NASA and DARPA will Demo a Nuclear Thermal Rocket Engine in Space by 2027

NASA and DARPA will work together to develop and demonstrate a nuclear thermal rocket engine in space by 2027. The Demonstration Rocket is part of the Agile Cislunar Operations (DRACO) programme.

DRACO’s proposed solid core nuclear thermal rocket temperatures could reach almost 5000 degrees fahrenheit, which will need new advanced materials. It will use high-assay low-enriched uranium (HALEU) fuel, rather than highly-enriched uranium, “to have fewer logistical hurdles on its ambitious timeframe”. Highly enriched uranium has many restrictions to prevent diversion to make nuclear weapons.

The plan is to engineer the system so the engine’s fission reaction will only turn on once it is in space. The goal is to test the nuclear thermal rocket engine enabled spacecraft in Earth orbit during 2027.

Nuclear thermal rocket could speed travel in space by about 2 to 5 times and reduce travel time. DARPA says that the nuclear thermal rocket offers a thrust-to-weight ratio 10,000 times greater than electric propulsion and two to five times greater efficiency than in-space chemical propulsion.

Nuclear thermal rocket programs have been on and off again for the past 60 years but no nuclear thermal rocket has made it to space.

24 thoughts on “NASA and DARPA will Demo a Nuclear Thermal Rocket Engine in Space by 2027”

  1. At the risk of displaying severe Dunning-Kruger effect, I have an idea. I commented before on what I see a very desirable fission reactor for space applications,

    So here’s a picture I drew.

    The basic idea is you use a molten salt with nuclear material in it held in a fuel tank, reactor/nozzle. The fuel tank, reactor/nozzle is spun to keep the molten reactor fluid in the container when heated up to a fluid state. In the picture, blue is the proton accelerator, yellow is the fuel and red the particles accelerating out the back. The advantages I see to this is long trusses can be built in space very lite. This means you could use electrostatic forces of low power and very long particle acceleration paths, and low stress on the truss, to end up with very high energies to start the fission. The reactor fluid itself could be used to capture neutrons, so reduce shielding requirements.

    Possible foibles.
    In the link on accelerator-driven subcritical reactor (ADSR) they use tungsten or lead as a spallation material for neutrons. Could Thorium fluoride salts be used? Or could the reactor fuel be lead mixed with uranium like Russian reactors? The goal I would think would be to have the fuel mixture act also as the spallation material for neutrons.

    Could the reactor bubble like boiling water throwing the reactor fuel out of the rotating container? Higher rotating speeds would maybe help this to not happen. In fact I’m not so sure how the Fission Fragment Rocket would not do something similar. Neutron embrittlement, eventually breaking off pieces of the aerogel and flinging them out the rocket. If the aerogel could work without this happening, then I would think a molten salt would also work.

    Another problem would be that all the lighter elements would be forced to the top because of centrifugal forces, causing the accelerated protons to not interact with the active part of the fuel. A possible fix would be to have another container higher, towards the center of the rotating fuel/nozzle, in which fuel is pumped from the far outer part of the container, (where the heavy elements would separate to). Any excess from the pump could just spill over into the main fuel holding tank.

    If you could use Thorium I would think this would be a great advantage. Start with a not so radioactive material when launched and react it in the drive.

    If this looks like it will work, I would like credit by calling it “Sam’s particle accelerated fission rotorocket”

    I forgot to add I would really like to see criticism of this idea, other than go kill yourself. Is it even in the ballpark of working?

    I really would like some criticism of this idea. One thing I do not understand is why does a fission-fragment-rocket have to be dust or aerogel. Is it because otherwise it would cause sputtering of the fuel. Like water bubbling in a pot sometimes flings out water with the steam if vigorously boiling?

    If you could use magnetic fields to contain the exhaust like the dust based fission fragment rockets you could, as they did get, power and have thrust with the same structure. Magsail? Another advantage of this is the molten salt could act as a heat barrier, as only the top of it would be reacting. Make the fluid deep enough and the temperature would be lowered enough by the time the heat soaked through to use less expensive casing materials.

  2. Even with DARPA military regulations, getting their hands on HALEU is stupid. Almost no substantial domestic production capability. All the terrestrial new nuclear designs that relied on HALEU for cheapness (because it was less of a regulatory hassle than HEU, and because they could source it cheaply from russia) got burnt by that. NASA wants a HALEU design because they don’t have access to the HEU regulatory bypasses the military has. They know they should be using HEU from a performance basis.

  3. 2027 may seem ambiguous to some, but not to me. We already built and tested NTR, see project NERVA. What we need to do now is modernize the NTR. And design it for mass production. We can get about a factor of four improvements on interplanetary launches. It would make manned missions to Mars much easier and cheaper.

    While I like the NTR, I would prefer nuclear power coupled to ion drives for orbiting missions to the outer planets. Slow and steady wins the race.

  4. Hi Brian
    NTR’s have a certain appeal, but they’re not the most effective use of nuclear power for rocketry if one is extracting off-Earth resources. Say you’re sourcing the propellant via electrolysis of water – using just the LH2 requires throwing away 88% of the propellant extracted.

    The only place in the Solar System with easily extracted H2 is Titan. Even at 0.1% abundance it’s a massive potential resource. If you want NTR’s flying all over the place.

  5. NB: For anyone new to the area, the reason you get twice the ISP with a project NERVA type nuclear rocket, despite not being able to match the temperature of old fashioned chemical rockets without melting your reactor, is that a pure hydrogen exhaust will give a much higher exhaust velocity at the same temperature, just because a H atom is lighter than a H2O molecule.

    So no good running water through the nuclear rocket. I mean it’ll work, but it won’t give the ISP of even a hydrolox chemical rocket, let alone a hydrogen NERVA one.

    If you DID want a more tractable propellant than H2, you still want something super light.
    He? Even worse than H2 for transport and storage.
    Li? Hmmmm…… Boils at 1342 degrees. That’s not TOTALLY impossible. I dont’ know how the ISP looks.
    Be? Nope. Just no.

    • This is why the hate against SLS needs to stifle.

      It gave rise to Clipper and Artemis that also helps Starship.

      SLS can launch hydrogen where Musk just won’t.

      Starship for supplies in the slow-boat approach. SLS for cyclers and nuclear drives. As Starship finds its footing, SLS ends its life as high energy upper stage/nuclear probes.

      • Think BUS. Any rocket company (including Musk’s) can lob liquid hydrogen to orbit. Not for its own propulsion, but as a payload. Trick is, getting a (a) large enough, (b) strong enough, (c) insulating enough, (d) light weight enough and (e) cheap enough Dewar on top of a launcher to make it to orbit, with enough liquid hydrogen to be of use.

        After that, launch a number of compositing shots … a NERVA styled reactor and business, a space-frame (truss) type holding platform, maybe 4 to 20 blobs of LH₂ in Dewars, another wiring harness full of chips and cameras, sensors and -ometers, and have both robotic and manned missions to LEO (low earth orbit) to assemble the tinker-toys into a fully functional interplanetary school bus.

        Musk’s brilliance in having fully-reusable launchers fits this perfectly. Even the aggregate cost of 20 missions is still a tiny fraction of a giant NASA-Apollo launch tub. In 2020…

        Anyway, the NERVA concept — at least as hasn’t much changed in 2023 — still is the best cobble-it-together option at this point. A pure graphite core can easily withstand 2,800° C (5,000° F ) nearly indefinitely. Having a modest flow of core-cooling propellant whiz through is pretty ordinary stuff. The exit cone isn’t even super-duper high tech. Doesn’t need to be. Diamond and graphite do all the heavy mechanical lifting.

        Getting an ISP of over 800 remains daunting, but ‘doable’.
        So do it already.
        Space Bus, here we come!


        • Problem with a solid graphite core is it has no uranium in it. If it did have uranium particles in it, that is where the heat would originate.

          Now if you imagine a rocket motor with a throat diameter of a coin and 30lbf thrust, then a “modest flow of core-cooling propellant whiz through is pretty ordinary stuff” indeed. That’s not what NERVA was or what the DARPA would be.

          • So GoatGuy,

            Yes a lot of the problems disappear if you bring NTP down from gigawatt to megawatt class. At low enough power the core becomes basically isothermal and a lot of mechanical issues are solved. Could probably even use a PD pump at that point.

          • Your followup comment captured my thought. Basically, once in LEO, all the need for high-enough-to-escape-Earth’s-gravity business goes away. So, having a porous graphite core plus graphite everything touching the hydrogen … makes it all work

  6. I’ve no doubt that we could have had one of these working in orbit in 5 years time, if we’d done it back in the 1970’s. In the early 1970’s they had flight ready hardware all set to loft.

    I have severe doubts that today’s US government can pull off anything like that in that short of a time frame.

    I mean, just the fact that they’re developing it from scratch, instead of dusting off the plans from NERVA and building an engine already proven to work is a bad sign.

    • They didn’t do NTP space missions in the ’70s because it’s not worth the trouble.

      These russian nuclear propelled torpedoes and missiles don’t fill some niche or gap… they’re not unusable per se, but they don’t provide a tangible tactical advantage…. they’re more like psychological terror weapons like the V1 buzz bomb.

      From the article: “which will need new advanced materials.”

      Read between the lines. That means they can’t actually get close to the operating point that would make it worthwhile. They can’t run it white hot like an incandescent light bulb filament, so why bother….?

      Go ahead and do it, but it’ll be a disappointing turd performance wise. The nuclear fog rolling in gets all of the fanboys impressed. They always forget that nuclear fission is just hot rocks.

      • The NERVA ground tests were running to over 700 seconds ISP at ground level, extrapolated to about 840 in vacuum. That’s real tests on a test stand.

        Sure, they never got to fly any of the hardware, the program got canceled before that happened, but we know NERVA worked, at about twice the ISP achievable with chemical propellants. And twice the ISP is a huge difference.

      • A 1.5 radial peak and 2.0 total peak would be a job well done in a core the size of a washing machine. A reasonable design would attempt to obtain uniform flow through the fluid channels of the fuel elements or frit because that would be more predictable than orificing lower powered channels…. IOW, live with the peak in uniform coolant flow field like all other single pass cores (like PWR). The temperature peaks in the center of a fuel particle, may be hundreds of degrees higher than the surface temperature. So needing to keep the centerline fuel temperature of the hot spot below 5000F while it operates at 150-200% of average power means the average exhaust temperature of the rocket will be hundreds if not a 1000F colder than combustion of HYDROLOX. In reality, the NTP cannot achieve its theoretical performance written on the napkin in the NASA cafeteria. If it worked, it would already be in use.

        • And I do understand that 800s is double the ISP of hydrolox. Doesn’t wash out the difficulties I highlight. Tell me how you’re gonna get giant tanks of bulky hydrogen into orbit for this unshielded rocket bomb to heat. Typical space nonsense daydreams.

          • Do we not already get tanks of bulky hydrogen into orbit for the hydrolox rockets that have been operating for a human lifetime?

          • The difference between 400 seconds and 800 seconds is the difference between a rocket that can go to Mars, and a rocket tug that can fling your Mars rocket towards Mars, and then turn around and return to Earth to refuel and do it again the next day.

            Dramatically increasing the payload to Mars.

            • I got it. Efficiency is king. That’s why all the general contractors drive their tools to the construction site in Hyundai Elantras that get 40 MPG instead of Silverados that get 20 MPG. Oh wait… they don’t do that.

  7. No doubt some hot spot will be near or greater than 5000F if you want to realize any benefit in ISP.

    “The plan” is to engineer the system so the engine’s fission reaction will only turn on once it is in space. Even Homer would include that in his plan, because there is a non-trivial probability for it to enter the ocean sooner or later.

    • Where the uranium will have come from in the first place. Uranium became a renewable resource after extraction from seawater was demonstrated as practical. The extractable uranium content in seawater is sufficient to power human civilization indefinitely since it’s replaced by leaching from new rock that emerges from the mantle.

      • My point was that if it gets dunked in the ocean, it would flood and turn on unless there was an absorbing insert (like control rod(s)) to keep it subcritcal in water.

Comments are closed.