NUCLEAR ROCKETS: To Mars and Beyond

Nuclear Rockets: Then and Now

In 1961, President John F. Kennedy in his address to Congress outlined a new and bold space program. What many Americans remember is Kennedy's national goal of "landing a man on the moon and returning him safely to Earth" by the end of the decade. Few Americans remember that Kennedy also outlined an effort to go beyond the moon, perhaps to Mars and beyond.

In this issue of National Security Science, we tell the story of this lesser-known effort, one in which scientists at Los Alamos National Laboratory successfully built and tested a variety of nuclear rockets. Although the program officially ended in 1972, research to further improve the basic design of nuclear rockets has continued in other organizations, with current designs based on the Pewee-2 engine (1969–1972) now having specific impulses of 925 seconds.

In 1961, President John F. Kennedy in his address to Congress declared a national goal of "landing a man on the moon and returning him safely to Earth" by the end of the decade. From 1961 to 1975, America's space program used Apollo spacecraft and Saturn rockets to explore the moon, establish the Skylab program, and support a joint United States–Soviet Union mission in 1975. Saturn rockets (Fig. 1) were chemically based, making them huge—the Saturn V rocket stood 111 meters (363 feet) tall. Fully fueled, the rocket had a total mass of 3,000 metric tons (6.5 million pounds).

Having missions successfully reach the moon, the National Aeronautics and Space Administration (NASA) scientists and engineers set their sights on Mars and beyond. Their goal was to develop the technology to visit such faraway places, and Los Alamos would play a key role.

Although chemical rockets took astronauts to the moon and could take them to Mars, there are many drawbacks to the technology. For example, chemical engines produce relatively little power, making astronauts rely on planetary alignments, or "launch windows," to provide an extra gravitational slingshot effect that helps catapult space vehicles into space. Moreover, chemical rockets are slow, making long trips to places like Mars impractical for manned missions.

A more feasible technology is nuclear propulsion. Nuclear rockets are more fuel efficient and much lighter than chemical rockets. As a result, nuclear rockets travel twice as fast as chemical-driven spacecraft. Thus, a nuclear rocket could make a trip to Mars in as little as four months, and a trip to Saturn in as little as three years (as opposed to seven years). Such condensed trip times would help reduce astronaut and instrument exposure to harmful radiation emitted from the cosmic rays and solar winds that permeate interplanetary space.


Fig. 1. The Apollo 11 lifts off. This Saturn V rocket was the first to land humans on the moon. NASA successfully launched 13 Saturn V rockets with no loss of crew or payload. This rocket remains one of the largest and most powerful launch vehicles ever brought to operational status (from the standpoint of height, weight, and payload). Image courtesy of NASA.

The concern with nuclear rocketry lies in the radioactive components of nuclear power and these inherent safety challenges. Such concern has discouraged, and even prevented, space programs from implementing nuclear-powered missions to Mars and beyond.

The Atomic Age

In 1945, the United States ushered in the Atomic Age by detonating two atomic weapons over Japan, thus hastening the end of World War II. For approximately four years only the United States possessed nuclear weapons, but in 1949 the Soviet Union successfully tested its own nuclear bomb, bringing about the beginning of what would eventually become the Cold War.

One of the principal problems related to early atomic weapons consisted of their delivery systems. These early atomic weapons weighed about five tons each, making it impossible for aircraft to carry them over intercontinental distances. Indeed, the B-29s that dropped the first weapons over Hiroshima and Nagasaki flew only a few hundred miles from Tinian Island to reach their targets in Japan. The radioactive components for the atomic weapons reached Tinian by ship.

To solve the problem of delivering nuclear weapons thousands of miles, the United States began to develop heavy-lift, long-range weapons-delivery systems. These systems included strategic bombers such as the Convair B-36 "Peacemaker" (Fig. 2), ground-based rockets, and nuclear-powered aircraft. The unique characteristics of these systems provide options for decision makers. Rockets provide prompt response (approximately 30 minutes) and superior accuracy. Whereas aircraft can be launched to demonstrate resolve and subsequently be recalled as events warrant. Another competing design was the nuclear ramjet, although this design did not come into being until the 1960s.


Fig. 2. Built by Convair, the B-36 "Peacemaker" was the first bomber capable of delivering all the nuclear weapons in the United States arsenal. Image courtesy of the United States Air Force.

Project Rover

From 1955 through 1972, Los Alamos Scientific Laboratory (as the Laboratory was known then) conducted Project Rover, a program whose goal was to develop the technology for a nuclear-thermal rocket for space applications. Project Rover was part of the NASA space program, with the nuclear reactor portion falling under the Atomic Energy Commission (AEC).

The genesis of Project Rover can be traced to 1942, when scientists began to address the idea of using nuclear energy to propel an aircraft or rocket. These ideas were formulated soon after Enrico Fermi and his associates conducted the first successful test of a fission reactor. As early as 1944, scientists at the University of Chicago's Metallurgical Laboratory and Los Alamos began to discuss the possibility of using a fission reactor to heat a gas to high temperatures and propel a rocket—the basic idea behind a "nuclear-thermal" rocket. These scientists published several reports that explored the potential for this type of rocket.

The reports attracted the attention of the United States Air Force, which funded secret, small-scale, studies of nuclear-thermal rockets at Oak Ridge, Tennessee, from 1947 to 1949. Interest in nuclear rockets waned until 1954, when Robert Bussard of Oak Ridge Laboratory (now Oak Ridge National Laboratory) published a detailed engineering study of several nuclear-thermal-rocket applications. The Air Force commissioned Los Alamos Scientific Laboratory, Lawrence Radiation Laboratory at Livermore (now Lawrence Livermore National Laboratory), and others to perform more theoretical studies of nuclear-thermal-rocket performance.

Early in 1955, Los Alamos summarized its investigations in a now-declassified report, on nuclear-powered second stages for intercontinental ballistic missiles. The report touted the performance advantages of nuclear-powered rockets, which garnered support for materials research in the field. This report and similar information from Lawrence Livermore led to the start of serious nuclear-rocket-reactor work at both laboratories.

In 1956, Lawrence Livermore was redirected to work on nuclear ramjets, with Los Alamos continuing to develop rockets under Project Rover. As atomic weapons became smaller and lighter, chemical rockets became a viable delivery system. In 1957, the Air Force stated that nuclear rockets no longer had any military value and recommended that space applications be pursued for them instead.

Project Rover consisted of three principal phases: Kiwi (1955 to 1964), Phoebus (1964 to 1969), and Pewee (1969 to the project's cancellation, at the end of 1972). Nuclear reactors for the Project Rover were assembled at Los Alamos' Pajarito Site. For each engine there were actually two reactors built, one for "zero-power critical" experiments conducted at Los Alamos and another used for full-power testing at the former Nevada Test Site (now the Nevada National Security Site). Fuel and internal engine components for the engines were fabricated in the Sigma complex at Los Alamos. Figures 3 and 4 show some of the reactors. The Project Rover illustration provides additional information for each phase.


Fig. 3. Robert Hanrahan stands between the Kiwi-A (left) and Phoebus-1 (right). Phoebus-1 is now at the Atomic Testing Museum in Las Vegas, Nevada. Hanrahan is now director of the Office of Nuclear Experiments for the National Nuclear Security Administration. Archival photo courtesy of Robert Hanrahan.


Kiwi a Test fire photo

Fig. 2. Built by Convair, the B-36 "Peacemaker" was the first bomber capable of delivering all the nuclear weapons in the United States arsenal. Image courtesy of the United States Air Force.

NERVA: Nuclear Engine for Rocket Vehicle Application

In 1961, NASA and the AEC embarked on a second nuclear-rocket program known as NERVA. Taking advantage of the knowledge acquired as scientists designed, built, and tested Project Rover research reactors, NERVA scientists and engineers worked to develop practical rocket engines that could survive the shock and vibration of a space launch. From 1964 to 1969, Westinghouse Electric Corporation and Aerojet-General Corporation built various NERVA reactors and rocket engines.

In 1969, NERVA's successes prompted NASA-Marshall Space Flight Center director Wernher von Braun to propose sending 12 men to Mars aboard two rockets, each propelled by three NERVA engines (Fig. 5). The mission would launch in November 1981 and land on Mars in August 1982.

conceptual illustration of a spacecraf

Fig. 5. A conceptual illustration of a spacecraft for a manned Mars mission proposed by NASA's Wernher von Braun in August 1969. Two spacecraft would make the trip in tandem, each one powered by three NERVA-type engines.
Image courtesy of NASA.

Although the mission never took place, engines tested during that time met nearly all of NASA's specifications, including those related to thrust, thrust-to-weight ratio, specific impulse, engine restart, and engine lifetime. When the Project Rover/NERVA program was canceled in 1972, the only major untested requirement was that a NERVA rocket engine should be able to restart 60 times and operate for a total of 10 hours.

There was one engine, however, that exceeded some NERVA specifications. Designed, built, and tested at Los Alamos, the Phoebus-2A Project Rover engine (Fig. 6) produced up to 4,000 megawatts of thermal power. In those terms, it was the most powerful nuclear propulsion reactor ever built.


Fig. 6. The Phoebus-2A being prepared for testing at Los Alamos Scientific Laboratory. Photo courtesy of Richard Malenfant.

Nuclear-Thermal Reactors

During the Project Rover/NERVA projects, scientists conducted 22 major tests of nuclear-thermal-rocket engines (Fig. 7). Many of these tests explored potential solutions to complex problems that arise when using reactors to propel rockets with hot hydrogen. Significant issues with materials stability, compatibility, and corrosion beyond those encountered in terrestrial power reactors had to be addressed to produce practical rockets.

Kiwi-B4A reactor awaits testing at the Nevada Test Site

Fig. 7. The Kiwi-B4A reactor awaits testing at the Nevada Test Site in 1962. At the left is Norris Bradbury, Laboratory Director from 1940–1970.

The principal difference between reactors used for space propulsion and electricity generation is the temperature of the cores. A reactor core consists of (1) fuel elements that contain the radioactive material to produce fission; (2) structures designed to hold the fuel elements in place; (3) structures that control the reactor's operation by absorbing, reflecting, or slowing ("moderating") neutrons produced by the fission reactions; and (4) a cooling system. The cooling system or "working fluid" (a gas or liquid) absorbs heat produced by the fuel elements and transfers the heat or energy to other parts of the system to generate propulsion or electricity.

Figure 8 shows a simplified schematic of a nuclear power reactor. The schematic shows uranium fuel elements, which cause fission reactions. The radiation-protection barrier limits radiation exposure to plant workers and the environment.


Fig. 8. This general schematic shows the major nuclear reactor components. Uranium fuel elements produce fission reactions and neutrons. The neutrons help produce more fission reactions, leading to a sustained chain reaction. Operators move control rods in and out of the reactor to absorb neutrons, and thus control the chain reaction. This controls the power at which the reactor operates. The moderator improves reactor efficiency by slowing down the neutrons, thus increasing their likelihood of producing fission reactions. The reactor heats the water, which passes through the core, thus generating steam. The steam turns an electrical generator, thus producing electricity. Original figure courtesy of the European Nuclear Society; redrawn by Los Alamos National Laboratory to enhance clarity.

In a nuclear-thermal-rocket reactor, the temperature must be as high as possible to achieve optimum performance (see the article "The Basics of Nuclear Rocketry" on page 25). Thus, the core temperature for the Kiwi-A test was 2,683 Kelvin (K) or 4,370°F, whereas the core of pressurized- water reactors used for nuclear power plants is only around 600 K (620.6°F).

A major difference between a nuclear-thermal-rocket reactor and a power-plant reactor is the cooling systems. Nuclear rockets use hydrogen, whereas U.S. power plants use water. Hydrogen is the best propellant gas for a nuclear-thermal rocket. Nevertheless, working with hydrogen at these high temperatures presents many challenges.

It Comes Down to the Cores

All the Project Rover/NERVA reactors had solid cores. As detailed in "The Basics of Nuclear Rocketry" on page 25, researchers designed liquid- and gas-core reactors for nuclear-thermal-rocket propulsion—and even conducted small-scale experiments on components for these designs, but only solid-core reactors were built. Few materials, however, remain solid at the temperatures in the core of a nuclear-thermal-rocket reactor. Structurally, several metals and ceramics with high melting points (so-called "refractory" materials) may be used to build a core, but the way these materials interact with neutrons also plays a key role in their selection.

For example, the metal with the highest melting point—tungsten, at 3,695 K (6,191.6°F)—strongly absorbs neutrons, particularly "slow" ones, which have energies much less than one electronvolt. Project Rover cores were capable of operating with tungsten as a fuel matrix. However, the development of tungsten required technology development that proved to be beyond the capabilities of the program at the time. Consequently, Project Rover focused on graphite core reactors.

As a crystalline form of carbon, graphite behaves well at high temperatures because it has the highest melting point of any element. Graphite not only retains its strength at high temperatures but also actually becomes stronger. Graphite has long been used in various high-temperature industrial applications; therefore, scientists began considering its use in the design of Project Rover reactors.

The first fission reactor—and many reactors built after it—consisted of graphite bricks stacked in a "pile," with rods of uranium dispersed throughout. Graphite was chosen to build piles mainly because of its good neutronics properties and because it was a weak neutron absorber and a good reflector and moderator of neutrons. However, graphite in a simple pile never encounters the extreme conditions as it would in the core of a nuclear-rocket reactor. Graphite's response to these extreme conditions was unknown.

Hot Hydrogen Complicates Things

Scientists suspected early on that graphite could pose a serious problem when used in a nuclear-thermal-rocket reactor. The best propellant gas for this type of rocket is hydrogen. In the Project Rover engines, large amounts of hydrogen passed over the rocket nozzle and some reactor components before being forced through channels in fuel elements within the reactor core. This process heated the hydrogen, and the hot hydrogen quickly corroded the graphite in the reactor.

To protect these channels, scientists coated their inner surface with a thin niobium carbide (NbC) film. At first, scientists used a gaseous mixture of niobium chloride (NbCl5), hydrogen chloride (HCl), and hydrogen (H) to deposit NbC onto the channel's inner surfaces, using a process called chemical vapor deposition (CVD). However, as core designs evolved, the lengths of the fuel elements increased. To meet these longer fuel-element designs, researchers vertically stacked shorter fuel-element sections to make longer fuel elements. Eventually, researchers perfected the technique for fabricating longer fuel elements.

The CVD process could not deeply penetrate the longer channels to coat evenly the inner surfaces of the fuel elements. Therefore, researchers developed a new coating method. Thin niobium tubes were inserted into the channels and heated in place under a hydrogen chloride gas, which converted the Nb to NbC. The outer surfaces of the fuel elements were also coated. Later, the program evolved to zirconium carbide (ZrC) coatings

Fueling the Reactor

The fuel elements in the cores of the Project Rover/NERVA reactors consisted of uranium-loaded graphite, made using a new method developed by Haskell Sheinberg, who is now a retired Los Alamos National Laboratory Fellow. Unlike the proven high-temperature method for making graphite parts, the new method worked at room temperature, thus making it easier to fabricate fuel. Also, the new method produced stronger, denser graphite than the traditional method. Sheinberg's method is still in use today.

The fuel elements in the Kiwi-A core consisted of flat plates molded and pressed at room temperature from a mixture of fine graphite powder called graphite flour, fine carbon powder, graphite flakes, a resin binder, and uranium oxide (UO2) particles. The fuel elements underwent baking from 318 K (113ºF) to 453 K (356ºF) over a period of 36 hours, followed by a heat treatment under vacuum at 1,073 K (1,472ºF). During this two-stage baking process, the resin decomposed into amorphous carbon and gas. The final baking of the elements was at 2,723 K (4,442°F) to crystallize the amorphous carbon into graphite. Graphite is stronger and a better heat conductor than amorphous carbon, which is brittle.

During this graphitization process, the UO2 particles are thermally converted to uranium carbide (UC2) particles. After graphitization, the fuel elements were machined to fine tolerances, as required for reactor operation. The flat-plate fuel elements in the Kiwi-A core were the only fuel elements used during Project Rover/NERVA that were not clad or coated to reduce hydrogen corrosion.

The fuel elements for all the Kiwi reactors after Kiwi-A (except the last one, Kiwi-B4E) were extruded into their near-final shapes and dimensions from a "green mix"—which consisted of graphite flour, fine carbon powder, resin binder (partially polymerized—or "set" furfuryl alcohol), UO2 particles, and a catalyst (maleic anhydride). After extrusion, the elements were baked to 523 K (482ºF) for approximately 56 hours to polymerize the resin. Then they were heated to as high as 1,123 K (1,562ºF) to remove gases produced during polymerization. Finally, the elements were graphitized around 3,000 K, and then machined to final specifications.

The extruded fuel elements had a diameter of approximately one inch, and were approximately four feet in length. In the early reactors, extruded fuel elements consisted of cylinders with a circular cross-section. Later, the cylinders had a hexagonal cross-section, so the assembled core resembled a honeycomb. The number of channels in each fuel element and the channel diameters also changed as reactor designs evolved. Figure 9 shows a cross-sectional schematic of a typical Project Rover/NERVA reactor.


Fig. 8. This general schematic shows the major nuclear reactor components. Uranium fuel elements produce fission reactions and neutrons. The neutrons help produce more fission reactions, leading to a sustained chain reaction. Operators move control rods in and out of the reactor to absorb neutrons, and thus control the chain reaction. This controls the power at which the reactor operates. The moderator improves reactor efficiency by slowing down the neutrons, thus increasing their likelihood of producing fission reactions. The reactor heats the water, which passes through the core, thus generating steam. The steam turns an electrical generator, thus producing electricity. Original figure courtesy of the European Nuclear Society; redrawn by Los Alamos National Laboratory to enhance clarity.

In early 1962, Project Rover scientists and engineers encountered a "back-reaction" problem with UO2 particles in the fuel elements that had been converted, during graphitization, to UC2 particles. Uranium carbide particles are extremely reactive and revert back to oxide in air, particularly humid air. During fuel-element processing, storage, and during reactor operation, the back reaction generated carbon dioxide (CO2) gas and degraded the fuel elements. Oxidation of the UC2 particles during storage of the fuel elements made them swell as much as 4%, so that their final dimensions exceeded the fine tolerances required for the reactor.

To resolve this problem, scientists replaced UO2 particles with pyrolytic-graphite-coated UC2 particles. Eventually, uranium carbide particles were developed that could withstand 2,873 K (4,550°F) for 30 minutes. The Kiwi-B4E, Phoebus-1, Phoebus-2, Pewee, and NRX-A reactors/engines all used fuel elements containing pyrolytic-graphite-coated UC2 particles. This type of fuel element became the "standard" fuel element. The NRX-A series of nuclear reactors were developed to demonstrate that Kiwi-B4 reactors could be adapted to withstand launch loads. During NRX-A6 and Pewee tests, the standard fuel elements operated for one hour at hydrogen exhaust temperatures between 2,400 K and 2,600 K (3,860.6–4,220.6°F).

Meanwhile, Los Alamos researchers continued to develop fuel elements capable of performing at ever-higher core temperatures. In 1972, Los Alamos tested two new fuel-element concepts in their Nuclear Furnace (NF-1) reactor, specifically designed to test such concepts. The NF-1 reactor was a heterogeneous, water-moderated, beryllium-reflected reactor for performing high-temperature nuclear tests. The first fuel element was a pure carbide (U,Zr)C. The second element was a "composite" fuel element.

Both the standard and composite fuel elements performed well when tested in the NF-1 reactor for 109 minutes at 2,450 K (3,950.6°F). Projections at that time indicated that the composite fuel elements would be good for 2 to 6 hours at 2,500–2,800 K (4,040.6–4,580.6°F). The researchers achieved similar endurance times at 3,000–3,200 K (4,940.6–5,300.6°F) for the carbide fuel elements, once an improved cross-section design reduced a cracking problem. For 10 hours of operation, fuel elements were limited to hydrogen exhaust temperatures of 2,200–2,300 K (3,500.6–3,680.6°F) for standard elements, nearly 2,400 K (3,860.6°F) for composite, and approximately 3,000 K (4,940.6°F) for pure-carbide fuel elements.

Mission: Cancelled

NASA's plans for NERVA included a visit to Mars in 1979 and a permanent lunar base by 1981. NERVA rockets would be used for nuclear "tugs" designed to take payloads from low-Earth orbit to higher, larger orbits as a component of the later-named Space Transportation System. The NERVA rocket would also be used as a nuclear-powered upper-stage component for the Saturn rocket (a chemical-based rocket), which would enable the upgraded Saturn engine to launch much larger payloads (up to 340,000 pounds) to low-Earth orbit.

In 1973, Project Rover/NERVA was cancelled. Although the projects proved very successful, the space mission itself never took place. No nuclear-thermal rockets were ever used to send explorers on long-range space missions.

It was the Mars mission that led to NERVA's termination. Members of Congress judged the manned mission to Mars was too expensive and that funding the project would continue to foster a costly "space race" between the United States and the Soviet Union.

By the time NERVA was cancelled, the NERVA-2 would have met all the mission's objectives. Two of these engines would have been fitted to a NERVA stage capable of powering a manned interplanetary spacecraft.

During its lifetime, Project Rover/NERVA achieved the following records:

  • 4,500 megawatts of thermal power
  • 3,311 K (5,500.4°F) exhaust temperature
  • 250,000 pounds of thrust
  • 850 seconds of specific impulse
  • 90 minutes of burn time
  • thrust-to-weight ratios of 3 to 4

Beyond proving the feasibility of nuclear space propulsion, Project Rover/NERVA enabled scientists to produce approximately 100 technical papers that covered the properties of graphite, graphite flour, and other forms of carbon. The program also produced several important spin-offs, including Sheinberg's room-temperature graphite-fabrication process and methods for coating graphite with thin films of metal carbides.

Moreover, the technology for coating UC2 particles with pyrolytic graphite eventually led to the TRISO fuel beads now used in commercial high-temperature, gas-cooled reactors to generate electricity. However, the program's most important spin-off—by any measure—was the heat pipe (see the " Inspired Heat-Pipe Technology" article).

The heat pipe is currently the centerpiece of the Los Alamos research program known as Heatpipe Power System (HPS) reactors. As envisioned by heat pipe inventor Los Alamos physicist George Grover, HPS reactors use heat pipes to transfer heat from a reactor core to thermoelectric elements or heat engines.

In 2000, NASA created Project Prometheus to develop nuclear-powered systems for long-duration space missions. This project was NASA's most serious consideration of nuclear power for space missions since the cancellation of Project Rover/NERVA in 1972. For the Jupiter Icy Moons Orbiter (JIMO), a spacecraft designed to explore Europa, Ganymede, and Callisto, NASA intended to use an HPS reactor. The JIMO (Fig. 10) design used a fission reactor to power a Brayton-cycle heat engine that ran an electrical generator. The electricity would then power scientific instruments and an ion-propulsion unit. In 2005, NASA canceled the Prometheus Project as a result of budget constraints.

An artist's conception of JIMO

Fig. 10. An artist's conception of JIMO as it nears Jupiter and its icy moons: Europa, Ganymede, and Callisto. Illustration courtesy of NASA.

Because scientists continue to look back and build upon the technical advances developed during Project Rover/NERVA, those enduring advances are likely to one day play a major role in humanity's exploration of the solar system and beyond.

-Brian Fishbine, Robert Hanrahan, Steven Howe, Richard Malenfant, Carolynn Scherer, Haskell Sheinberg, and Octavio Ramos Jr.

In this issue...

Contact Us | Careers | Bradbury Science Museum | Emergencies | Inside LANL | Maps | Site Feedback | SSL Portal | Training

Managed by Triad National Security, LLC for the U.S. Department of Energy's NNSA © Copyright Triad National Security, LLC. All Rights Reserved. | Terms of Use | Privacy Policy