A fuel element is where fission actually occurs. Fuel elements for solid core nuclear reactors come in many different varieties, and the differences can seem impenetrable to someone first learning about nuclear reactors. This may be because the fuel element looks like a lump of material, usually ceramic or metal, sometimes with some holes in it.

This simple shape can hide huge amounts of complexity, though. Nuclear fuel is one of the only times when you create a material knowing that this material’s elemental composition is changing, and that atoms are going to be throwing pieces of themselves apart at speeds a good fraction of the speed of light. This means that there’s enormous stresses on the fuel element at every level you look, from the microstructure of the material to the overall structure getting heated very intensely, causing major thermal stresses across the fuel element.

The challenges that a nuclear fuel designer faces are intense, and explaining the fine details of the design of fuel elements are well beyond my abilities. However, the basics of the different fuel elements are relatively easy to understand, and the implications for the differences can be huge.

Fuel elements for other reactor types tend to be more refueling mechanisms rather than the object that fission occurs within, and so will be focused on in each reactor type’s page (as they come online).

Back to Basics: Fission

The Gloriously Pesky Neutron

Fission is the process of splitting the nucleus of an atom. Because the nucleus of the atom is positively charged, a proton can’t be used to split the atom, and the electron cloud repels any electrons (which are tiny anyway). The best tool for splitting the nucleus is the neutron (both photons and protons at accelerator energies are able to split nuclei as well, but they are USUALLY less convenient to use), and while it makes fission possible, it also leads to many of the difficulties of nuclear power.

In order for the nucleus to absorb a neutron, they need to be in very close proximity for a given period of time. This distance and time period are different for different isotopes of different elements. To make matters more complex, neutrons can travel at different speeds, and the same neutron interacting with a nucleus at one speed can have a different result than the same two bodies interacting at a different speed. Once the nucleus absorbs a neutron, it usually just hangs onto it, and often undergoes either alpha or beta decay (another common reaction is the n, gamma reaction, where a different isomer, or nuclear configuration within the same isotope, is created; this is indicated by a letter after the atomic mass number – for our purposes, most commonly seen in the Am242m isomer). However, in certain cases the atom splits through nuclear fission, and these isotopes of particular elements are called fissile.

The interactions between neutrons and other matter is called neutronics, and it’s one of the most challenging realms in nuclear engineering. These interactions can be complex, and can only be understood well using incredibly difficult mathematical models running on supercomputers for extended periods of time to figure out the bare-bones sketch of what could happen.

The basics are understandable, however. Neutrons interact with different materials differently, and the number of neutrons compared to protons in the nucleus affects this behavior. Just as in the electron cloud, there are energy levels to the nucleons (protons and neutrons) in a nucleus, with valence states that need to be filled completely for stable behavior of the nucleus. The more stable the nucleus, the less likely it is to absorb the neutron, and the more likely it is to reflect it away instead.

Neutron reflectors act almost as mirrors, and tend to be heavier elements, such as tungsten, but beryllium is another common (and less massy) choice. Each element has a certain statistical chance to interact with a neutron, and one of the ways that it can interact is through reflection – where the neutron isn’t absorbed, and transfers a given amount of energy (depending on the relative atomic mass of the reflector material to the neutron’s mass). Especially for in-space reactors, the core is generally speaking surrounded by a neutron reflector in order to ensure that the minimum number of neutrons is lost. Each of these neutrons is going slightly slower than they were BEFORE they were reflected, so the use of a reflector “softens” the neutron spectrum. This makes an interaction more likely in two ways: the neutron is passing back through the reactor, giving it another chance to be captured and cause fission in the fuel, and the neutron is moving slower, so there’s a greater chance of neutron capture.

The mechanism behind this reflection is known as elastic scattering. In the case of very small atomic nuclei, such as protium (hydrogen with no neutrons), the mass of the two particles is basically the same. Since they are the same size, kinetic energy is more effectively transferred from the neutron to the proton (about half the energy going to the H nucleus that the neutron collides with), and the neutron slows, in this case almost exactly like a pair of billiard balls colliding. The larger the atomic nucleus, though, the less energy is transferred, although some still is, because the larger nucleus has more inertia. As such, it takes more scatterings for the neutron to be slowed the same amount by higher atomic number nuclei than smaller ones.

As a general rule, the slower the neutron, the more likely it is to be absorbed by a nucleus, but the energy of the neutron as it enters the nucleus has an effect on whether fission happens or not (neutron capture), and if it does it changes what the fission products that result are.

This slowing is called moderation, and is something that occurs in all reactors to a certain degree. One concern with moderating materials (and construction materials within the core as well) is their neutron absorption cross section. Every isotope of every element has a different neutron absorption cross section, measured in “barns,” as in “broad side of a barn.” The goal of the reactor designer is to ensure that there are enough neutrons, going the correct speed, with enough fuel, in the core of the reactor to keep a sustained reaction going, and the fewer neutrons you lose to anything other than fuel, the better. This is where isotopic enrichment of the various materials that aren’t the fissile fuel itself comes into play: as a general rule, the heavier the isotope of a particular element the less likely it is to absorb another neutron (although, this is where quantum mechanics and fundamental physics stick their badfinger in, and there’s no real substitute for careful, detailed observational experiment to get the best numbers possible). A good moderator will slow the neutrons, and keep them in the reactor core, without absorbing many. The most common ones are either water (H2O), or heavy water (D2O), which has hydrogen with one neutron instead of protium with none, and can generally burn less-enriched fuel than a light water reactor. Graphite is also commonly used in many reactor designs, although nuclear-grade graphite is special in that it can’t have any boron (which has a high neutron absorption cross section and is actually used in control rods and drums), and it degrades over time.

However, these aren’t the only materials used for moderation. Beryllium plays an important role in moderating some reactors, such as the TORY-II and BWXT LEU CERMET reactors. Beryllium and tungsten are used as reflectors, and also as components in CERMET – type fuels. Other materials, such as lithium hydride, are proposed for their high hydrogen content (which can moderate more with less mass), ammonia has been proposed before, and parrafin was used in many early experiments (unfortunately, it turns into goo and clogs pipes after a short time). However, the organic coolant that was more preferred in earlier reactors was known as Santowax, which was used in the Whiteshell Reactor #1.

When the “spectrum” of the reactor is being discussed, this is what they’re referring to: how much energy do the neutrons have? A fast reactor slows the neutrons as little as possible as a general rule, but because they’re going so fast a lot of fuel needs to be available to make up for the fact that each individual chance of interaction is lower (on the flip side, the likelihood of the fissile fuel absorbing a neutron and not splitting is far lower). For a thermal reactor, the neutrons have been slowed significantly, into the thermal region; these types of reactors are popular because they’re usually considered easier to control (i.e. prevent a runaway reaction), require less fissile fuel in the fuel elements (lower enrichment), and loss of moderator (as long as it’s not coolant) stops the reaction from occurring. There’s a mid-range as well, known as the epithermal region, which is where the NERVA rockets operated, but this is often a compromise region of the neutron spectrum because there’s a much higher chance of neutron capture of the fuel without fission occurring, leading to production of excess transuranic elements (which are all radioactive, quite a few fairly highly so, and most of which won’t split if another neutron gets captured), something highly undesirable in terrestrial reactors where radiation from TRU is a problem in spent nuclear fuel elements (although, since this is primarily in the form of alpha radiation, the (small) concern is mainly in the waste heat generated.

The major consequence of the neutron spectrum is how much fissile fuel (more on that below) is in the reactor overall, and therefore in each fuel element. The faster the neutron, the less likely an individual capture is, and therefore more atoms of the fuel need to be present. Often, this is done by enriching the uranium to increase its’ 235U content. Epithermal reactors (which are rare in most cases, but the NERVA engines were epithermal) also have a high fuel load, but it doesn’t need to be as high as for a fast reactor. Thermal spectrum reactors require the least amount of fissile fuel, since the neutrons have been slowed the most, and this fuel is usually of low-enriched uranium or natural uranium.

Another concern for a nuclear fuel element is power distribution through the element, especially if the fissile material and moderator aren’t evenly distributed through the reactor. Neutron reflectors, control rods, propellant or coolant running through the reactor, all of these things affect the neutronic environment the fuel element experiences, and have to be accounted for when designing and manufacturing a fuel element. The ideal depends on what you’re trying to achieve: for an electrical system you want a flat power distribution, where pretty much every part of the reactor core has about the same energy level. This keeps things simple, and the simpler something is the better. Thermal systems, on the other hand, are more complex beasts. In the case of an NTR, often the power distribution is heavily skewed to fading at the edges of the reactor core. This is typical in small reactor cores, as all NTR designs are, and is known as “buckling.” There are limits to how much you can play with this however: in order to get the most thrust possible out of the system, you want the propellant leaving each fuel element to be the same temperature and pressure, travelling in the same direction, to prevent turbulence from building up in the gas flow through the engine, and therefore maximizing the available thrust (these variations in pressure and density can also have both neutronic effects and mechanical ones).

Nuclear Stability, and Its’ Implications For Nuclear Fuel

Looking back at the nucleus for a moment: if you have one nucleon at a higher energy level with an available spot at lower energy, then the nucleon will drop into that lower energy state and let off a photon, just as an electron would. The key difference is the energy and wavelength of the photon: instead of a particular color, this is a gamma ray of a particular energy level and wavelength. This can be pinpointed so precisely that particular wavelengths of gamma radiation are used for remote sensing on orbiter missions around various bodies in our solar system, and have even been used to analyze the elemental composition and nuclear reactions occurring within distant stars and galaxies.

Unstable nuclei will go through radioactive decay, and the type depends on the element and its’ isotopic identity. Reducing the amount of instability is why a nucleus will shed a helium nucleus (alpha), dropping two spots on the Periodic Table; or it can go through beta decay (which usually ends up subsequently going through what’s known as “isometric decay” – in practice it emits a gamma ray as well), a very strange process, where a neutron doesn’t really (but seems to) split into a proton and an electron (or a positron), or it can under go fission, splitting into two roughly equal portions of nucleons (there’s a statistical spread to what’s created). This can occur without any influence, which is called spontaneous fission and occurs in some elements, but these obviously don’t stick around very long so they’re (generally) only found as man-made elements (although natural 238U also has a relatively high spontaneous fission rate).

Some fissile elements will spontaneously go through fission, as well, due to the extreme instability in their nucleus. These fuels offer many attractive options, including for nuclear propulsion designs that don’t require a reactor to maintain the nuclear propulsive mechanism (fission sails, for instance, a subject of a later post). The rate at which this occurs varies depending on the isotope in question, and can be either a boon or a problem for reactor designers. Generally speaking, those with a high spontaneous fission rate also tend to have very short half-lives, limiting their use as nuclear fuel.

For a nuclear reactor, there needs to be a good middle ground: stable enough that it’s controllable, but unstable enough that when power is wanted, the fuel will undergo fission and provide that power. Generally speaking there are two common choices for this, and three other less-common but also considered fissile fuels.

  1. Uranium 235: The most common form of fissile fuel. Used in most nuclear reactors and some nuclear explosives.
  2. Plutonium 239: Not nearly as common as a fuel for reactors due to complications both with non-proliferation concerns (isotopic enrichment isn’t necessary) and with reactor designs, but still used in MOX (mixed oxide) fuel.
  3. Uranium 233: Becoming more widely known due to the recent popularization of the Molten Salt Reactor Experiment at Oak Ridge, in particular in its’ two-fluid thorium breeder design (more on that below). Not as well characterized as either 235U or 239Pu. One issue with 233U is that it has fewer delayed neutrons in comparison to prompt ones, which can lead to reactor control and stability issues.
  4. Americium 242m: Am242m is interesting because it’s got the smallest “bare sphere critical mass” (i.e. the hardest it is to make something achieve criticality while actually being able to go critical), and as such is of interest to those looking to get a LOT of power out of a small amount of mass. However, it has a much shorter half-life than either U or Pu (432.2 years), so for some extremely long duration missions the decay of the fuel will possibly cause problems (for in-solar-system it’s fine, though).
  5. Californium 252: This is really the red-headed stepchild of the fissile fuels, almost nothing has been published on it except as a contaminant in spent fuel. However, it also can be weaponized, and therefore be made into fuel. However, in the fast neutron spectrum, this has the largest fissile cross section, and by definition fast reactors don’t need moderator, so there are mass savings to be had. Remember, every gram counts in spaceflight!

As a very important note, these fuels can’t just be swapped out with each other. This is something that is often missed in discussions of Gen IV terrestrial reactors, and the recent upsurge in popularity of the thorium breeder concepts (which burn 233U, bred by neutron capture and beta decay from 232Th). Each has a different neutron capture cross section, different fission products (each with their own neutron cross section and radiological profile), and different chemical properties (which can have a major effect on the ways that the fuel can be used). This isn’t plug-and-play, the reactors for each could look significantly different – even with the same power output for the same purpose.

Some other elements will undergo fission as well, but aren’t typically used for fuel in a reactor, either because they are too unstable and decay before they’re used, or because they don’t produce enough neutrons to keep up a sustained chain reaction, known as criticality (a reactor going critical is a good thing, supercritical not so much). This opens up some exotic options for propulsion that bubble to the surface regularly with a very interesting, but difficult to pull off, proposition: lithium-7 microfission, which doesn’t produce neutrons (and some nuclear physicists consider the p-B11 reaction to be fission, rather than fusion). These more exotic options will get their own page, because they’re different enough that including them in here doesn’t make as much sense.

Chart of Fissile Isotopes
Image via Wikipedia

There is more to this story than just those fissile isotopes, though. Of all those mentioned, only one of them (235U) is found in any significant quantity in nature, and it makes up a small percentage of the uranium found on Earth (and presumably elsewhere due to its shorter half-life). The vast majority of the rest is Uranium-238, which is not fissile. It can absorb a neutron, though, go through beta decay a couple times, then become Plutonium 239,a fissile isotope. This process is called breeding, and can also be done with thorium, which goes through two beta decays to become 233U, which is also fissile. Similar processes happen with Am and Cf as well.

What Do You Want?

Nuclear fuels are not a one-size-fits-all type of system. Each one has its own chemical and thermal properties that greatly affect everything from what kind of reactor it can be used in, to the materials that touch the fuel are made out of, to the temperature limits of the fuel.

For in-space nuclear power, there are generally two steady-state goals from a reactor: either the most efficient electricity production for an electrically propelled spacecraft, or very high temperatures for a thermal rocket. The first is what we do with nuclear power on Earth anyways, but there requirements are different enough that you can’t use typical fuel used in a light-water reactor and throw it in an astronuclear reactor and have it work.

Let’s have a look at the two concepts separately, to see the different goals for each different type of design.

Gimme the Juice!

DUFF Test setup McClure 2013
DUFF Experimental setup, image courtesy DOE

The thing about a nuclear electric power source is that it’s not just the fuel temperature that matters: the heat transport system and power conversion system need to be matched to the reactor core, and the entire thing needs to be balanced between these systems to work properly. This can very easily be seen in the DUFF experiment, where the working fluid for the experimental heat pipe had to be changed in order to work at the power level the reactor provided, and the Stirling convertor used had to be changed as well. Because of this, while higher temperatures mean more efficient operation on a fundamental level, on the practical side an electricity-producing reactor doesn’t need to operate at nearly as high a temperature as a thermal rocket does. This is a good thing, because it means that there’s less thermal stress on the fuel itself, meaning less swelling, less cracking, and less chance for the fuel to melt down.

Nuclear_Fuel_Pellets_(14492225000)
Uranium oxide fuel pellet, image courtesy Wikimedia Commons

Another type of fuel that’s sometimes used in space-based designs, but is exceedingly common on Earth is the oxide-based fuel. This comes in many different types, from many different manufacturers, in many different enrichments, and is perhaps one of the most mature (if not the most mature) form of nuclear fuel element in the world today. However, as is widely known, they tend to not be able to handle very high temperatures, and for a nuclear thermal rocket that temperature is everything. As such, except for some electricity-producing designs, these types of fuel elements aren’t seen much in in-space reactor designs.

CANDU Fuel Bundle
CANDU fuel bundle, image courtesy Canadian Nuclear Association

Finally, there’s the metal fuel element. This is an alloy of fissile and non-fissile metals, with the non-fissile metals picked either because they do moderate or don’t moderate the neutrons in the reactor. This can be a remarkably simple fuel element (two halves of a sphere of 235U, each just over half the critical mass for a bare sphere, for instance), or can involve changing the alloy as the fuel element has cast to tweak the distribution of elements within the fuel element. Perhaps the most famous type of metal fueled reactor is the CANDU reactor, one of the most popular nuclear power plant types worldwide. Often, though, these also are only seen in electricity-producing reactor designs, since they simply can’t handle the temperatures needed for thermal operation AS A SOLID. The LARS molten NTR actually counts on this melting of the fuel element, keeping the molten uranium in place by spinning the coolant channel. But that’s an NTR, not an electric design.

Bring on the Heat!

A nuclear thermal rocket, on the other hand, has one goal: to produce as high a temperature as possible in the propellant, and in order to do this the fuel has to get as hot as possible. This means running your reactor as close to the limits of its materials as the system allows. To make matters more entertaining, the most common propellant used is hydrogen, which does terrible things to most materials at high temperatures (both erosion and hydrogen embrittlement are major issues that were faced during Project ROVER).

Coated Particle Diagram
Later ROVER fuel element composition diagram, image courtesy NASA

To deal with the heat of a nuclear thermal rocket, Project Rover and NERVA rockets primarily used a graphite composite structure, with pellets of 235U between the sheets of graphite, which acted as a moderator. The size and spacing of these microparticles was the subject of a lot of research at Los Alamos during Rover, and by the end of the hot-fire testing, beads of uranium carbide coated with pyrolitic graphite (an early form of TRISO fuel) of varying sizes and enrichments were used. This improved the fuel element effectiveness significantly. All of these fuel elements were very susceptible to erosion from the hot hydrogen propellant, though, so cladding was needed. Often this is molybdenum, but there are many options out there. I’ll be going into this more in-depth in The Engines that Were, so we won’t go more into it here for now.

Reactor cell w carbide, Finseth

 

Another fuel form that has been researched both in the US and the USSR is the use of carbides instead of graphite composite. Here, instead of sheets of carbon, the carbon is mixed with a less electronegative element to make an incredibly hard material that can withstand very high temperatures. There are various options for the carbide structure, but since there’s more than just fissile fuel and carbon in these fuel elements, choosing which element is best is a matter of balancing high heat tolerance and resistance to the hot hydrogen with the neutronic effects of the rest of the matrix (everything not fissile atoms) of the fuel element. Testing on carbide fuel elements has been conducted in the US and the USSR, later in Russia. There are different challenges to the carbide fuel than the graphite composite fuel, as was demonstrated during the Nuclear Furnace test in 1973 (23 carbide fuel elements were integrated into the reactor for testing).

CERMET micrograph, NASA
W-UO2 CERMET fuel micrographs, image courtesy NASA

One of the more talked about concepts in nuclear thermal propulsion right now is the use of low-enriched uranium (LEU) for a nuclear thermal rocket. Because there’s less fissile material in the fuel elements themselves, the neutron spectrum needs to be softer (i.e. the neutrons need to be slowed more with a moderator). In order to do this, a different sort of fuel element is being investigated. Rather than have the fuel be in a carbon matrix, here the fuel is held in a ceramic metal, called CERMET. This is an odd chemical composition, combining the thermal properties of metal with the materials properties of ceramic, but it allows for incredibly safe high-temperature operation. Because of this, it is being investigated for accident-tolerant fuels in current nuclear reactors. Another advantage is the ability to control its composition very precisely, changing the ratios of various elements as the production process proceeds. This means that moderators, fission poisons (for advanced waste management designs), and fuel can be distributed within the fuel element to make a very complex, but more efficient or reliable or less waste-producing fuel element. This is the real advantage of CERMET: the flexibility that it offers as a material. However, the fuel elements also tend to swell enough to cause problems for certain designs. As with everything in the nuclear industry, this can be addressed, but in order to address it it needs to be understood. With the greater ease of working with LEU fuel, as NASA’s new reactor uses, many more institutions (both private and public) will be able to participate in, and contribute to, the testing effort required to fully characterize these new fuels.

This does not exclude carbides from being used in LEU thermal fuel elements, as can be seen in multiple designs currently under investigation. One of the more interesting designs is known as SULEU, or the Superior Use of Low Enriched Uranium. Not only does this design utilize high-temperature tricarbide fuel elements, but the control drums actually remove a portion of the core when they rotate, further limiting the critical capabilities of the core. However, due to funding priorities and apparent lower technical maturity, NASA has chosen to pursue the CERMET fuels more aggressively at this time.

The Shape of Things

The purpose of the reactor, and its operating principles, define the shape of nuclear fuel more than the other way around. In a civilian light water reactor, the individual fuel pellets are carefully manufactured with different levels of enrichment for different locations throughout the reactor core, and these locations are changed over time. Kirk Sorenson compares this to stirring a campfire, and there’s a lot of truth to that metaphor: there’s more reactivity in the middle of the core than on the outside edge, so more fuel will be burned up the closer to the center of the core it is.

In an in-space reactor, this reshuffling of fuel isn’t an option: in-space power reactors have to operate properly for years without maintenance or repair, in incredibly difficult environments. Nuclear thermal rockets have much shorter lives, with operational lifetimes measured in hours, due to the fact that this maintenance can’t take place on current designs, and the realities of the extreme environment of an NTR reactor core on top of the space environment. The same fuel element that was designed for use in a terrestrial HTGCR for hundreds or thousands of hours of operation would only last one or two in the harsher chemical, thermal, and fluid dynamic conditions of an NTR.

Electrical FE

There are a number of different options that have been proposed for the shape of fuel elements for in-space electric production. The shape mainly depends on the physical size of the reactor, how the fuel elements are being cooled, and how electricity is being produced.

TOPAZ Core configuration, Bennett
Image via NASA (Bennett, 1989)

The vast majority of the reactors that that have flown have used in-core thermionic fuel elements. Often, this is just a simple cylinder of fissile fuel inside a thermionic housing, sometimes called a flashlight configuration. Electricity and heat are extracted through the thermionic elements. These can be oxides, carbides, metals… this is a very flexible fuel form, and that makes it very attractive.

Other systems use sodium coolant running to a power conversion system, either a Stirling, Brayton, or Rankine cycle. Other coolants have been proposed as well, depending on the temperature of the working fluid and the power conversion system. Sometimes, the coolant runs through the fuel element using internal channels. Other times, fuel is held in rods or pins, and coolant is run around them in a setup called a calandria, similar to a CANDU. Some designs have fuel elements with interspersed channels as well.

The development of NTR fuel elements was paralelled in the development of the high temperature gas cooled reactors, and as such there are more similarities than differences between these fuel elements. The main difference is that the working fluid in the HTGC reactors is typically helium, not hydrogen, since the exhaust velocity isn’t the driving force in power plant design as it is in rocketry. Instead, a lower temperature, more steady operation allows for much longer fuel element life and far fewer concerns about chemical reactions between the coolant and fuel element. Many bimodal designs exploit this common heritage to produce electrical power.

DU Core FE
KRUSTY depleted uranium surrogate core, image courtesy DOE

Current NASA development of Fission Power Sources (FPS) is focusing on the Kilopower program, which uses a monolithic (single-piece) fuel element of Uranium-Molybdenum alloy, cooled by heat pipes. The first round of testing (DUFF) went incredibly well, leading to the non-nuclear, and in November 2107 nuclear, testing of KRUSTY, culminating in the first full-power fission test of a new nuclear reactor design in the US in over 50 years. You can read more about the experiments here.

There’s a lot more flexibility in fuel element design for electricity production, so this is an area that’s almost as varied as the number of reactor systems out there.

Thermal FE

A nuclear thermal rocket is a specialized form of a high temperature gas cooled reactor, some of which have been built and operated for electricity production around the world, mainly in the UK. In this case, the temperature of the fuel elements, the amount of heat that the gas can absorb, and the amount of gas that needs to pass through the reactor to keep it from melting itself all have to be balanced. In the case of the nuclear thermal rocket, the gas leaving the core has to be as hot as possible, which means running the system as close to the margins as you can safely get away with at full thrust. Getting this balance right was a major focus of Rover, and is a major problem for solid fuel elements.

Because of this, having the fuel not be solid extends this upper temperature limit, making the rocket more efficient. These fuel forms are covered in the different reactor pages, with the implications of the various design choices looked at there.

For a solid fuel element, the name of the game is to maximize the surface area of the fuel element while minimizing the mass to maximize the amount of energy that can be transferred to the propellant. This can be done a number of ways, but two common shapes tend to be most common: either a hexagonal fuel rod with different numbers and sizes of channels along the long axis of the fuel elements; the other option is a twisted ribbon of fuel in a channel.

pebbleBed03
TRISO Fuel Element via Project Rho

Another popular option is the use of a pebble bed reactor, where the fuel is in the form of individually coated beads or pebbles, which was used in the Timber Wind reactor project for the DOE, as part of the Strategic Defense Initiative. These are the main choices of the US and Russia, respectively, but other options have been proposed as well. A good example would be the pulsed NTR, designed by Dr. Arias at the University of Catalonia, which uses plates of fuel and propellant passing between them. This is a very unusual and unique reactor type, because it can generate bursts of reactivity due to its core structure – which allows this reactor to adjust the thrust of the rocket (but, as with most rocket propulsion systems, this means that your specific impulse, or “fuel efficiency,” suffers). This fuel form (with unclad graphite plates) was also seen in the first KIWI reactor during Project ROVER, but due to chemical stability issues this fuel form was discarded.

Another limitation on the shape of the fuel element is the material properties of what the fuel element is made out of. Each material is strong and weak in different ways, shrinks and swells differently… and all of this has to be accounted for in the design of the rest of the components of the reactor to account for this.

Composite fuel element matrix drawing
ROVER composite fuel diagram with cladding (NbC or ZrC), image courtesy NASA

Finally, most of the materials used for fissile fuel react with the hot hydrogen used as propellant, and if hydrogen isn’t used other reactions rear their heads. To avoid this, various different materials are used as coatings for the fuel elements; these materials are collectively called cladding. There are various methods to place the clad between the fuel and the material, either by having the clad be a separately manufactured piece that is then installed during construction, or by using a method to coat the fuel element with a layer of metal. Depending on the material used, the temperatures that the clad has to respond to, and the swelling of the fuel element, choosing the proper clad is an important decision. Continuous development of various clad options were conducted both at Los Alamos and at the National Security Test Site, and development has continued (albeit at a slower pace) since.

This is Where the Action is At

Fuel element design drives the entire nuclear rocket engine design, sometimes in large ways – such as the size of the reactor core or the maximum temperature of the exiting coolant – and sometimes in small ways, for instance the size of turbopump you need, or how much electricity is available for the rest of the ship. Perhaps most importantly, when the fuel elements are no longer usable, due to either damage or to fission poison buildup, the reactor no longer works, and except for possibly capturing some waste heat nothing else can be done with that engine unless the reactor can be refuelled.

With all of the options available, there are several (including oxides, metal fuel, and others) that are being looked at for electricity generation, and three that are of interest to various space agencies for nuclear thermal propulsion. Due to the more challenging nature of the thermal fuel elements, these get far more attention, but the electricity-producing designs are just as important, and just as varied in their options.

DU Core w clamps
KRUSTY core with installed heat pipes, image courtesy NASA

On the side of electricity production, NASA currently has a program called Kilopower, which together with the Advanced Stirling Radiothermal Generator forms the backbone of the planned, lower-power electricity production systems. Russia still has the BES-5 reactor, and China has expressed interest in this reactor as well. Since it’s the most well-established, reliable system, given the seal of approval by not only Rosatom and Roscosmos, but also the Department of Energy and NASA, and Russia is very willing to export nuclear technology, this design could once again see use.

There are also larger designs on the books, but most of them have largely been paper reactor studies due to the challenges of testing systems on the scale that these reactors require. These include the Fission Surface Power reactor and Project Prometheus on the NASA side. While it’s certain that there are designs of similar size classes in the Russian catalogue of designs, I’m still working on finding out information about them.

For thermal rockets, the goals of increasing the thrust-to-weight ratio and the specific impulse of the rocket are directly impacted by the temperature that the fuel can be heated to, and how much of that heat can be transferred how quickly to how much propellant. High-temperature fuel forms developed for nuclear thermal rockets (namely CERMET) were first being looked at for highly accident-tolerant fuels on Earth (Al-matrix fuels are very attractive for the lower temperatures needed for lower-power-density reactors that are good for terrestrial applications), and this knowledge has in return fed back into rocket design.

Because fuel elements are often compact, and are able to be thermally tested for swelling and damage by heating them electrically (and even passing hot hydrogen through them to check for reactions), much work has been done over the years at various centers for various fuel elements that have reached a relatively high level of technological maturity. Nuclear rocket designers can use these fuel elements to design new systems, but this means that the resulting rocket engine is fundamentally limited to what the fuel element is able to provide. Due to the cost and time savings associated with using off-the-shelf technology, however, these legacy fuel elements (or variations on them) may yet fly in nuclear spacecraft in the future.

More to Follow!

Each reactor design on this page will eventually have a section on the fuel element used for the reactor in question. In addition, as time allows there will be additional pages for each of the fuel element types. Make sure to keep checking back for more updates!

Special thanks to all that have contributed (rather heavily in many cases) to this page. Not having formal training in nuclear engineering, this page had MUCH more help than average! Special thanks to Jaro Franta for his time and patience answering MANY questions about fuel element design!