The solid core nuclear thermal rocket (NTR-S) is the only type of nuclear thermal rocket system ever tested as a full system, and was even flight rated for manned missions. In these rockets, a nuclear reactor using solid fuel (in the form of pellets, rods, plates, or other options) is used to heat a propellant gas (usually hydrogen, but ammonia, methane, and CO2 have all been suggested). This is then expanded out of a nozzle to produce thrust, carrying away as much heat as possible in the exhaust gas stream.

NTR Flow Diagram via NASA

Solid fuel elements are the most familiar type of fuel for reactor designers, and as such these designs require less (but still significant) work to make ready to fly, and so are likely the first to be flown. There are a wide range of different types of fuel, in different shapes and sizes, though, which means that even in just this category there are many different options. These include more traditional hexagonal prism shapes, fuel pellets held in rods, TRISO (tristructural isotropic, referring to the three parts of the pellet: fuel, moderator, cladding) fuel pellets in a pebblebed reactor design, plates, and others, notably the twisted ribbon design used by Russian reactor designs. All have the same goal, though: to maximize the amount of contact between the fuel and the propellant, which also acts as coolant for the reactor core.

Because the coolant is leaving the reactor as propellant, these designs usually don’t need large heat radiators – which is almost unique among reactor designs. Also, because the gas both cools the reactor and is propellant, the flow of fuel is usually started before the reactor itself starts up and continues for a time after the reactor is shut down, for the needed moderation from the hydrogen to occur and to get rid of the decay heat from the reactor after shutdown. This is one of the unique characteristics of NTRs from a guidance, navigation, and control perspective compared to chemical engines (thrust buildup time of up to a minute is another, if GNC considerations for NTRs interest you I recommend this paper as a good starting point).

Schematic view, Finseth
NERVA NRX A6 Cutaway, Finseth 1991

The most common types of NTR-S are variants of a number of reactors designed by the US during the 50s through 70s, as part of Project Rover, a joint program by the Atomic Energy Commission (later the Department of Energy) and the US Air Force (transferred shortly after the start of the program to NASA). This program tested four basic reactor types in hot-fire tests at the National Nuclear Test Range in Nevada, at the Jackass Flats facility, and in reactor engineering tests at Los Alamos Scientific Laboratory (later Los Alamos National Laboratory). These were the KIWI-A, KIWI-B, Phoebus, and Pewee reactors, with another two tested at Los Alamos only – the Dumbo and MITEE reactors. As the program advanced, the Phoebus reactor design was chosen for the NERVA program (Nuclear Energy for Rocket Vehicle Applications) for the specific goal of building a flight ready, man-rated nuclear thermal rocket for the post-lunar missions to Mars that were planned as a follow on to the Apollo program.

Undated NERVA test, via LANL

However, with the cancelling of the later moon missions, much less missions to Mars, and the increasing public resistance to nuclear power, the program was wound down, with hot-fire tests stopping in 1972, but work continued after and the program’s final report wasn’t released until 1991. Work on at least two of these designs, the Pewee and MITEE reactors, continues today at a much lower level, with a derivative of Pewee, the Small Nuclear Thermal Rocket (SNTR), being the most advanced design in terms of development, and likely the one that will be used first if a nuclear thermal rocket is used in the relatively near future.

RD-0410 via Astronautix

The USSR also had a number of development programs for NTRs, most based on uranium carbides, which offer incredibly high heat tolerances (for solid fuels). Initially a first stage rocket was proposed, but was cancelled shortly after. Other programs came and went over the next few decades, and finally culminated in a pair of engines: the RD-0410 and RD-0411. These reactors are unique in the shape of the fuel elements: these reactors used fuel elements shaped like a twisted ribbon in order to maximize the amount of surface area of the fuel that is in contact with the propellant, allowing more heat to transfer into the propellant while also allowing more propellant to move through the reactor.

Other fuel forms have been proposed for NTRs, in fact NASA’s current focus is on CERMET, or cermaic-metal composite, fuel for over 30 years. This is a fuel form that, while it doesn’t have as high a heat capacity as carbides, allows for much more flexibility in fuel composition. Because of this, lower enrichment is required for an NTR to function (see the Fuel Elements page), and so the US is pursuing this option for near-term NTR development. For more, see my series of posts on the NASA NTP program.

Pebblebed cutaway via Project Rho

Pebblebed reactors are another concept that is popular for NTRs. The most developed of these designs was for Project Timberwind, a partner program to the Strategic Defense Initiative (Star Wars). The advantage to the design is that, since the fuel isn’t held in a static rod, it can move through the reactor as it heats and cools, and as the fission reaction occurs the fuel also changes density and moves differently as well. This reactor has a lot of advantages, since it increases both the temperature that the reactor can run at and also increases the surface area between the fuel and the propellant. However, in order to hold the fuel in place the reactor must spin quickly enough to hold the fuel pellets in place with centripetal force, which adds complexity to the reactor design.

Pulsed NTR, image via Wikipedia

Another design uses plates of fuel rather than pellets or pebbles. This design is a much more advanced design, proposed by Dr. Arias at the University of Catalonia. This design is based on the TRIGA (Training, Research and Isotope General Atomic) reactor, which was designed for students to be able to experiment with a reactor that is very difficult to cause a meltdown in. This reactor can pulse up to 33,000 MW for brief periods of time without causing damage to the reactor, due to the very short time that the pulse occurs there’s very little energy that’s able to transfer into the body of the reactor itself. This pulsing caught the attention of Dr. Arias of the Polytechnic University of Catalonia, who designed the Nuclear Thermal Pulse Rocket. This engine has two modes: the first maximizes thrust at the cost of specific impulse, the second mode maximizes specific impulse at the cost of greatly reduced thrust.

BNTEP Diagram McMurdy 1994
Bimodal Nuclear ThermoElectric Propulsion diagram, McCurdy et al NASA GRC 2014

There are other variations on each of these reactors. Two of note are the Binary Nuclear Thermal Rocket, which includes a power conversion system in addition to the thermal propulsion equipment, and after the initial burn using the thermal rocket the reactor is switched over to producing electrical power for an electric propulsion system. This allows for fast orbit changes that require high thrust, but also allows for increased speed by using the continuous burn of electric propulsion. The second variation is the LANTR, or Liquid Oxygen Augmented Nuclear Thermal Rocket. This is effectively an afterburner for an NTR. Since the hydrogen that is used as propellant isn’t combusted, burning it can increase thrust significantly the same way as an afterburner in a jet engine (although there, it’s the unburned oxygen that fuel is added to, not the other way around). Both of these options are available to any NTR system, but the solid fuel systems are the ones where the concept is most developed.

LANTR diagram borowski
LANTR concept, Borowski 1995

Reactivity and Neutronics Control: Fuel Elements 101

Almost all reactor designs use some form of moderator, to slow the neutrons enough to increase the chance that they’ll be captured by an atom of fissile fuel and split. In the case of nuclear rockets, this moderator needs to be as light as possible, and usually minimally bulky as well (although that is usually less of a challenge). Many designs don’t even use a moderator, instead using external neutron reflectors (made out of depleted uranium or beryllium) to keep the neutrons in the core for longer.

NTR FE photo
NTR Fuel elements, via NASA

In the initial reactor designs for the US, and all of the reactor designs for Russia, the fuel was a graphite composite structure, with bits of the uranium fuel suspended in graphite. This not only moderated the neutron within the fuel itself, but the high thermal loading tolerances (it can get hot or cold very quickly with minimal damage) also make this an attractive option. Because the hydrogen propellant (or hydrogen in the propellant, if H2 isn’t used) will react with the carbon, causing significant erosion of the fuel elements. Because of this, a cladding substance is needed, usually zircalloy (a special zirconium based alloy used in terrestrial reactors for the same purpose). This is another level of moderation.

Other materials are used as well. In the US, most recent and current development has focused on ceramic-metal composite known as CERMET. This combines the advantages of ceramic fissile fuel with the heat conductance and toughness of metal. The other common option is to use carbides, in various forms, because they have incredibly high heat tolerances. This is an avenue that is being pursued in Russia, and has been extensively experimented with in the US as well, with some proposals continuing to this day.

For a more in-depth look at the types of fuel elements and the tradeoffs involved, check out the Fuel Elements Page [Coming soon!].

TRITON cutaway with control drums Borowski
Cross section of NTR, bottom arrow points to control drum, neutron poison on small section facing hexagonal core. Image via NASA

Rather than control rods that pass through the reactor, the majority of the neutron control in most designs comes from a set of drums mounted around the outside circumference of the reactor. These drums are coated on one side with a neutron absorbing material, usually boron, and on the other with a neutron reflecting material, usually beryllium. By changing the orientation of the drums in relation to the reactor, the neutron flux can be absorbed to prevent the nuclear reaction, or can be reflected to start and throttle the reaction. Depending on the orientation of the drums, more or less neutrons are reflected or absorbed.

Some designs do still have a control rod or two, but because the control rod has to be physically removed from the reactor this requires an additional housing for the retracted control rod(s) to move into, taking up more volume and requiring additional motors and drive systems. Because of this, and the demonstrated effectiveness of the control drums during Project Rover in the US, this design choice is less common than it once was.

Propellants: Advantages and challenges

NTRs of all types use propellant heated in a reactor to make thrust. Because the goal is the highest exhaust velocity possible for maximum efficiency, hydrogen (H2) is one of the most efficient propellants due to its low atomic mass, only able to be beaten by monatomic hydrogen. Unfortunately, solid fuel cannot reach the temperatures required to dissociate H2 into monatomic hydrogen, so we won’t look at it here, but more advanced designs can do this for an increase of 1000 seconds of specific impulse.

Other options have been proposed, and are looked at more in depth on the NTR main page.

Size and Power Output

LASL Reactor Family, Houts NASA
The Rover family of test reactors, image courtesy NASA

Many of the early designs, both for the US and the USSR, were large reactors, sometimes 1,000 MW or more. The first design the Russians considered was an ammonia-propelled reusable nuclear first stage (in the late 1950’s by OKB-1), but safety concerns caused that to be dropped rather quickly. The largest American reactor ever tested, the Phoebus 2A, was 4000 MW. However, reactor sizes dropped rapidly, with the size of proposed reactors in the USSR dropping from 200MW in 1963 to the two more current designs tested in the 1980’s producing at most 70 MW (NPO Luch RD0411). The current frontrunner for a solid-core NTR in the US is the Small Nuclear Thermal Rocket, or SNTR, designed and extensively researched by Dr. Stan Borowski of NASA’s Glenn Research Center, is rated at 367 MW, similar to the Pewee class in Project Rover in size.

There are a number of reasons for this. First, the reactor’s power output also has a direct effect on the amount of radiation mitigation needed for the engine, and second multiple engines can be used for a single spacecraft. For radiation shielding, this means a larger, heavier craft: either you can extend the payload (including habitation modules) farther away from the reactor, making the ship longer and adding some weight, or adding more dense, heavy shielding at the same place that a good percentage of the mass of the ship was concentrated anyway: right next to the engine.

Criticality Limited NTR Diagram, Borowski 2014
Criticality limited SNTR, Borowski et al 2014

Another consideration is that in the case of both the US and Russia, unmanned probes are proposed as the first use of this technology: for the US, a 145 MW rocket has been proposed as a forerunner to the SNTR, called the Criticality-Limited NTR; the Russian RD0411 is designed for unmanned probes (392 kN thrust, unable to find reactor power).

As a general rule, nuclear power scales better big than it does small, and the same can be said as a general rule of thumb for nuclear rockets from an energy density point of view. However, limitations in the engine system as a whole will limit the size that an engine can reach (although engines significantly larger than 4000 MWt are easily doable).


More to come soon!