thermal power and 500 W of electricity for an operating life of just 1 year, and the failure of a voltage regulator caused the system to shut down after 43 days (Wilson et al., 1965).

Beginning in 1983, NASA, the DOD, and the Department of Energy invested approximately $500 million in the SP-100 space nuclear power reactor. This system was intended to generated 2 MW of thermal power and 100 kW of electricity, but because of high costs, schedule delays, and changing national space mission priorities, the SP-100 program was suspended in the early 1990s and later canceled. The Soviet Union launched dozens of short-lived space nuclear power reactors during the 1970s and 1980s, and several unfueled Soviet systems were purchased by the United States in the early 1990s. These systems were extensively ground tested by a joint team of U.S., British, French, and Russian engineers using electrical heaters in place of the nuclear cores. Although the test program was successful, the United States did not use the Soviet equipment or technology in a flight program (NRC, 2006).

Project Prometheus was the most recent U.S. attempt to develop space nuclear power reactors. This project began in 2002, and it’s initial focus was on the Jupiter Icy Moons Orbiter mission. The project selected a nuclear electric propulsion reactor concept that was scalable from 20 kWe to 300 kWe. A nuclear electric propulsion system for a deep-space mission would need to be validated for reliable operation for a mission lifetime of 10 to 20 years, with no maintenance or repair. However, as with the SP-100 program, Project Prometheus did not proceed to the point of demonstrating the ability of system designs or available technology to meet required performance or lifetime specifications. Instead, it was terminated in 2005, after it became clear that it would have cost at least $4 billion to complete development of a spacecraft reactor module, and a total of at least $16 billion to develop the entire spacecraft and complete the mission, not counting the cost of the launch vehicle or any financial reserves to cover unexpected cost growth (JPL, 2005).

The performance and reliability of space nuclear power reactor systems using current technology remains unproven, especially for missions with long lifetimes. In addition, the committee is not aware of any substantive effort currently under way anywhere in the world to develop space nuclear power reactor systems. The history of space nuclear power reactors suggests that space nuclear reactors, if successfully developed, could meet the needs of some missions and could enable other missions that are not now under consideration because of power limitations. However, history also shows that the development of high-power, long-life space nuclear power reactors would be very time-consuming and expensive.


Three U.S. spacecraft with RPSs on board have inadvertently returned to Earth. In all cases, the RPSs performed as designed; the cause of the mission failure lay with other, nonnuclear systems.

The Transit 5BN-3 spacecraft with one SNAP-9A RPS on board broke up and burned up on reentry after a launch-vehicle upper stage failure. The design philosophy at that time was to require that the 238Pu oxide fuel totally burn up during reentry into Earth’s atmosphere, which it did.

As a result of that accident, the RPS design philosophy was changed to require full containment of the fuel (i.e., no fuel burn up) during an inadvertent reentry from or to Earth orbit. This design philosophy is still in effect.

The Nimbus B-1 weather satellite, the first NASA satellite to use an RPS, was intentionally destroyed during launch due to the erratic ascent of the launch vehicle. The launch vehicle, upper stage, and payload were totally destroyed by the explosion initiated by the destruct action, and the debris fell into the Santa Barbara Channel off Vandenberg Air Force Base. The two SNAP-19B2 RPSs were recovered intact (i.e., no 238Pu oxide fuel release occurred), and the fuel was used on a later mission.

The last accident involving a U.S. RPS was the Apollo 13 mission, which has been well documented. The SNAP-27 heat source assembly was stowed in the Lunar Excursion Module, which returned to Earth after the mission was aborted. It reentered over the South Pacific Ocean. Air and water sampling detected no 238Pu oxide fuel, indicating that the SNAP-27 heat source assembly survived reentry intact (as designed) and came to rest at the bottom of the Tonga Trench under more than 7,000 feet of water, where it still remains.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement