Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Clean lithium fission saltwater rocket

  1. Mar 23, 2016 #1
    Had been talking NSWRs on a spaceflight forum and a thought occurred to me. Lithium-6 fission can be triggered with relatively low-energy neutrons and releases 4.78 MeV, a helium-4 atom, and a tritium atom. Without a neutron flux, however, lithium-6 is completely stable.

    With a small electrically powered neutron source and a magnetic nozzle, a lithium-hydride-fueled thruster could have specific impulse approaching 3% of the speed of light, making it an extremely promising low-thrust engine. But that's not what I'm interested in.

    In order to generate significant thrust, you need a high neutron flux...on the order of what you get inside a nuclear reactor. So why not?

    Natural uranium cannot sustain a chain reaction without a neutron moderator. Thus, a mass of natural uranium inside a tungsten vessel (for neutron reflection and for containment) is quite safe. Add a neutron moderator, however, and you'll get a chain reaction and very high neutron flux.

    Lithium hydroxide dissolves in water at 129 g/L. If you dissolved enriched lithium-6 hydroxide in heavy water, you would have a propellant with a density of 1,239 kg/m3 which, when pumped through a natural uranium cylinder, would immediately moderate neutrons, trigger a neutron flux, and undergo clean fission from that neutron flux. The reaction is inherently self-limiting, as the water is converted to steam by the lithium fission energy and thus stops moderating the neutrons. Such a propellant mixture has a theoretical exhaust velocity on the order of 2,000 km/s.

    The propellant serves as the reactor coolant, the moderator, AND the fuel source:


    This design avoids the problems with a nuclear thermal rocket because the reactor pile only needs to get excited enough to generate a high neutron flux, not generate the energy to heat the propellant. It also avoids the "giant nuclear bomb tank" and "fissioning radioactive death spray" problems inherent in an enriched uranium saltwater rocket a la Zubrin. There are also no moving parts anywhere close to the nasty bits.

    If water proves too corrosive, then you could use lithium-6 hydroxide dissolved in ethanol with a graphite suspension. Varying levels of coolant performance, neutron moderation, thrust, and specific impulse could be achieved by mixing pure heavy water or pure light water in with the saturated saltwater; this could be done dynamically at the turbopump since none of the propellant components are particularly dangerous.

    I foresee a few major challenges. I chose natural uranium since it is cheap and safe, but it may require a very high mass to achieve moderated criticality, which would put a high threshold on engine size and thrust. A different fissile neutron source might be better; I don't know how far down you can scale something like the HFIR. Also, you don't really want your propellant directly in contact with naked fissile material, so you'd want to have the fissile material separated from the propellant by some material that is neutron-transparent and has high strength and heat resistance but is also thermally conductive. Refueling the fissile mass would be tricky if you wanted to reuse, though I could conceive of a replaceable fuel rod arrangement.

    The exhaust isn't entirely safe; it releases a good bit of tritium. But that's far, far better than the nucleotides released by a uranium NSWR. Basically, you don't want to be anywhere near this exhaust stream (not the least of which because it will melt any launch pad) but it shouldn't be prohibitive. It may be useful to include a supercharger to mix atmospheric air in with the exhaust stream to increase thrust and reduce the energy level of the exhaust.

    An overly-optimistic discussion of a lithium saltwater rocket was discussed https://www.linkedin.com/pulse/20140724165847-39571567-nuclear-salt-water-rockets-revisited [Broken], but with the unrealistic expectation that Jetter-cycle D-T fusion would take place within the exhaust stream and that a lithium deuteride suspension would be used rather than dissolved lithium hydroxide. This is unnecessary; lithium fission provides plenty of energy on its own.
    Last edited by a moderator: May 7, 2017
  2. jcsd
  3. Mar 23, 2016 #2


    User Avatar
    2017 Award

    Staff: Mentor

    You get a really hot plasma that will evaporate the uranium. Unless you make the lithium density so extremely low that fission is negligible. Both options don't seem favorable. Remember: your reactor has to be supercritical to account for neutrons lost to Li6 fission.

    Nozzles won't work properly at those exhaust speeds.
  4. Mar 23, 2016 #3
    I like the idea but I don't want to be around when someone tests it.
  5. Mar 23, 2016 #4


    User Avatar
    Staff Emeritus
    Science Advisor

    That seems rather speculative, especially in the absence of engineering calculations. What does one mean by 'electrically powered neutron source'.

    Neutron flux determines power level in conjunction with the fissile atom density. More importantly, one needs a high mass flow rate and exhaust velocity.

    Natural U requires a moderator with a low absorption coefficient, e.g., heavy water. Natural U would be impractical for a small core.

    Direct thrust or so-called nuclear thermal rockets typically contain highly enriched fuel in a carbide or nitride matrix. Liquid hydrogen provides some moderation, but as it heats, it is less effective as a moderator. The core mass flow rate must match the nozzle exit mass flow rate.
  6. Mar 24, 2016 #5
    I depicted a conventional de Laval nozzle for convenience, but you're completely right; I would almost definitely need a long cylindrical reaction chamber with a convex-section diverging nozzle. It won't have the near-perfect efficiency achievable by a properly choked converging-diverging nozzle, but nozzle design here will be less about efficiency and more about not blowing up.

    Erosion/evaporation of the uranium is a big no-no. That's why I said "You don't really want your propellant directly in contact with naked fissile material, so you'd want to have the fissile material separated from the propellant by some material that is neutron-transparent and has high strength and heat resistance but is also thermally conductive." Upon review, this could simply be a thinner layer of tungsten; it should let enough neutrons through if it's not very thick. Other materials could potentially be better, though. Erosion and heat flux can also be mitigated by using a boundary layer of unsalted water.

    Using a convex-section diverging combustion chamber as the nozzle will focus the highest neutron flux near the nozzle throat, which will further reduce heat flux within the chamber and keep the point of highest fission rate well away from the point of highest chamber stress:


    Lithium-6 hydride salt has a specific energy of fission of 6.6e13 Joules/kg; corresponding to a theoretical maximum exhaust velocity of 11,500 km/s which is just under 4% of c. Of course this is more particle physics than nuclear engineering and is most definitely a first-order approximation at best; the actual exhaust velocity would not be nearly that high.

    The neutron source could be any sort of applicable neutron generator, from a tiny neutristor to a photoneutron generator to a D-T accelerator.

    Lithium hydride is already an ionic compound, so a thruster would probably vaporize the salt and then use a very strong magnetic field to contain the lithium ions. Neutrons would then be fired into the cloud of ions; the system would need to be designed in such a way that only fission products would have enough kinetic energy to escape out the "nozzle end" of the magnetic confinement region.

    I suggested natural uranium because it is safe (it cannot go critical without a moderator), cheap (compared to enriched fissile material), and less concerning in terms of proliferation. The enriched lithium saltwater would thus need to use heavy water to provide a moderator; the moderator being the reaction mass adds another level of inherent safety. But it's very possible that a different fuel would be better for building a lightweight, safe, high-neutron-flux reactor.
  7. Mar 24, 2016 #6


    User Avatar
    Staff Emeritus
    Science Advisor

    What is the basis of the particular value for specific energy.
    Actually, it is nuclear engineering. Particle physics has little to do with the problem of using a solution of lithium as a propellant.

    Lithium hydride, or lithium hydroxide? Using only the products of ostensibly a (n,α) would be rather inefficient, and would require a way to remove a bulk of the lithium solution. A neutron flux/beam in a vapor or gas would produce a low reaction rate. A lithium solution as moderator means depleting the moderator, which means a large reservoir of moderator. Expelling the HW moderator as propellant would be an expensive proposition. It seems that the concept as proposed is rather impractical.

    Graphite moderation with high enriched U or U,Pu is the most practical approach to a NTR core.
  8. Mar 24, 2016 #7
    4.78 MeV / 7.023 amu = 6.567e13 J/kg.

    Definitely. I was initially pointing out that neutron flux on pure lithium hydride could produce a low-power ion thruster with a stupidly high exhaust velocity, but I also that that wasn't what I was primarily interested in. More interested in the lithium-fission NSWR, which will have a much lower exhaust velocity (but still much higher than chemical propellants) and very high thrust.

    This is not a nuclear thermal rocket (NTR); this is a nuclear saltwater rocket. The fissioning material and the moderator are expelled together as reaction mass; you can afford to do this because the specific energy is so high (in my proposal, enriched lithium-6-hydroxide-saturated heavy water, the specific energy is 1.94e12 J/kg).

    Nuclear saltwater rockets have been proposed before (specifically Zubrin's in the early 90s) but they are highly unfeasible because they use enriched uranium saltwater that is mixed in the reaction chamber to form a critical mass. The solution must be stored in thin boron tubes to avoid predetonation. My proposal allows the use of non-radioactive, non-volatile fuel which is activated by a lower-temperature high-flux reactor.
  9. Mar 24, 2016 #8


    User Avatar
    2017 Award

    Staff: Mentor

    Your lithium fuel environment has to be dense enough to thermalize neutrons, so it also will thermalize the fission products.

    It does not matter which solid material you use. Ions impacting with MeV energies impacts will always erode the surface. A liquid (and evaporating) material might work, but probably reduce thrust per reaction mass massively.
  10. Mar 24, 2016 #9
    I absolutely want the fission products to be thermalized. MeV energies are far, far too high to handle; distributing this energy to the oxygen and deuterium in the heavy saltwater reduces the velocities well out of the relativistic range while massively increasing the momentum and therefore the thrust.

    Adding the annular unsalted-water flow will do the same, increasing thrust-specific propellant consumption (because the unsalted heavy water now counts as part of the propellant) but significantly increasing thrust/weight. We are looking at a specific impulse thousands of times greater than the best chemical rockets and hundreds of times greater than the theoretical limits of solid core NTRs. Diluting the propellant mixture with more reaction mass is not going to be a problem.
  11. Mar 24, 2016 #10


    User Avatar
    2017 Award

    Staff: Mentor

    Well, not any more. You have a factor of 3000 for direct fission exhaust, if you thermalize your fission products to have some nice temperature gradient (solid uranium at the side, superhot plasma in the center) you'll lose a large part of that. Sure, a factor 10-100 would still be a massive improvement, and make missions in the solar system much easier and faster if everything else works out.
  12. Mar 25, 2016 #11
    The ion thruster version using nothing but pure lithium-6 has a maximum theoretical specific impulse of 1.17 million seconds. The lithium hydride ion thruster has a maximum theoretical specific impulse of 1.09 million seconds.

    Cutting it down to the high-thrust versions, a saturated lithium-6 hydroxide heavy saltwater application would have a theoretical specific impulse on the order of 200,000 seconds. But we will definitely need an unsalted heavy water boundary layer to protect the combustion chamber and diverging nozzle while also helping to moderate fast neutrons. The resulting loss of specific impulse (and corresponding increase in T/W ratio) will depend on the ratio of salted water to unsalted water. At a 1:1 ratio, specific impulse drops to 142,000 seconds; at a 1:4 ratio it drops to 91,200 seconds. But this is still 150-200 times the specific impulse of the most efficient chemical rockets, and the thrust would make Shuttle SRBs look like model rockets.

    On the other hand, the use of a neutron flux reactor with a more enriched fuel type would require less moderation and could increase specific impulse over that previously noted, up to a theoretical maximum of 210,000 seconds.

    I think these are manageable numbers. If we can get rid of the tritium release problem then this would make it trivial for even small vehicles to achieve orbit and come back with fuel to spare, while also enabling partial brachistochrone trajectories for nearby interplanetary destinations.
  13. Mar 26, 2016 #12
    Engineering the propellant stream to pass around the reactor on its way out rather than through it could serve to partially shield the rest of the ship from neutron flux, as the lithium would suck up a lot of the neutrons.

    I wonder if adding sulfur or chlorine or fluorine or some other compound to the propellant would cause it to react with the tritium produced and keep it from being biologically active for long enough that it would decay.
  14. Mar 26, 2016 #13


    User Avatar
    Staff Emeritus
    Science Advisor

    What concentration in solution is one considering, what isotopic ratio (Li-6/(Li-6+Li-7)), and what energy spectrum. Look at the solubility of LiOH in H2O. One could assume a super-saturated solution. Depending on the concentration and how fast the coolant is passed through the core, one may find a low fraction of Li-6 will fission. One will have to determine the time of transit through the neutron field and integrated of the (n,α) over the energy spectrum and time to determine what fraction of Li-6 will fission between introduction to and exit from the core.

    Li hydride (LiH) was mentioned (that was probably incorrect), however its melting point is 692°C, and it is very reactive with water, especially as temperature increases.

    In what environment does one propose to deploy such a system? The earth's atmosphere, or space? If one fissions Li-6 with a neutron in an aqueous propellant, one will not be separating the tritium from the helium by the time the solution is exhausted.
  15. Mar 26, 2016 #14
    It would be entirely possible to keep the lithium metal or lithium salt in a solid form and simply react it with water as needed, but then you end up pumping hot gas into the reactor core rather than liquid, and that's problematic. Not that it isn't going to flash to gas and then to plasma soon enough, but you kind of want it in liquid form when you are pumping it about.

    The length of the reaction chamber, the variance in neutron flux, the specific salt used...it would all have to be sorted out. I would assume 95%+ enriched lithium-6; don't know how much could be made to fission before it left. There will be a scaling thing, where the larger the engine is, the more efficient it will be.

    Lithium hydroxide has very nice solubility in water, at 129 g/L, but there are better options. Lithium chloride is 883 g/L and ends up being twice as much lithium by weight per unit volume. There is also a possibility that another solvent would allow greater mass percentage of lithium-6.

    I was thinking of including the compound either in the coolant stream or in the main fuel stream, so that free tritium in the exhaust ends up reacting with that compound rather than being generally released. Might reduce tritium "fallout" enough that they would allow it.
  16. Mar 26, 2016 #15


    User Avatar
    Staff Emeritus
    Science Advisor

    This is somewhat faulty logic. Whatever tritium is produced in the propellant stream will be exhausted, so it doesn't really matter whether it's combined with S, Cl or F, which would all dissociate in an aqueous solution, especially one heated to several 100 C.

    It is important to calculate the probability of inducing an (n,α) reaction in a Li-6 nuclide from the time it enters the neutron flux to the time it leaves. If it is something like 0.1, the approach may be questionable, but if it is something like 10-3 or 10-4, then the concept would be impractical.

    One would not produce much of a plasma, if by plasma one means complete ionization of Li (and other atoms), otherwise, one may find a plasma being ions (cations and anions) in a water vapor. In LOX + H2, rich in H2, the exhaust is basically H2 and H2O.
  17. Mar 27, 2016 #16


    User Avatar
    2017 Award

    Staff: Mentor

    You won't be able to bind the tritium chemically for years. In water, hydrogen atoms get swapped between the molecules on a timescale of milliseconds to seconds. Some chemical bonds can lead to larger timescales, but tritium will get in the environment if you want to use the rocket in the atmosphere. Anyway, it won't work, see below.

    As a rough approximation, the cross-section for Li6(n,##\alpha##)t is 1 to 10 barn for MeV to 100 eV neutrons, increasing to 1kbarn for thermal neutrons (source). Let's take the 1 kbarn value as upper estimate.
    If we want to shoot out fuel at 10 km/s (slow) with a reactor length of 10 meter (long), and want to get 1% fission efficiency (low), we need a fission timescale of 100 milliseconds. That needs a neutron flux of 1022/(cm2 s). Oops. That is about the neutron flux you get in a supernova. Nuclear reactors are 11 to 12 orders of magnitude below that. If you just get a fission rate of 10-13 to 10-14, the concept does not work.
  18. Mar 27, 2016 #17

    What about increasing the path length of the lithium? I know the High Flux Isotope reactor produces a flux of roughly 1e12 neutrons per second per square cm, but it is intentionally poisoned to keep the flux variance as low as possible. I wonder whether it would be possible to set up a helical flow path to give the lithium fuel a much larger time under high flux.

    Neutron reflection is a very picky and inefficient thing but it is still conceivable that the reactor exhaust channel could be shaped in such a way as to focus the region of maximum flux to be at least a few orders of magnitude higher.

    The total free neutron output of the HFIR (if the neutron poisons were removed) is on the order of 1e20 neutrons per second, IIRC. While 100% neutron capture is of course extremely unrealistic, this can at least give us a means of estimating upper-bound power output. 1e20 s-1 times 4.8 MeV is 77 MW; at an arbitrarily assumed 10% fission, this would be a mass flow of 38 mg/sec of lithium hydroxide. With fully saturated saltwater in a 1:4 ratio with moderating coolant, that's 1.8 grams per second. With 40% Carnot efficiency that would be an exhaust velocity of 185 km/s and a thrust of 333 Newtons.

    Gives us ballpark numbers to work with, I suppose.

    EDIT: If it simply wouldn't be possible to get a high enough neutron flux to derive all the energy required, this could still potentially have promise as an accelerated adaptation to a standard NTR. The lithium fission would absorb a lot of the dangerous neutron flux from a typical NTR, enabling the use of less shielding, as well as generate additional energy so the core could run cooler with higher exhaust velocities.
    Last edited: Mar 27, 2016
  19. Mar 27, 2016 #18


    User Avatar
    Staff Emeritus
    Science Advisor

    The 10 m length cited by mfb is pretty long already, and that doesn't work, primarily due to the high flow rate.
    One cannot simply take 40% of the energy from the nuclear reaction, but one must use the product of the energy of a reaction and the reaction rate, and the reaction rate per unit volume is very small.

    Commercial power reactors have fluxes on the order of 1013 n/cm2-s.

    The lithium-6 will not absorb a lot of neutrons.
  20. Mar 27, 2016 #19
    What about using some kind of neutron multiplier that would be consumed by the reaction?

    A sufficiently active layer or layer set of neutron multipliers might even make a D-T fusion deuterium ion accelerator design feasible.
  21. Mar 28, 2016 #20


    User Avatar
    2017 Award

    Staff: Mentor

    The required neutron flux would make your uranium explode in milliseconds. You cannot both use uranium for neutron production with a low fission rate and lithium for neutron absorption with a high fission rate. It just does not work.
  22. Mar 28, 2016 #21
    Did a bit more digging.

    Quoting from this paper:
    The paper goes on to discuss the high probability of reaching fluxes in excess of 1017 neutron/cm2*sec. It seems likely that with the use of molten salt reactors, pebble-bed reactors, or other more specialized designs, further improvements to this neutron density could be achieved.

    Of course, even 1018 or 1019 neutrons/cm2*sec falls a bit short of the 1022 figure derived by mfb above. But it is a lot closer, and could open the door for alternative geometries to allow high lithium fission levels.

    I was thinking less along the lines of uranium and more along the lines of an accelerator-to-target design. I don't know whether there are any large isotopes which can be triggered to release numerous thermal neutrons at once when struck by a fast fusion neutron, but if so, then these could be used to multiply the yield of a D-T accelerator. Another possibility would be spallation; spallation sources routinely produce beam fluxes on the order of 1017 neutron/cm2*sec. If these could be multiplied (perhaps focusing the proton beam onto an initial mercury target with a tantalum or depleted uranium cylinder at the center, as shown in the attachments) by a few orders of magnitude, it could become feasible to create the desired neutron flux.

    spallation 1.png spallation 2.png
  23. Mar 29, 2016 #22
    Discussion on forum.nasaspaceflight.com already focused on the main problem: you need a neutron source which is too good to be true. But do not despair. You stumbled over one of the most promising engine ideas for interstellar travel - a fission fragment rocket. The problem is, how to let some significantly nonzero fraction of fission fragments leave engine and fly as exhaust without damaging the engine, and without much thermalization? (Any fission will do, no need to be fixated on Li).

  24. Mar 30, 2016 #23

    Are there any fissionable but non-radioactive isotopes with a net positive prompt or delayed neutron output? Lithium-7, of course, fissions endothermically and releases a thermal neutron when struck by a very fast neutron, but that's just break-even.

    Barring that, which fissile isotope has the highest total neutron multiplication factor? Even if a pure lithium fission saltwater rocket isn't possible, it could be possible to run a subcritical lithium-salted NSWR with a far better ISP due to lithium-6's higher specific energy compared to the heavy radioactive isotopes. It would also be safer since the fuel would never be critical on its own.

    For an accelerator-based design, I wonder whether lithium+deuterium fusion (without fission) to 2 He-4 would hold any promise. You can store the fuel easily (lithium deuteride), accelerate the ions directly, and there is no neutron radiation or radioactive exhaust. And the specific energy is four times higher than lithium-6 fission.
  25. Mar 30, 2016 #24


    User Avatar
    2017 Award

    Staff: Mentor

    Shooting Li into D or vice versa mainly produces a lot of heat from thermalization. Fission reactions are really rare, and there is nothing you can do about it.
    Accelerator-driven DT fusion doesn't work as power plant for the same reason. Accelerator-driven fusion gives quite compact neutron sources, but no relevant energy release.
  26. Mar 31, 2016 #25
    Probably sound like I'm clutching at straws here, but I'm honestly just curious about what the possibilities are...

    Are there any fission events or decay chains which produce deuterium? 2He and 6He both have a low probability of decaying to deuterium, but that's all I've found. Alpha decay, beta decay, and proton decay all seem to be a thing, but deuteron decay doesn't seem to happen...or am I wrong?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted