A nuclear thermal rocket (NTR) is a type of thermal rocket where the heat from a nuclear reaction replaces the chemical energy of the propellants in a chemical rocket. In an NTR, a working fluid, usually liquid hydrogen, is heated to a high temperature in a nuclear reactor and then expands through a rocket nozzle to create thrust. The external nuclear heat source theoretically allows a higher effective exhaust velocity and is expected to double or triple payload capacity compared to chemical propellants that store energy internally.
NTRs have been proposed as a spacecraft propulsion technology, with the earliest ground tests conducted in 1955. The United States maintained an NTR development program through 1973, when it was shut down for various reasons, including to focus on Space Shuttle development. Although more than ten reactors of varying power output have been built and tested, , no nuclear thermal rocket has flown.
Whereas all early applications for nuclear thermal rocket propulsion used fission processes, research in the 2010s has moved to fusion approaches. The Direct Fusion Drive project at the Princeton Plasma Physics Laboratory is one such example, although "energy-positive fusion has remained elusive". In 2019, the U.S. Congress approved US$125 million in development funding for nuclear thermal propulsion rockets.
In May 2022, DARPA issued a request for proposal (RFP) for the next phase of their Demonstration Rocket for Agile Cislunar Operations (DRACO) nuclear thermal engine program. This follows on their selection, in 2021, of an early engine design by General Atomics and two spacecraft concepts from Blue Origin and Lockheed Martin. The next phases of the program would focus on the design, development, fabrication, and assembly of a nuclear thermal rocket engine.DARPA moving forward with development of nuclear powered spacecraft. DARPA News, 4 May 2022. URL: https://www.darpa.mil/news-events/2022-05-04 In July 2023, Lockheed Martin was awarded the contract to build the spacecraft and BWX Technologies (BWXT) would have developed the nuclear reactor. A launch was expected in 2027,DARPA Kicks Off Design, Fabrication for DRACO Experimental NTR Vehicle. DARPA News, 26 July 2023. URL: https://www.darpa.mil/news-events/2023-07-26 but this was put on indefinite hold due to nuclear reactor test requirements, later compounded by proposed cuts in the FY2026 budget by the second Donald Trump administration before being cancelled.
In June 2025, the European Space Agency proposed its own NTP engine, called Alumni. At the same time, another form of nuclear thermal propulsion, called centrifugal nuclear thermal rocket, uses liquid uranium for fuel.
Principle of operation
Nuclear-powered thermal rockets are more effective than chemical thermal rockets, primarily because they can use low-molecular-mass propellants such as hydrogen.
As thermal rockets, nuclear thermal rockets work almost exactly like chemical rockets: a heat source releases thermal energy into a gaseous propellant inside the body of the engine, and a nozzle at one end acts as a very simple heat engine: it allows the propellant to expand away from the vehicle, carrying momentum with it and converting thermal energy to coherent kinetic energy. The specific impulse (Isp) of the engine is set by the speed of the exhaust stream. That, in turn, varies as the square root of the kinetic energy loaded on each unit mass of propellant. The kinetic energy per molecule of propellant is determined by the temperature of the heat source (whether it be a nuclear reactor or a chemical reaction). At any particular temperature, lightweight propellant molecules carry just as much kinetic energy as heavier propellant molecules and therefore have more kinetic energy per unit mass. This makes low-molecular-mass propellants more effective than high-molecular-mass propellants.
Because chemical rockets and nuclear rockets are made from refractory solid materials, they are both limited to operate below , by the strength characteristics of high-temperature metals. Chemical rockets use the most readily available propellant, which is waste products from the chemical reactions producing their heat energy. Most liquid-fueled chemical rockets use either hydrogen or hydrocarbon combustion, and the propellant is therefore mainly water (molecular mass 18) and carbon dioxide (molecular mass 44). Nuclear thermal rockets using gaseous hydrogen propellant (molecular mass 2) therefore have a theoretical maximum specific impulse that is 3 to 4.5 times greater than those of chemical rockets.
While liquid hydrogen provides high specific impulse, its density (about 71 kg/m³) is far lower than that of common chemical rocket fuels such as liquid methane (about 422 kg/m³) or RP-1 kerosene (about 810 kg/m³). Large propellant tanks are therefore required, adding structural mass that partially offsets the engine's efficiency advantage. Liquid hydrogen must also be kept at cryogenic temperatures, which requires heavy insulation or active cooling to limit boil-off, particularly on long missions. Nuclear electric propulsion, by contrast, uses electric thrusters that can operate on dense, easily stored propellants such as xenon.
Early history In 1944, Stanisław Ulam and Frederic de Hoffmann contemplated the idea of controlling the power of nuclear explosions to launch space vehicles. After World War II, the U.S. military started the development of intercontinental ballistic missiles (ICBM) based on the German V-2 rocket designs. Some large rockets were designed to carry nuclear warheads with nuclear-powered propulsion engines. As early as 1946, secret reports were prepared for the U.S. Air Force, as part of the NEPA project, by North American Aviation and Douglas Aircraft Company's Project Rand. These groundbreaking reports identified a reactor engine in which a working fluid of low molecular weight is heated using a nuclear reactor as the most promising form of nuclear propulsion but identified many technical issues that needed to be resolved.H. P. Yockey, T. F. Dixon (1 July 1946), "A Preliminary study on the Use of Nuclear Power in Rocket Missiles", Report NA-46-574R. Gomog (3 August 1946), "Rocket Computations", Report NEPA-508 L. A. Oblinger (13 August 1946), "Pilot Plant for Nuclear Powered Aircraft", Report NEEA-505L. A. Ohlinger (21 November 1946) "Controls for Nuclear Powered Aircraft", Report NEPA-511 Feasibility of Nuclear Powered Rockets and Ramjets, Report NA 47-15, February 1947"Nuclear-Powered Flight", LEXP-1, 30 September 1948
In January 1947, not aware of this classified research, engineers of the Applied Physics Laboratory published their research on nuclear power propulsion and their report was eventually classified.A. E. Ruark ed. (14 January 1947) "Nuclear Powered Flight", APL/JEU-TG-20 In May 1947, American-educated Chinese scientist Qian Xuesen presented his research on "thermal jets" powered by a porous graphite-moderated nuclear reactor at the Nuclear Science and Engineering Seminars LIV organized by the Massachusetts Institute of Technology.
In 1948 and 1949, physicist Leslie Shepherd and rocket scientist Val Cleaver produced a series of groundbreaking scientific papers that considered how nuclear technology might be applied to interplanetary travel. The papers examined both nuclear-thermal and nuclear-electric propulsion.
Early NASA engine development Through Project Rover, Los Alamos National Laboratory began developing nuclear thermal engines as soon as 1955 and tested the world's first experimental nuclear rocket engine, KIWI-A, in 1959. This work at Los Alamos was then continued through the NASA's NERVA program (1961–1973). NERVA achieved many successes and improved upon the early prototypes to create powerful engines that were several times more efficient than chemical counterparts. However, the program was cancelled in 1973 due to budget constraints. To date no nuclear thermal propulsion system has ever been implemented in space.
Nuclear fuel types A nuclear thermal rocket can be categorized by the type of reactor, ranging from a relatively simple solid reactor up to the much more difficult to construct but theoretically more efficient gas core reactor. As with all thermal rocket designs, the specific impulse produced is proportional to the square root of the temperature to which the working fluid (reaction mass) is heated. To extract maximum efficiency, the temperature must be as high as possible. For a given design, the temperature that can be attained is typically determined by the materials chosen for reactor structures, the nuclear fuel, and the fuel cladding. Erosion is also a concern, especially the loss of fuel and associated releases of radioactivity.
Solid core
Solid core nuclear reactors have been fueled by compounds of uranium that exist in solid phase under the conditions encountered and undergo nuclear fission to release energy. Flight reactors must be lightweight and capable of tolerating extremely high temperatures, as the only coolant available is the working fluid/propellant. A nuclear solid core engine is the simplest design to construct and is the concept used on all tested NTRs.Solid Core NTR Beyond Nerva. Retrieved 4 May 2022
Using hydrogen as a propellant, a solid core design would typically deliver specific impulses (Isp) on the order of 850 to 1000 seconds, which is about twice that of liquid hydrogen-oxygen designs such as the Space Shuttle main engine. Other propellants have also been proposed, such as ammonia, water, or LOX, but these propellants would provide reduced exhaust velocity and performance at a marginally reduced fuel cost. Yet another mark in favor of hydrogen is that at low pressures it begins to dissociate at about 1500 K, and at high pressures around 3000 K. This lowers the mass of the exhaust species, increasing Isp.
Early publications were doubtful of space applications for nuclear engines. In 1947, a complete nuclear reactor was so heavy that solid core nuclear thermal engines would be entirely unable to achieve a thrust-to-weight ratio of 1:1, which is needed to overcome the gravity of the Earth at launch.Alvarez, Luis, "There Is No Obvious Or Simple Way To Use Atomic Energy For Space Ships", U.S. Air Services, January 1947, pp. 9-12 Over the next twenty-five years, U.S. nuclear thermal rocket designs eventually reached thrust-to-weight ratios of approximately 7:1. This is still a much lower thrust-to-weight ratio than what is achievable with chemical rockets, which have thrust-to-weight ratios on the order of 70:1. Combined with the large tanks necessary for liquid hydrogen storage, this means that solid core nuclear thermal engines are best suited for use in outer space (outside the earth's lower atmosphere), and not for liftoff, not to mention avoiding the radioactive contamination that would result from atmospheric use (if an "open-cycle" design was used, as opposed to a lower-performance "closed cycle" design where no radioactive material was allowed to escape with the rocket propellant).
One way to increase the working temperature of the reactor is to change the nuclear fuel elements. This is the basis of the particle-bed reactor, which is fueled by several (typically spherical) elements that "float" inside the hydrogen working fluid. Spinning the entire engine could prevent the fuel element from being ejected out of the nozzle. This design is thought to be capable of increasing the specific impulse to about 1000 seconds (9.8 kN·s/kg) at the cost of increased complexity. Such a design could share design elements with a pebble-bed reactor, several of which are currently generating electricity. From 1987 through 1991, the Strategic Defense Initiative (SDI) Office funded Project Timberwind, a non-rotating nuclear thermal rocket based on particle bed technology. The project was canceled before testing.
Pulsed nuclear thermal rocket