Hello, and welcome to the Beyond NERVA blog! Today, we continue our in-depth look of NASA’s new nuclear thermal rocket. We briefly looked at the history of NTP (as NASA calls it, “nuclear thermal propulsion”) in part one, and in part two we took a deep dive into the materials that NASA is investigating for its’ new design, ceramic metal (CERMET) fuel elements. Today, we look at the stage and spacecraft itself, with a brief look at some information about the proposed engine design. The next post will focus on the testing and launch safety considerations for an NTP system (as well as some unique guidance, navigation, and control considerations), and we’ll close with a post about other options for using low enriched uranium to fuel a nuclear thermal rocket, this time using advanced carbide fuels.
As we saw in the first post, nuclear thermal rockets are nothing new. The US has built and tested them before, and even successfully tested one in flight configuration. In the second post, we looked more closely at the new materials technologies that are being used to make an even more capable NTR, CERMET fuels, but we also saw that there’s a problem: in order to use low enriched uranium (LEU), the fuel needs large amounts of isotopically separated tungsten, and this has been a major challenge for the supplier. To date, I have been able to find no information about deliveries of even 50% enriched 184W (the needed isotope), much less the more than 90% enriched tungsten needed for the fuel elements that NASA has designed.
So how does NASA plan to address this problem? Well, 184W would be useful for more than NTRs (tungsten is also used as a neutron reflector in the core of certain thermonuclear weapons designs, hence the lack of details on the development process and difficulties associated with it), so just because NASA has not been able to have the process developed doesn’t mean that it won’t be in the future (weapons programs have an easier time getting money than NASA’s nuclear program).
Advances in LEU Nuclear Propulsion
Even if this doesn’t pan out, there’s still other options. The one that caught the public’s attention last year was the signing of a new contract with BWXT. While far from a household name, in the DOE, US Navy, and NASA they are well-known. They helped build the USS Nautilus, and were an early (and currently are the major) supplier of fuel for the US Nuclear Navy. They offer commercial and research fuel resupply and disposal contracts on a number of reactor designs. Since the early 1980’s they have fabricated all of the Department of Energy’s experimental fuel elements (with the exception of KRUSTY, which was fabricated by Y12). They also are a prime contractor for many of NASA’s nuclear-related activities, participating in environmental impact assessments, technical consultancy, and other areas.
They have proposed a new, and thus far poorly described, design for an NTR, using CERMET fuels of a varying composition at different points in the core, to better manage and moderate the neutrons produced during fission. Based on what little information I’ve been able to gather since NETS 2018, according to Michael Eades using molybdenum/tungsten (MoW) as a matrix material for CERMET fuels is apparently as good as using tungsten-184 for both moderation and thermal limits, and this appears to be the path that BWXT will be using moving forward. However, I haven’t had the chance to go over the research yet, and apparently there are some significant changes, so today we’ll focus on the rest of the spacecraft.
It’s likely that the design will be similar to the one proposed by BWXT last year, however, which uses a technique known as “zoned moderation,” where different parts of the reactor are exposed to different neutron flux energies due to the distribution of moderator and reflectors throughout the core. There is no reason that this technique will not work using natural uranium as a fuel element matrix material rather than the beryllium and tungsten that was proposed for the earlier design.
Even this isn’t the end of the options, though… another fuel form, advanced tricarbides, also offer the potential for LEU use, in particular in the Superior Use of Low Enriched Uranium (SULEU) reactor design, which we’ll cover in a future blog post (not only is it a carbide-fueled reactor, but there are enough other nifty design features that this reactor definitely needs its’ own post).
The Beginnings of the Modern Astronuclear Thermal Era
Beginnings are important in nuclear engineering. A retired Lawrence Livermore engineer once told me “nuclear design is evolutionary,” and that’s especially the case with this system.
In many ways, the new dawn for nuclear propulsion was in 1990, at “Nuclear Thermal Propulsion: A Joint NASA/DOE/DOD Workshop,” held in Albuquerque, NM from June 10-12. This conference happened during the death throws of the Strategic Defense Initiative (SDI, Reagan’s Star Wars program), when funding was being cut for every single program associated with SDI. There was a nuclear thermal rocket design that was part of SDI, Project Timberwind, which used a pebblebed reactor to increase fuel surface area, but this was a relatively early casualty of Congressional budget cuts. In addition, as noted by the DOD Office of the Inspector General, the program not only was over budget and consistently failed to meet benchmarks, but there were questions about the predicted performance of the engine as well.
In order to continue moving forward with nuclear thermal propulsion, the three main stakeholders in the US came together to present their concepts. The conference started by establishing what had come before, and also a “baseline” was established to compare new ideas to the legacy NERVA designs that were available at the time. New subsystems, techniques for handling cryogenic hydrogen, and materials all would combine to make even an NTR using the same fuel elements and reactor geometry would be greatly improved over what was available in 1973. After this, presentations were made about many different aspects of nuclear thermal propulsion, from launch safety concerns to materials advances to advanced concepts for liquid, vapor, and plasma fueled reactor designs. The focus was on getting the most bang for the very few bucks that would be coming down the pipeline, and on the difficulties of testing any design under the regulatory regime that was in place at the time (which was not hugely different from what we face now).
The only way at the time to be able to test an engine was to capture ALL of the exhaust that passed through the reactor, which meant that you had to be able to store it somewhere – a very big somewhere. It also meant that more exhaust translated rather directly into greater expense for testing, so the thrust of the designed engine was specified to be in the 25,000 klbf range, similar to what the Pewee engine provided during Project Rover. We aren’t going to be getting into testing options in this post (that’s the next one), but keep in mind that the ability to fully test an NTP system on Earth is going to be a critical requirement, and the size of the engine (and the amount of propellant that needs to be captured) directly affect how difficult (and expensive) it will be to test as a full system.
This conference also can be seen as the birthplace of the immediate predecessor of the LEU NTP system: Stan Borowski’s Small Nuclear Rocket Engine (updated version of the design available here). This engine is a Pewee-class, graphite composite modern design, and still remains an option for a small NTR, although one that would require HEU rather than the LEU that NASA is currently focusing on. This engine also allows for bimodal operation, where oxygen is injected into the hot hydrogen stream and then ignited, giving a big boost to the amount of thrust available (at the cost of specific impulse), which became the LANTR, or Lunar Oxygen Augmented Nuclear Thermal Rocket, for faster trips to and from the Moon (which is so close the increased thrust is a big boost to mission capabilities).This design was investigated for more than a decade as NASA’s primary NTP concept, and remains an area of active research at both NASA’s Glenn Research Center and at Oak Ridge National Laboratory, where many of the fuel fabrication techniques are being investigated in depth.
Many of the parts of the SNRE stage remain in the LEU NTP stage. These include non-nuclear components, the basic shape and volume of the stage, nuclear and thermal shielding, and while slightly changed the mission requirements are largely the same. Perhaps the only major-ish change is the difference in the proposed launch vehicle: in the early days of NASA’s Design Reference Mission for Mars 5.0, the Ares V rocket was still on the drawing boards – and in the mission plans. This rocket ended up being canceled with the end of the Constellation program, and a slightly smaller replacement, the Space Launch System, was proposed. The difference between the rockets necessitates a re-juggling of what is launched on each orbital launch, due to the decrease in payload capacity from 140 mt to low Earth orbit to 110 mt, but this is something that can be addressed relatively easily by using slightly smaller modules (although often it ends up requiring an extra launch). So, looking back over the design proposals leads to a lot of insights into NASA’s thinking and requirements for their new nuclear rocket design.
Since there’s still a lot of questions about the exact form that the engine itself will take, let’s go ahead and look at the rest of the NTP stage: shielding, non-nuclear components, propellant tankage, size, and mission requirements.
Radiation shielding is essential on any nuclear system, but nuclear propulsion presents a number of challenges that are unique. Of course, the biggest part of this effort is to reduce crew dose during operation, but the engine components that aren’t in the core of the reactor (such as turbopumps, actuators, etc) will also be materially attacked by the radiation flux coming off the reactor, and the fuel itself can be heated as well (which causes local boiling and cavitaton in the turbopumps – and both are bad news). For a deep dive into this subject, I cannot recommend Winchell Chung’s Atomic Rockets page on the subject highly enough.
Because space is pretty much the definition of the middle of nowhere, the only thing that really needs to be shielded is the spacecraft itself. To save mass, the easiest thing to do is to stick your nuclear reactor on one end of the ship, your crew quarters on the other, with the fuel tanks in the middle. Then, place a radiation shield between the nuclear reactor and the rest of the ship. This is called a shadow shield, because the ship stays in the shadow of this radiation shield.
What is Radiation, and How is it Shielded?
There are four main types of radiation that come off a nuclear reactor: alpha, beta, and neutron radiation form the group known as particle radiation, and high energy photons like hard UV, x rays, and gamma rays, form the ray portion of the radiation flux. (These, obviously, are ionizing radiation types. Non-ionizing radiation, on the other hand, is not a danger to the crew, and is something to just be dealt with or exploited by the ship – infrared, for instance, also copiously comes off a nuclear reactor, but that heat energy is the entire point of running the thing!) This second type of radiation is made up entirely of photons, but of much higher frequency than visible light. The first type, however, is a salad of different particles: alpha particles are bare helium-4 nuclei (and as such have a charge of +2), beta radiation is a high-energy electron (charge -1), and neutron radiation is made up of – surprise! – neutrons (and as such have no charge)
The easiest way to consider shielding is to split the two types of radiation up and deal with them separately, since they have almost opposite requirements for stopping them. So, let’s look at particles first, and then rays.
Particle radiation is stopped through a process called “elastic scattering,” which is most easily pictured by a pair of balls, one moving and one stationary, hitting each other. Depending on the mass and velocity of each ball, they will reflect off each other, and momentum from the ball that WAS moving gets at least partially transferred into the ball that was stationary. How much is transferred depends on the relative masses of the balls: the closer the masses, the more energy can be transferred. So, to stop any of the particle radiation types, low-atomic-mass (low-Z) materials are ideal, usually something chock full of hydrogen. This lends itself to water, hydrates, and organic materials. However, the atoms in the material will obviously be bounced around, and over time the material will become degraded. As an additional challenge, ray-type radiation will break the hydrocarbon chains that make up organic shielding, and as such will degrade these types of materials even further (for a look more into these effects, check out the organically moderated reactor concept, only think of the challenges of slowing these particles to a stop). The other option doesn’t work for neutron radiation, but works on the other two: electromagnetic confinement. This is the approach used by the concept of a mini-magnetosphere for a ship, being explored by NASA and Rutherford-Appleton Laboratory, and diverts the particles before they come in contact with any materials using powerful electromagnets. This is a very advanced concept, and often is far more massy than using a passive material. In addition, alpha and beta particles aren’t able to leave the reactor’s pressure vessel anyway, so they generally aren’t a concern.
There ARE particles that are a concern, though: neutrons, which are uncharged but slowed and stopped the same way, and galactic cosmic rays, or GCRs. These are higher-Z nuclei that have been ejected from some high energy event, like a supernova, and come tearing through space at a significant percentage of the speed of light. They cause a large amount of damage on the atomic level, and are a major source of the radiation flux that astronauts receive. Unfortunately, because they’re moving so fast they’re virtually impossible to stop or divert, unless you have a strong electromagnetic field blanketing whatever you’re protecting (and even then, because they have so much mass, it’s hard to get them completely diverted, just slowed a bit).
Neutrons are basically unheard of in space, however, so dealing with those is easier: you just have to worry about blocking the reactor from the rest of the ship. This can be done using a number of materials, often very high in hydrogen. During NERVA, a lot of study was put into neutron shielding, and many of the concepts were discovered to be impractical, either due to manufacturing difficulties or material mass, but two stand out: lithium hydride (LiH) and boron carbide (B4C). LiH is the most effective neutron shield per unit mass, and if the lithium is enriched such that only 6Li is used (lithium with an atomic mass of 6), it becomes a very effective neutron shield as well, since it wants to capture another neutron to become 7Li. The downsides are that it doesn’t work nearly as well in high-neutron flux environments, does not conduct heat well, is thermally limited to prevent dissociation of Li and H, and is highly reactive so requires some sort of cladding material to prevent chemical reactions. Boron carbide, on the other hand, is the most effective shield per unit volume, especially when 10B is used (one of the best neutron poisons available). The lack of hydrogen makes it less effective as a moderator of the neutrons that aren’t captured by the boron, though, and it has 20% more mass than a LiH shield of similar shielding characteristics. This is already an off-the-shelf product, though, and the use of 10B will not change these well-established manufacturing procedures, so it remains a very attractive option, especially if smaller, individual shields (spot shields) are needed for individual components, such as the stepping motors used to control any control drums used in the design.
Gamma and X-Ray Shielding
Rays, on the other hand, tend to be simpler to stop – assuming you can handle large amounts of mass! For lower-energy photons, when they come into contact with an atom, they are absorbed by an electron in the electron cloud, which jumps to a higher energy state, then drops down, emitting a slightly lower energy photon in the process. This effect is how neon lights are produced, or how chemicals can be identified by spectral emission. This is where lead shielding comes in for terrestrial reactors (and magnetite-heavy concrete, along with other design features), and lead is commonly used in shadow shield designs for the same reason. However, any high atomic mass element (HZE) can be a reasonably effective shadow shield, and depleted uranium (238U) is sometimes used as a shield for compact reactors due to its greater density and atomic number. The down side to this method of shielding, however, is that it’s heavy, and heavy is the LAST thing that you want on your spaceship. Unfortunately, for complicated reasons I’m not going to get into here, there’s no way to effectively reflect these high-energy photons, so this really is the only way that we are able to deal with them.
Keep in mind, for the main payload (in this case the crew quarters) there is plenty of other mass in the way. This includes the tanks of propellant, the material the tanks are made out of, structural components to transfer the thrust from the engine to the payload without destroying the ship, support equipment… all of these will absorb or reflect radiation to a greater or lesser extent. We’ll look at the propellant and tanks separately, but keep in mind that while the majority of the shielding is provided by the radiation shield, this doesn’t mean that this is the only shielding available.
The exact size and composition of the shield is going to change, depending on the final design of the engines that will be used, but there shouldn’t be a huge variation in the type or quantity of radiation coming off one form of solid core NTR versus another, so only minor tweaks to the shield composition should be necessary. For a good example of the types of changes that may be necessary, an analysis of the Kilopower reactor by McClure and Poston shows how shielding requirements change as fuel type changes [Insert link].
Starting in 2014, a team of researchers at Oregon State University and Marshall Spaceflight Center led by Jarvis Caffery has been examining NASA’s shielding requirements for NTP. In a nutshell, the goal is to REDUCE the overall radiation exposure to the crew for the length of the mission by reducing the overall mission flight time, reducing the crew’s exposure to much more damaging galactic cosmic rays and HZE particle radiation from events like supernovae and neutron star collisions. This doesn’t mean that there aren’t short-term radiation limits that NASA has to work within, or career doses of radiation that are a severe limitation to current mission planners. Given current NASA radiation dose limits, it’s actually impossible to use chemically propelled rockets, because the crew would reach their lifetime dose limit either on the surface of Mars, or on the trip back. NASA is re-examining these limits, and recent legislation that has been proposed to study low-dose radiation exposure may end up significantly changing these requirements in the future.
Caffery et al suggest that in order to maximize the benefit of radiation shielding for the available mass budget, it may be best to concentrate on combined shielding around the crew habitat, to deal with the radiation flux coming off both the reactor and from the environment, rather than concentrating more mass in the shadow shield. However, they also note that using HZE shielding (like tungsten, lead or uranium) near the crew habitat is something to be avoided, since this is how you get brehmstrahlung, and either gamma or x-rays flood your crew cabin.
For the shielding between the reactor and the rest of the ship, this is certainly going to be more than one shield, and the main one is likely to be a composite shield, for a variety of reasons. Various parts of the rocket engine itself will need to be shielded to ensure the more sensitive components are exposed to as-low-as-practicable neutron fluxes. Perhaps the two most important are the stepping motors that will likely be used for the control systems and the turbopumps. Depending on the design that ends up being used, the turbopumps may be on the “hot” side of the main shadow shield, or if a significant part of the shielding occurs in the main structure of the ship and so the shadow shield is reduced in mass, these pumps may be exposed to too high a neutron (or gamma) radiation flux. These can be shielded by secondary shields – in this case possibly B4C, because it not only is a more effective shield per unit volume, but also moderates the neutrons that interact with it less than something rich in hydrogen, leading to lower neutron absorption rates into the mechanical assembly.
The main shield has many options, but there are definite limits to what can be done. Any one concept isn’t going to be good enough, there’s going to need to be a solution that addresses many tradeoffs and problems. This leads to the composite main shadow shield, a concept we’ve seen before in the Kilopower system.
Looking at the Kilopower shield, there are layers of an HZE material (in this case tungsten, but DU is another good option), with thin layers of LiH sandwiched between. This means that the neutron moderation benefits of the LiH – and therefore the likelihood that the neutron will be slowed enough to be absorbed – are spread through the bulk of the shield.
LiH is one of the best neutron shields out there (the best by mass), especially when enriched with Li6, but it has a number of problems, including chemical and thermal stability. Especially in the case of the Li6-enriched variety, a lot of energy will be deposited here as the neutrons are first slowed, and then absorbed, which means that heating can be significant, and unfortunately LiH isn’t the best conductor out there. Dissociation of LiH into lithium metal and H2, which then will either form pockets of gas that weaken the surrounding material, or is lost through outgassing, can occur if the thermal load gets too high.
In order to mitigate this, and also increase the chances of neutron capture and energy deposition in the more thermally conducive HZE shielding plates, the LiH is spread through the shield. This allows for the LiH to be in a small enough sheet to allow for needed thermal dispersion into the more thermally conductive (and less sensitive) metal components. It also means that the secondary gamma emissions from the neutron moderation and capture have plenty of shielding to stop them before they reach the end of the shield.
A design using B4C would have less volume, but more mass. This material is already something that’s commonly used in machine tools all over the world, and even enriching the boron to increase its’ likelihood of absorbing neutrons won’t change those manufacturing techniques significantly. One option studied by McCafferty et al was a pebblebed design, where spheres of B4C would be packed into a casing made out of some structural material. This allows the already better thermal properties of B4C to be maximized, while maintaining the shielding properties of the material by minimizing available ray paths for radiation through the material. Due to its’ higher mass, this material hasn’t been studied as extensively as LiH, but offers some distinct advantages, and so this is was explored more thoroughly in the 2015 paper I linked above. With its’ machinability (and long industrial application), thermal conductivity and resistance, and lower-volume shielding properties, this is a material that will likely show up in many designs, if not necessarily for a main shield then definitely for secondary shields.
While the flux is going to be highest when the reactor is operating, this does not mean that the only radiation flux coming off the reactor is during operation. Once fission occurs in the reactor core, the entire reactor becomes irradiated – the longer it operates, and the higher power it operates at, the more radioactive the partially used fuel and exposed reactor components will become. This means that the highest radiation flux coming off the reactor will likely be in the final burn of the nuclear fuel’s life (and the reactor itself if it’s not designed for refueling), when it will also likely be pretty much exhausted of fuel. This is the worst case that must be designed for, and unfortunately the one most sensitive to decisions that haven’t been made yet.
Unfortunately, at the moment the design for the shield will be up in the air. Until a number of decisions have been finalized in the engine design process, including fuel type, enrichment, neutron spectrum, and others, only options and broad outlines are able to be proposed. Another challenge brought up by the authors is that the primary tools used to model time-dependent dosing calculations, the MCNP code released by Los Alamos National Labs, isn’t exactly the best at these sorts of calculations. Because of this, testing of any shielding system will be needed.
The Propellant Tanks
Any propulsion stage needs propellant to work, and in the case of NTRs the ideal propellant is one of the most difficult to work with: hydrogen. Hydrogen is the lightest of the elements, and as such has a bad habit of being able to seep through just about everything, weakening it in the process. Cryogenic cooling can significantly reduce its’ bulk, but it remains incredibly bulky even under cryogenics, and its’ very low liquefaction point means that maintaining it in cryogenic storage is a major challenge.
This is the hydrogen boil-off problem, and it’s something that has vexed every rocket designer to use LH2 since the beginning of the space age (and many chemical engineers in the decades since it has been discovered). In low Earth orbit (LEO), H2 tends to boil off at predictable rates due to a number of factors which define the quirks of various systems. In addition, as the hydrogen seeps through the structures of the tank and spacecraft, these get weakened and brittle, known as hydrogen embrittlement. Add in the large volume that H2 requires, and this can be one of the most challenging propellants to use in a rocket, and many of the challenges dealt with in the Rover program were actually related to using H2 propellant, which hadn’t been done in the US before.
There are two ways to deal with this problem: first is with a purely passive system, as is done in launch vehicles, and the second is to use an actively cooled system to minimize or eliminate hydrogen boiloff. The first option, unfortunately, isn’t an option for the NTP stage (due to very long mission times), but the passive cooling technology is still used, based on the LH2 tank design for the Space Launch System. This tank is made out a special aluminum/copper/lithium alloy (Al 2195), which is a high-strength, weldable alloy. Currently, new welding techniques are being used with this alloy on the construction of the SLS main tank at NASA’s Michoud facility by Boeing, which will also improve the quality of an NTP propellant tank as well.
Surrounding the main tank is a thermal protection system (TPS), commonly a foam insulation (in the Apollo SIV-B third stage, this was polyethylene foam internal to the tank, which resulted in less than 10% boil-off during the LEO insertion and TMI burn phases of the mission), and a sun shield as well (likely, in the case of a longer-term mission, seen as a gold foil coating on the stage around the propellant tanks). Additional TPS techniques have continued to be investigated, including use of H2 gasses during the boiling being used as a vapor to cool the rest of the TPS through careful venting of the overpressure coming off the H2 tank, and the use of cryocoolers to further cool the thermal shields and mitigate heat transfer. However, it seems unclear exactly which materials would be used for a long-term cryogenic storage TPS for the NTP, and this could be a major problem for such a long-duration mission as a manned Mars mission, which would require the H2 to be maintained for over a year. An example of what has been flown would be the Power Reactant Storage Hydrogen Tank, which lost 2.03% of its’ reactant per day over the 21 day lifetime of the system. This leads to a huge increase in the needed H2 for an extended mission, and a corresponding loss of payload capacity. Even more modern systems lead to a boil-off rate of about 6%/month, which is incredibly prohibitive for as extended mission as a Mars crewed mission.
Nuclear + Rockets: Always Complicated
When I started this project, it seemed (relatively) easy: nuclear power, while complex, isn’t unknowable. Rocket propulsion is far more complicated in so many ways than nuclear thermal rockets: only heat is necessary, not the finicky balance between fuel and oxidizer. Sure, there are a thousand details, but that’s what engineering is for.
There is a truth to this, but one of the simplest systems is a wonderful example of why this subject is so difficult to address. Propellant tanks are, in theory, fairly simple: it’s a fancy thermos, with given rates of boil-off that can be adjusted by improving the insulation of the system. Heat mitigation is primarily needed from the solar environment, which is a task that all spacecraft have to address
With a nuclear reactor, there are two additional vectors for thermal heating, both from the reactor. First, there’s gamma ray heating, caused by the remaining gamma radiation after the primary shadow shield and support equipment. A small amount of this is coming from the fission reactions themselves, but the bulk of the shield will absorb these particles. The larger component comes from the neutron flux coming off the reactor, either through elastic collisions of the neutrons and hydrogen in the tank – which slow the neutron and accelerate the hydrogen, heating it – or through the secondary gamma radiation caused by these collisions. As each neutron is slowed (thermalized), it is more likely to interact with the next atomic nucleus, increasing the number of reactions while reducing the energy of each of those interactions, until the neutron is finally captured. These resulting gamma rays are far more easily absorbed by heavier (higher-Z) atoms than lighter ones, so while it’s more unlikely that they will be absorbed by the propellant, the support structures and tank itself will be heated by these interactions, transferring the heat to the propellant through conduction.
According to a 2015 paper by B.D. Taylor et al of NASA’s Marshall Spaceflight Center, neutron interactions with liquid H2 drop to effectively zero after less than 50 cm of penetration into the tank itself (due to hydrogen’s excellent moderation properties), and gamma heating becomes the major source of nuclear-caused thermal heating after about 15 cm.
Due to the unique nature of the internal heating caused by the radiation flux (rather than the still-present external heating caused by background radiation that is a more well-understood problem), thermal stratification and complex convective cycles are far more likely to develop in an NTR’s propellant tanks. This can be mitigated by careful construction of baffling, and possibly with mixing equipment internal to the tank itself.
Enter Active Cooling: The Zero Boil-Off Tank
If LH2 boiloff was eliminated, not only would your propellant not be leaking away constantly, but the hydrogen embrittlement would be reduced or eliminated as well. In addition, according to some studies the launch mass of a LH2 system could be reduced by 20% or more if boiloff was eliminated for cislunar space missions, between the mass of the H2 and the required larger tankage requirements. While thermal shielding is a huge help (such as the gold foil seen on many spacecraft), the ambient temperature of space is still higher than the boiloff temperature, so active cooling is needed. In addition, the propellant will be warmed by both gamma rays and neutrons that weren’t absorbed by the shadow shield, and so needs to be actively cooled to prevent even faster boiloff. This problem is so severe, in fact, that NASA no longer plans to try and use just passive cooling techniques to get to Mars.
Enter the Zero-Boiloff Tank, a design that NASA began researching in 2006 with the Florida Solar Energy Center. This system uses a multi-stage cryocooler and hydrogen densification system to ensure continuous cooling of the cryo H2. This design started as a small (150 L, due to facility safety regulations) dewar, built at the FSEC, surrounded by a storage vessel. Later tests used a much larger tank, closer to what would be used for a rocket propellant tank, either for a stage or a propellant depot.
This is a system that we’ll go more into depth on in its’ own post, so we’re going to look at it more briefly than we typically do here in the interests of blog post length.
In short, a ZBO tank uses integral cryocoolers to maintain the propellant below the boiling temperature of the H2. This definitely adds dry mass and complexity to the system, but by significantly reducing or eliminating boil-off, the overall mass needed for the system to complete the mission requirements is reduced by a large amount. This can be paired with vapor-cooled shielding and passive TPS to optimize the mass of the system.
This is still a very active area of research, since it directly impacts chemical as well as NTR systems. From the development of a small, desktop breadboard system, to a larger, outdoor system, an on-orbit technology demonstration mission (CRYOTE), and continued research into system components and optimization, much research is still being done to optimize system mass and usability. As such, the final design of the propellant tanks is still very much up in the air.
There is one advantage that the ZBO designs have over traditional tank designs for NTR use: the internal support structure will act as an additional shield down the center line of the spacecraft, protecting the payload more than just the LH2 remaining in the tanks.
LH2 Shielding Goes Away Through the Mission
As propellant is expended during a burn, there will be less mass between the payload and the reactor, meaning that secondary radiation protection will decrease the longer the engines burn.
This is a problem for the payload, because the flux coming off the reactor increases the longer its’ burned (due to fission product decay, ambient delayed neutron flux, and increased reactivity requirements to overcome neutron poisoning in the fuel elements). As mentioned above, the internal structures of a zero boil-off tank mitigate this problem somewhat, but they aren’t so large that they would completely fill the entire center line of the spacecraft between the reactor and the payload. However, there have been some designs that retain a column of H2 in the tanks, even when “empty,” which mitigate this. There is a mass loss if this is done, but depending on acceptable radiation dose to the payload, and the radiation flux coming off the reactor, this may be a good decision for some spacecraft designs, especially smaller ones where the distance between the reactor and the payload bus is smaller than on larger spacecraft (for instance, a lunar shuttle vs. a Mars spacecraft).
LEU NTP Mars Vehicle
The LEU NTP stage’s primary mission is going to be a crewed mission to Mars. This doesn’t mean that the stage can’t be used for other missions, but every project needs a mission, and in this case that mission is NASA’s Mars Design Reference Mission 5.0 with an expected mission date for the 2037 launch window to Mars.
In order to complete this mission, a number of components are going to need to be assembled on-orbit: the core propulsion stage (CPS, containing not only the main engines, but reaction control systems, avionics, a solar electrical power system – no bimodal plans for the basic design – and cryogenic fluid management hardware), an in-line propellant tank (essentially the same as the CPS, but with the engines and shielding replaced with more LH2 tank, and a smaller RCS), a saddle truss with up to 5 LH2 drop tanks (one in-line, the rest attached to the outside of the truss), a smaller saddle truss for payload, a deep space habitat (based on the TransHab design), and an on-orbit manned spacecraft (the Orion module).
The basic design for the propulsion bus hasn’t changed since the design of the Nuclear Cryogenic Propulsion Stage, the immediate predecessor to the LEU NTP stage. One retired DOE engineer of my acquaintance loves to point out “all nuclear design is evolutionary in nature,” and this is one of those times that it clearly shows.
The numbers that are used in this section are based on the HEU version of this stage, optimized for Mars DRM 5.0. They will likely be slightly different, not only due to the new engines but also due to advances in mission design and vehicle optimization, by the time this system is taking its’ first crew to Mars, but they will be very close.
Core Propulsion Stage
Being NASA, NTP design is modular in nature, with the idea being that the same propulsion and power module can be used for multiple mission types by adding additional propellant tankage and support equipment depending on the mission profile and destination. So, a Lunar shuttle may only need the core stage, while extensive additional tankage and payload would be necessary for a Mars mission.
The core propulsion stage for the NCPS (and likely the LEU NTP stage) is approximately 25 meters long and 8.4 meters wide, carries three CERMET-fueled 25 klbf NTP engines (Pewee class) for main ship propulsion, 47.2 metric tons of LH2 propellant, and 15.6 metric tons of reaction control system fuel and oxidizer (NTO/MMH). When launched, it will be fully fueled, with a wet mass of 109.5 mt (dry mass 46.2 mt). This pretty much maxes out the payload capabilities of the Space Launch System, which is the preferred method for lofting the stage into orbit. A composite truss structure provides the structural strength for the stage. The reaction control system is nitrous oxide/monomethel hydrazine hypergolic fueled, based on the Fregat RCS, with 328 s of specific impulse. The main engines are planned to be rated at 900 s isp, but as we’ve seen there are many questions remaining about the actual design of the engine that will be used.
In-Line Propellant Tank
Moving up the spacecraft structure, the next module to be launched will be an in-line LH2 tank (ILT), which at 25.7 m is just slightly longer than the CPS, but with the same diameter. This module is very similar to the CPS, but replacing the engines and shadow shield with additional tank volume. The RCS is also smaller on this stage, since it’s not on the end of the stack and therefore needs to apply less force than the rear-most section of the spacecraft. With a dry mass of 29.7 mt, 79.2 mt of usable LH2, and just over 2 mt of RCS fuel/oxidizer, the total wet mass of this module while on the ground is 108.2 mt – once again, constrained by the capabilities of the Space Launch System. A similar composite truss structure is used on this portion as the CPS, and docking adapters on each end are used to secure this module to the CPS aft, and the saddle truss forward.
Saddle Truss with Drop Tanks
The third portion of the spacecraft is a long (27.8 m) saddle truss, which means that the structural components form a cylinder around a central hollow. In this case, that hollow holds an additional in-line drop tank, another part of the RCS, and has the capability to mount additional external drop tanks (this part of the spacecraft is far enough forward that these will be shielded by the shadow shield). With a total dry mass of 29.75 mt, and a total wet mass of 118.4 mt, this portion of the stack carries a minimum of 84 mt of LH2 propellant. Since this is a drop tank, and will be used for trans-Mars injection burns, the ZBO tank will not be used here, leading to LH2 boiloff of approximately 1.54 mt. Once again, this will take up the full launch capabilities of the SLS, and will be the second-to-last module launched.
The final portion of the spacecraft is the mission payload. In this case, it consists of a smaller saddle truss (containing mission specific payload, an RCS, and a canister for holding cargo, approx 12.14 mt), a fully stocked deep space habitat (TransHab, 51.85 mt fully stocked), and the crewed spacecraft (in this case the Orion spacecraft, but the original design called for the MPCV, Orion’s predecessor, which massed 14.49 mt without fuel). This is the lightest weight of all the launched modules, with 78.8 mt of mass on the pad. It’s possible that this launch may carry additional fuel, but instead it may just take advantage of using a less capable (and therefore less costly) launch vehicle.
The Integrated Stack
While in low Earth orbit, and once fully assembled, the Mars crewed spacecraft will mass approximately 414.15 metric tons, delivered by four launches of the Space Launch System. Once assembled, the crew will be delivered to the spacecraft for the beginning of the trip to Mars. This will be the largest spacecraft ever meant to travel ANYWHERE except in low Earth Orbit, and will only be smaller than the International Space Station.
Another nice thing about this spacecraft is that, because it’s so long, and the mass is well-distributed, it will also be the first to use centrifugal artificial gravity. By rotating it end over end, it is possible to induce 1 gee of centrifugal acceleration after the trans-Mars injection (TMI) burns, and slow the rotation down to 0.38 gee by the time of Mars orbital insertion (MOI). Then, the rotation will be stopped, and the MOI burn will take place.
Variants of this design have been proposed since the middle of the 1900’s, both for pure nuclear thermal and bimodal thermal and electric propulsion. The bimodal variant (named by its’ creator, Stan Borowski at NASA’s Glenn Research Center) is the Copernicus – B, and has a single large Hall thruster mounted on the center of mass of the spacecraft. After TMI, and spacecraft spinup, the electric thruster is activated for the Earth-Mars cruise period, turning around at the midway point (RCS and navigational correction on a spinning spacecraft has been demonstrated before, it’s more difficult but completely doable). This reduces the travel time to Mars significantly over pure NTP, but at the cost of a much more complex reactor system (of a type that the US isn’t currently investigating strongly, although most astronuclear companies have considered the idea, including NASA’s prime contractor for NTP, BWXT), a power conversion system, added heat rejection equipment, and the electrical thrusters and propulsion. This design is more complex, however, and the current contracts for NTP focus heavily on the pure reactor core.
Most mission designs for crewed Mars missions assume more than one vehicle: at least one, often two, NTP powered cargo ships are sent to Mars before the crewed vehicle described here. These cargo missions are usually planned for arriving at Mars, and having systems verification completed, before the manned mission. This does impose an approximately 22 month delay on the manned mission (the time it takes for another launch window to open from Earth to Mars), but on the other hand it ensures that the supplies and resources needed by the astronauts have been delivered safely. These designs use the CPS as described above, as well as the in-line fuel tank, but the additional saddle truss with drop tanks may or may not be necessary, depending on the mass requirements for delivery to Mars and the number of cargo missions. These follow a slower, minimum-energy (Hohmann transfer) TMI profile, whereas the crewed mission will follow a faster transit (both to reduce crew exposure to the interplanetary radiation environment and to maximize surface stay time).
An early (2009) construction plan for a two cargo ship mission (available here, based on the Ares V, the predecessor to the SLS) involved launching two core propulsion stages, to be mounted to two uncrewed cargo ships for a minimum energy transfer to Mars. This involved a total of four launches for the two craft, each of which would have a mass in LEO of about 236 mt. However, based on the launch estimates more recently provided compared to the launch requirements for this version of the mission’s manned vehicle (which requires three launches as opposed to the more recent estimate of four), it is likely that each of these vehicles may require three launches instead of two (the Ares V was designed for 140 mt to LEO, significantly more than the SLS). One other change, however, is that the overall mass of the crewed interplanetary transfer vehicle is only 326 mt, indicating that a significant amount of mass that current plans assume is on the crewed vehicle would be transferred by the cargo missions instead (my guess is that this is because they were planning on 140 mt to LEO for this design study, not 110 mt). These modules would be assembled in LEO before TMI, and until the first burn for leaving Earth orbit, the reactors would not achieve criticality. This makes the reactor effectively radiologically inert, and not a concern to operate around during launch and construction.
These modules could be assembled at the ISS (assuming it’s still around by the time crewed Mars missions are being launched), or independently in LEO. Details on specific construction methods are sketchy, however with extensive experience in multi-module construction on orbit by most international players involved in the ISS, it shouldn’t pose too great of a challenge – just one with many technical details to work out.
LEU NTP: The Latest Plan to Get to Mars
Nuclear thermal propulsion offers the chance to open far more distant places than humanity has ever set foot to human exploration. While it’s theoretically possible to use chemical or electric propulsion, nuclear thermal propulsion offers far higher efficiency than chemical engines, with high thrust making orbital and interplanetary maneuvering far more rapid than the slow but steady burn of electric thrusters.
Currently, NASA’s plans to go to Mars heavily rely on this promising technology, which was demonstrated over 50 years ago (as we saw in part 1). New requirements in the types of fuel that are able to be used have led to major advances in materials engineering, and open up the possibility of using low enriched uranium (as we saw in part 2). By this point, the basic design for the interplanetary spacecraft is (hopefully) clear.
There remain issues to be dealt with, though: First, the engines need to go through a testing regime that will minimize radiological release to the environment, and be demonstrated to be able to be launched safely, and survive a launch failure without causing an environmental disaster or accidental criticality event; second, the core propulsion stages need to not only be launched, but also be used to their maximum effectiveness to get us to Mars. These will comprise the next two blog posts, which research is already well underway on. After that, I hope to address a different popular fuel form, carbide fuels, which offer even higher operating temperatures, and also address the Russian version of NTR, the RD-0410 “twisted ribbon” architecture, which China has also been experimenting with in recent years.
Nuclear Thermal Propulsion
Radiation and Shielding
In-Space Radiation Environment and Crew Quarters Shielding
In-Space Reactor Shielding
Radiation Shielding Materials Containing Hydrogen, Boron, and Nitrogen: Systematic Computational and Experimental Study – Phase I NIAC Final Report; Thibeault et al, Advanced Materials and Processing Branch NASA Langley Research Center, 2012
Auxiliary Support Systems for NTR
Propellant Tanks and Zero Boil-Off
Passive Thermal Protection
Nuclear Cryogenic Propulsion Stage
Ares V Launch Vehicle (Early NCPS and LEU NTP Launch Vehicle)
Space Launch System (Current NASA Super-Heavy Lift Vehicle)
NASA’s Mars Design Reference Mission 5.0 and Associated Considerations
I hope to have these blog posts released in a more timely manner. Unfortunately, these posts often have me searching for weeks for obscure information that is difficult to find even when paper titles and authors are known, and this last year has been more… fulsome with events in my personal life, let’s say. Hopefully, the greatest challenges are now behind me, and I hope to be able to post more frequently.
Unfortunately, with the difficulty in putting out just the blog (and associated pages), the YouTube channel is now on indefinite hold. There are draft scripts for many different videos, which will likely be edited into pages for the site in the coming weeks and months, but I can’t reasonably see myself being able to edit those scripts, record them, and do the video editing, much less the animations required for the scripts, at any point in the near future.
On the bright side, as some of you may have seen, the Facebook group has hit over 100 members! Feel free to come join the conversation if you’re on FB! (At some point I may branch out onto other platforms as well, but for now it’s difficult enough just keeping up with the blog and FB groups!)