Since before the beginning of the Space Age, engineers have sought to develop increasingly efficient propulsion systems. Chemical propulsion systems that burn a fuel and oxidizer to produce thrust were the first to be developed. With their high thrust-to-mass ratios (i.e. a small size engine can produce a large amount of thrust), liquid fueled chemical rockets were the first to allow us to overcome the bonds of gravity and pass the threshold into space.
The Nuclear Option
The most efficient chemical propulsion systems today burn liquid hydrogen and oxygen and have an Isp of up to about 450 seconds. Called “specific impulse”, Isp is a measure of the efficiency of a propulsion system. It can be thought of as the amount of thrust you get from a unit mass of propellant. For those who like to work in Imperial units, an Isp of 450 seconds, for example, one pound-mass (0.45 kilograms) of propellant yields a thrust of 450 pounds-force (2,000 Newtons) for one second. The Isp also gives an engine’s exhaust velocity when it is multiplied by the acceleration due to gravity. As every rocket scientist knows, a higher exhaust velocity translates proportionally to the faster a rocket of a given mass and propellant load will travel. Conversely, higher exhaust velocities can mean a larger payload for a given rocket.
Higher exhaust velocities can be achieved by increasing an engine’s operating temperature since the velocity of the exhaust products is proportional to the square root of temperature. But limitations in the strength of available materials used in an engine’s combustion chamber restricts how high these can go. The best of today’s chemical propulsion systems are already close to the theoretical maximum Isp. Use of the most energetic chemical propellant combination, liquid hydrogen and fluorine, could provide a modest increase in engine Isp. But the engineering difficulties of using dangerously reactive liquid fluorine offsets any performance advantages. Today, rocket engine developers are more concerned with maximizing the engine’s thrust-to-weight ratio and minimizing manufacturing costs. Significant new developments in engine efficiency lie elsewhere.
Another family of propulsion systems that offer significantly higher Isp are based on ion or plasma technology. Here electromagnetic fields are used to accelerate an ionized working fluid to very high velocity (see “The First Ion Engine Test in Space”). Although such systems can have an Isp in excess of thousands of seconds, they have minuscule thrust-to-mass ratios. With the addition of the mass of the power generation system required to run these engines, these systems are only capable of tiny rates of acceleration. While these propulsion systems do have their applications, those seeking high acceleration rates combined with high Isp have to look elsewhere.
A schematic of a nuclear thermal propulsion system. Click on image to enlarge. (LANL)
One of the most promising possibilities within the reach of our technology is nuclear thermal propulsion. Unlike a chemical rocket that uses combustion to heat the reactive mass (i.e. the combustion products, in this case) that are expelled to generate thrust, a nuclear rocket uses a nuclear reactor to superheat a lightweight propellant – ideally hydrogen. Although chemical and nuclear engines share similar engineering limitations in terms of operating temperatures and pressures, the much lower molecular weight of hydrogen compared with the combustion products of a hydrogen-oxygen rocket engine (which are largely water vapor) results in much higher exhaust velocities for a given engine temperature and pressure. This yields an Isp that can be on the order of 1,000 seconds. But can such an engine be built?
The Birth of Nuclear Rocketry
Not long after the first successful atomic bomb tests, scientists and engineers began to ponder the potential peaceful uses of this potent source of energy. As early as 1944, Stanislaus Ulam and Frederick de Hoffman at the Los Alamos Scientific Laboratory (LASL – today the Los Alamos National Laboratory) considered how nuclear detonations might be used for space travel. While such a scheme was later studied in detail as part of ARPA’s (Advanced Research Project Agency) Project Orion and the British Interplanetary Society’s Project Daedalus, it was felt that a slower, controlled release of nuclear energy would be more suitable.
In July of 1946, North American Aviation and the Douglas Aircraft Company’s Project RAND each delivered secret reports on their internal nuclear propulsion studies to the USAF. These landmark reports identified the “heat transfer” nuclear rocket (where a reactor heats a working fluid which acts as the reaction mass) as the most promising form of nuclear propulsion. Such a propulsion system could, in principle, be incorporated into an ICBM to lob nuclear warheads across the globe. But despite the glowing report and the promise of the technology, it was recognized that there were still many technical issues that needed to be resolved.
The American educated Chinese scientist named Hsue-Shen Tsien proposed a nuclear engine or “thermal jet” concept in 1948 during a lecture at MIT.
Not aware of the earlier secret studies, a group of engineers from the Applied Physics Laboratory at Johns Hopkins University openly published the results of their own independent studies in January 1947. In 1948 and 1949, two British space enthusiasts, A.V. Cleaver and L.R. Shepherd, also published a series of ground breaking papers in the Journal of the British Interplanetary Society on the same topic. But even before this series of papers was published, an American educated, Chinese scientist named Hsue-Shen Tsien (or Qian Xuesen, using the more modern Chinese transliteration, who later went on to head the Chinese atomic bomb program) gave a talk at the Massachusetts Institute of Technology about nuclear powered “thermal jets”. In all these studies, it was concluded that nuclear propulsion seemed to be viable. And given the number of people who independently arrived at the same conclusions, it was clear that the USAF would not have a monopoly in nuclear propulsion studies.
But all this early enthusiasm for nuclear rockets was dampened by a subsequent technical report done by North American Aviation. This report concluded that nuclear powered ICBMs were not practical. North American scientist felt that the reactor of a nuclear rocket would have to operate at the fantastically high temperature of 3,400 K – many times that of existing reactors. No known material could withstand such temperatures and maintain the strength required in a rocket engine. With this and other problems identified, interest in nuclear rockets faded noticeably as the 1950s began.
An Idea Resurrected
But not everyone agreed with the apparently bleak prospects for nuclear rockets. While development of nuclear rocket engines was largely abandoned after the North American report, work on nuclear-powered jet aircraft engines continued. In the early 1950s Robert W. Bussard who had been working on these nuclear aircraft propulsion systems at AEC’s (US Atomic Energy Commission) Oak Ridge National Laboratory in Tennessee reexamined nuclear rockets. Based on his work he concluded that the earlier reports were far too pessimistic and that nuclear rockets were probably practical after all. Bussard felt that they could effectively compete with chemical rockets especially on long flights with heavy payloads. Based on Bussard’s calculations and salesmanship, the USAF decided to reopen studies on the concept for possible use in ICBMs in 1955.
The work of Robert W. Bussard in the early-1950s showed that nuclear rocket propulsion was practical leading to the USAF starting work on the concept in 1955.
As part of the new AEC-USAF program, the Nuclear Propulsion Division headed by Raemer E. Schreiber was formed at LASL. A similar group was also formed at AEC’s Lawrence Radiation Laboratory operated by the University of California. But budget cutbacks in the June of 1956 resulted in an elimination of duplicate efforts and a consolidation of the various nuclear propulsion groups. The result was Livermore taking on the task of developing a nuclear ramjet under the code name “Project Pluto”. The nuclear rocket program went to Los Alamos under the code name “Project Rover”.
Raemer E. Schreiber giving a briefing on the Kiwi-A reactor in 1959. (LANL)
A series of different paper studies with such fanciful names like “Dumbo” (an engine reactor design) and “Condor” (a proposed nuclear rocket) were studied. Eventually a reactor design named “Kiwi” was selected as a first step for a nuclear rocket engine. Like its flightless namesake from New Zealand, the Kiwi test reactors would not fly but were nonetheless essential to the development of a practical nuclear rocket engine.
Kiwi-A was a series of “battleship” test reactors that would use compressed hydrogen gas to perform ground-based studies of potential nuclear rocket engine components. In the first Kiwi reactor, the 960 uranium oxide (UO2) infused graphite fuel plates (which was transformed into uranium carbide during manufacture) were stacked along with 240 plain graphite plates inside of a 43-centimeter thick, annular graphite reflector that could operate at temperatures as high as 3,000 K. Not only could the graphite withstand temperatures up to 3,300 K before beginning to weaken, it was also an excellent moderator that could slow fission-producing neutrons so they could maintain a nuclear chain reaction inside the core. The reactor core itself was 84 centimeters in diameter and 137 centimeters long. In the center of the reactor was a 46-centimeter in diameter “island” filled with heavy water (D2O) which not only served as a moderator (further reducing the required mass of uranium-235 to go critical) but cooled the moveable control rods located there as well.
A cutaway drawing of the Kiwi-A reactor. Click on image to enlarge. (LANL)
The reactor and its graphite reflector were housed inside an aluminum pressure vessel. The nozzle, manufactured by Rocketdyne, was a double-walled, water-cooled design made of nickel. This nozzle did not include a bell since the test objectives centered on the performance of the reactor itself. The first Kiwi-A reactor was intended to produce 70 megawatts of thermal power at a gaseous hydrogen flow rate of 3.2 kilograms per second for 300 seconds.
The Kiwi-A Tests
But even before the first Kiwi-A was built, there were already changes in the wind. Towards the end of 1957 it had become apparent to USAF planners that the Atlas missile would provide the US with an ICBM capability without the need to resort to exotic technologies like nuclear rockets. The infant nuclear rocket program would have died for a second time were it not for the launch of Sputnik on October 4, 1957 (see “Sputnik: The Launch of the Space Age”). The competitive pressures produced by the new Space Race meant that advanced technologies like nuclear rockets would be aggressively developed to give the country an edge in space exploration.
With the formation of NASA on October 1, 1958, the joint AEC-USAF nuclear rocket program was transformed into a joint AEC-NASA activity. While no longer needed for defense, nuclear rockets were ideal for space applications. In August of 1960 the joint AEC-NASA Space Nuclear Propulsion Office (SNPO) was formed with Harold B. Finger (who seven years later would become the Associate Administrator of NASA) as its manager. The goal of SNPO was to develop nuclear rockets that would aid the country’s effort to beat the Soviet Union to the Moon and planets.
The test firing of the Kiwi-A reactor on July 1, 1959. The yellow color of the plume is caused by a methane burner igniting the hydrogen exhaust of the reactor. (LANL)
While all these administrative changes were taking place, engineers were busy preparing for the first actual hardware tests. The first Kiwi-A reactor firing took place on July 1, 1959 at the Nuclear Rocket Development Station in Jackass Flats, Nevada about 150 kilometers outside of Las Vegas. It successfully fired for five minutes producing 70 megawatts of thermal power. But the test was not without its problems. During the test, a graphite closure plate above the reactor’s central island shattered with its debris being ejected from the engine. The damaged caused by the incident altered the flow of gaseous hydrogen through the reactor allowing temperatures to reach as high as 2,900 K. A post mortem inspection of the engine showed much cracking in the structures holding the reactor components in place caused by the unintended high radial thermal gradient inside the reactor. The graphite-rich fuel plates also experienced more hydrogen corrosion than expected. Despite the issues uncovered, the Kiwi-A test was considered a success with many practical lessons learned.
Diagram showing the modified Kiwi-A’ reactor. Click on image to enlarge. (NASA)
The next reactor, called Kiwi-A’ (pronounced Kiwi-A Prime), incorporated a number of improvements based on the experience with its predecessor. Instead of the fuel plates being used, the UO2 fuel was embedded in a matrix of graphite which was extruded into the form of long, rod-like cylinders which were then coated with niobium carbide to help reduce hydrogen corrosion. A total of six of these 23-centimeter long fuel elements were placed into each of the seven holes of a graphite module to produce a 137-centimeter long fuel module with these modules placed inside the reactor core.
The Kiwi-A’ reactor in its test cell ready for its firing on July 8, 1960. (LANL)
The first attempt to startup the Kiwi-A1’ was aborted when problems with the data channels resulted in an automatic reactor shutdown. A second attempt was also aborted when the methane flare system designed to ignite the hydrogen exiting the engine nozzle failed to operate. The next attempt on July 8, 1960 was successful with the thermal power output reaching 88 megawatts and an average temperature of the hydrogen exiting the nozzle of 2,178 K during an almost six minute run. But as with the Kiwi-A test, problems were encountered. Major power output perturbations were noted during the firing with debris seen exiting the nozzle. A subsequent inspection of the reactor showed that while the majority of the fuel elements had survived the test with little or no damage, 2.5% of them showed moderate to severe thermal damage from graphite corrosion and blistering of the niobium carbide coatings. Four of the fuel modules also experienced transverse cracking leading to their failures with the subsequent changes in power output and ejected debris observed.
The Last Kiwi-A Test
A third and final Kiwi-A reactor, designated Kiwi-A3, was built to further refine the Kiwi-A’ reactor design with the objective of operating at a 92 megawatt power level for 250 seconds. As before, improvements were made to this reactor based on previous experience. The short, 23-centimeter fuel elements were replaced with longer 69-centimeter elements. Different types of graphite components using various manufacturing techniques were also employed to determine which would provide the best performance inside the reactor. The various fuel module components underwent more extensive inspections to help eliminate those with hidden flaws.
The first test firing was attempted on October 7, 1960 but called off because the winds were blowing in the wrong direction for the deployed fallout detectors on the test range. The next attempt on October 10 was started successfully with the plan to fire at half power for 106 seconds before ramping up to full power for 250 seconds. During the half-power portion of the test, hydrogen exhaust temperatures reached 1,833 K or 305 K hotter than intended. After 159 seconds at half power, the reactor output was ramped up to an expected 92 megawatt level for the full-power portion of the firing. Again, the hydrogen exhaust temperature was much higher than expected with the gaseous hydrogen flow rate increased to 3.8 kilogram per second in order to maintain an exhaust temperature of 2,173 K. During the full-power plateau, several swings in temperature and thermal power output with an amplitude of up to 13 megawatts were observed.
The Kiwi-A3 during its test firing on October 10, 1960. (LANL)
Afterwards, it was discovered that a neutron monitoring instrument calibration error had led to an underestimation of the reactor’s thermal power output in realtime. Instead of running at the indicated 90 megawatt level for 259 seconds, the reactor was actually running at 112.5 megawatts – 122% of the engine’s power rating. As before, a post mortem inspection showed damage to the reactor components including the same type of damage to the fuel elements. But overall, the damage to Kiwiw-A3 was not as extensive as that found in Kiwi-A’.
Despite the many problems uncovered during the Kiwi-A firings, the reactor tests largely met their objectives demonstrating that a high-power density nuclear reactor could be controlled and heat hydrogen gas to high temperatures. With more design and engineering changes to come, efforts turned to the Kiwi-B series of reactors which would use liquid instead of gaseous hydrogen as a coolant.
Follow Drew Ex Machina on Facebook.
Related Video
Here is a documentary film produced by the Los Alamos Scientific Laboratory in the early 1960s about Project Rover giving a primer on nuclear thermal propulsion with footage from some early Kiwi reactor tests.
Related Reading
“The First Nuclear Reactor in Orbit”, Drew Ex Machina, April 3, 2015 [Post]
“The First Ion Engine Test in Space”, Drew Ex Machina, July 20, 2014 [Post]
General References
William R. Corliss, Nuclear Propulsion for Space, AEC, 1967
J.L. Finseth, “Rover Nuclear Rocket Engine Program: Overview of Rover Engine Tests – Final Report”, Prepared for NASA MSFC by Sverdrup Technology, Inc., February 1991
Daniel R. Koenig, “Experience Gained from the Space Nuclear Rocket Engine Program (Rover)”, LA-1006-H, Los Alamos National Laboratory, May 1986