There are some potentially valid criticism about the Nuclear Salt Water Rocket (NSWR). The criticisms highlight real physical constraints. A hydrogenous mixture will moderate neutrons, preventing a fast-spectrum detonation and likely limiting the reaction to a thermal fizz. However, Zubrin’s NSWR isn’t necessarily aiming for a weapons-grade explosion but rather a controlled, continuous supercritical reaction. The term “detonation” might be a misnomer—Zubrin could be describing a rapid, sustained fission process that’s more energetic than a typical reactor but less violent than a bomb. Still, the neutron moderation issue undermines the efficiency and thrust potential unless addressed.
Criticism 1: Thermal Neutrons and “Detonation”
The critic argues that the neutrons in a hydrogenous uranium mixture (like uranium bromide in water) would be thermalized (slowed to energies around 0.025 eV) due to moderation by hydrogen atoms in the water. Thermal neutrons are excellent for sustaining fission in reactors like TRIGA or thermal molten salt reactors (MSRs), but they don’t produce an explosive “detonation” in the traditional sense. Instead, the reaction would self-limit (like a TRIGA reactor excursion), fizzling out rather than producing a sustained high-energy detonation.
This is a valid point. In a hydrogen-rich medium, neutrons scatter off hydrogen nuclei and lose energy rapidly, becoming thermal. For an explosive yield (a “boom”), you need fast neutrons (around 1-2 MeV) to drive a rapid, supercritical chain reaction, as seen in nuclear weapons or fast-spectrum reactors. A thermalized neutron population would lead to a slower, self-regulating reaction, not the continuous detonation Zubrin envisions. The result might be more akin to a “dirty water vapor” exhaust with a lower Isp, closer to NTP (Isp ~800-1000 seconds) than the NSWR’s theoretical Isp of 10,000+ seconds.
Criticism 2: Fluorine and UF6
The critic also mentions uranium hexafluoride (UF6) and fluorine’s neutron interaction properties. Fluorine has a significant scattering cross-section for neutrons in the 0.5-1 MeV range, which could thermalize them, making it less suitable for a fast-spectrum reaction. This is why fluoride-based molten salt reactors (e.g., using LiF or FLiBe) operate in the thermal spectrum, while molten chloride salts are being explored for fast reactors. This critique suggests that even alternative formulations (like UF6 instead of uranium bromide) wouldn’t solve the fast-neutron problem.
This is also valid. Fluorine’s neutron moderation properties would hinder a fast-spectrum detonation, reinforcing the challenge of achieving the NSWR’s intended performance with hydrogenous or fluorine-containing mixtures.
Potential Solutions to Ensure Criticality and Efficiency
To make the NSWR viable, even with reduced efficiency, several modifications could be explored:
Minimize Neutron Moderation:
Replace water with a low-moderating solvent or carrier (e.g., a heavy-metal liquid like lead or bismuth, though solubility of uranium salts becomes an issue). This would preserve fast neutrons, increasing the reaction rate.
Use a deuterated solution (heavy water, D2O) instead of H2O. Deuterium moderates neutrons less effectively than hydrogen, allowing a spectrum closer to intermediate or fast, though this still wouldn’t fully achieve a 2 MeV neutron population.
Increase Fissile Concentration or Enrichment:
Push enrichment beyond 90% (e.g., 99% 233U or 235U) to compensate for thermalization losses. Higher fissile density could sustain criticality even with slower neutrons, though this increases cost and safety risks.
Magnetic Confinement or Flow Control:
Introduce magnetic fields to confine the plasma or reaction zone, potentially enhancing reaction efficiency by keeping fissile material concentrated. This is speculative and complex, as magnetic fields don’t directly influence neutrons, but they could manage charged fission products or plasma dynamics in the exhaust.
Add Fast-Neutron Reflectors or Boosters:
Surround the plenum with a neutron reflector (e.g., beryllium or graphite) doped with a material that scatters neutrons back into the reaction zone without thermalizing them excessively. Alternatively, inject a small flux of high-energy neutrons (e.g., via a fusion source or accelerator) to “seed” the reaction, though this adds significant engineering complexity.
Hybrid Fast-Thermal Design:
Accept a partially thermalized spectrum and optimize for a hybrid reaction. This might not achieve the full “detonating” ideal but could still exceed NTP performance by leveraging both thermal and fast fission. Chloride-based salts (e.g., UCl3) could be explored, as they’re suited to fast-spectrum reactions and might reduce moderation.
Accelerate Interaction Speed:
Increase the flow rate or turbulence of the solution into the plenum to reduce neutron moderation time and enhance criticality. This could mimic a “standing wave” of fission, though it risks instability and incomplete fission yield.
A hydrogenous NSWR as described would struggle to achieve a fast-neutron-driven “detonation” due to moderation, likely resulting in a less efficient, self-limiting reaction. However, the concept isn’t entirely nullified. By rethinking the medium (e.g., reducing hydrogen content), adjusting fissile parameters, or accepting a hybrid reaction, a modified NSWR could still outperform NTP, though it might not hit the extreme Isp Zubrin predicts. Magnetic fields, accelerated particles, or alternative materials could help, but they introduce complexity that strains practicality.
Temperature of a U-235 Supercritical Reaction
In an NSWR, the goal is to sustain a continuous supercritical nuclear reaction in a fluid medium (e.g., a uranium salt solution) within a reaction chamber (plenum). If we imagine discrete critical masses of U-235 colliding—say, two subcritical hemispheres brought together rapidly—the reaction mimics the initial moments of a fission bomb before disassembly.
Physics of the Reaction
Critical Mass of U-235: For bare, unenriched U-235 metal, the critical mass is ~52 kg. With a neutron reflector (e.g., beryllium or steel), this drops to ~15-20 kg. In a solution or with moderation, it varies further, but let’s assume metallic U-235 for simplicity, as a “collision” implies solid masses.
Supercriticality: When two subcritical masses (e.g., 10 kg each) collide to form a supercritical assembly (>20 kg with reflection), the neutron multiplication factor (k) exceeds 1, leading to an exponential chain reaction.
Energy Release: Each U-235 fission releases ~200 MeV (3.2 × 10⁻¹¹ J). In a supercritical state, the reaction rate skyrockets. For a prompt supercritical event (k ~ 1.1-1.5, typical of a bomb), 10¹⁵-10¹⁷ fissions occur in microseconds before the assembly blows apart.
Temperature Estimation
Energy Density: Assume 20 kg of U-235 achieves criticality. If 1% of the mass fissions (a conservative estimate for a brief supercritical excursion), that’s 0.2 kg or ~5 × 10²³ atoms. At 200 MeV per fission, the energy released is:
4 tons of TNT equivalent in a microsecond explosion.
Plasma State: This energy ionizes and heats the uranium and surrounding material into a plasma. The temperature depends on the volume and heat capacity of the system. For a fission bomb’s fireball, temperatures reach 10⁷-10⁸ K (10-100 million K) momentarily, driven by radiation and kinetic energy of fission products.
NSWR Context: In an NSWR chamber, the reaction is diluted by water or other coolant, and the goal is continuity, not instant disassembly. If the energy is spread over a larger volume (e.g., a 1 m³ plenum) and moderated by water, the temperature might drop to 10⁴-10⁶ K (10,000-1 million K) as the water vaporizes and absorbs energy. The exact value depends on the rate of mass collision, confinement time, and cooling effects.
Key Takeaway
For a single collision event, the temperature could spike to millions of Kelvin in the reaction zone, but the presence of water and chamber dynamics would rapidly quench it to thousands of Kelvin as the plasma expands and cools. A continuous process would average lower, perhaps 10,000-50,000 K, depending on flow and dissipation.
Propulsion from the Supercritical Reaction
Exhaust Characteristics
Fission Products: Each fission splits U-235 into two fragments (e.g., barium and krypton) with ~170 MeV of kinetic energy, moving at ~10⁷ m/s. In an NSWR, these mix with water vapor or other carrier fluid, forming a high-velocity exhaust.
Specific Impulse (Isp): Isp is proportional to exhaust velocity divided by gravity.
Continuous vs. Pulsed
If critical masses collide repeatedly, propulsion becomes pulsed, not continuous. Each “bang” ejects exhaust through a nozzle, akin to a rapid-fire Orion drive. A truly continuous NSWR would require a steady-state supercritical fluid, not discrete collisions, which is where the water-moderation critique bites—thermal neutrons slow the reaction, reducing velocity and Isp.
Comparison to a Multi-Gun Fission Bomb and Project Orion
Multi-Gun Fission Bomb Analogy
Gun-Type Bomb: In a gun-type fission bomb (e.g., Little Boy), a subcritical U-235 mass is fired into another to achieve supercriticality. Multiple collisions in an NSWR could resemble a “multi-gun” setup if subcritical slugs are slammed together repeatedly.
Differences: A bomb seeks maximum yield in one explosion, disassembling in microseconds. An NSWR aims for controlled, sustained thrust. Water in the chamber slows the reaction, preventing bomb-like efficiency (yield ~1-20 kt vs. NSWR’s distributed energy).
Project Orion Similarity
Orion Concept: Project Orion used nuclear bombs detonated behind a pusher plate, with each pulse (~0.1-10 kt) accelerating the craft. Isp was ~3,000-6,000 s, limited by plate ablation and pulse frequency.
NSWR Variation: Colliding critical masses in a chamber with water and nozzles is like a “wet Orion.” The water moderates and protects the chamber (unlike Orion’s exposed plate), while nozzles direct exhaust for continuous thrust. If pulsed (e.g., 10 collisions/s), it’s Orion-like but smoother; if continuous, it’s distinct.
Advantages: NSWR could achieve higher Isp (10,000+ s) by avoiding plate inefficiencies and leveraging direct exhaust. The chamber’s shielding reduces radiation compared to Orion’s external blasts.
Nuclear Explosion and Critical Masses Can Happen – Precise Control to Get Desired Propulsion
Temperature: A U-235 supercritical collision in an NSWR chamber could hit 10⁶-10⁸ K locally, cooling to 10⁴-10⁵ K with water and expansion—hot enough for propulsion but not a bomb’s fireball.
Propulsion: Isp of 10,000-50,000 s and thrust of 10⁶ N are feasible with high exhaust velocity, though water moderation may cap performance below Zubrin’s ideal.
Orion Parallel: Colliding masses evoke a pulsed Orion-like system, but the water-buffered chamber and nozzles suggest a hybrid—part continuous reactor, part pulse drive. It’s not a multi-gun bomb unless criticality is uncontrolled, which defeats the NSWR’s purpose.
To make this work, you’d need precise control of collision timing, minimal moderation (e.g., deuterium or no water), and robust chamber materials.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
From Wiki:
“For “thermal” (slow-neutron) fission reactors, the typical prompt neutron lifetime is on the order of 10−4 seconds, and for fast fission reactors, the prompt neutron lifetime is on the order of 10−7 seconds.[16] These extremely short lifetimes mean that in 1 second, 10,000 to 10,000,000 neutron lifetimes can pass.”
From me:
So take a 5 gallon bucket of water and some half inch PVC from the hardware store. Fill the PVC tubes with BORAX from the grocery store and cap the tubes. Place those tubes with regular spacing in the bucket (say 5 or 10 of them). Now, dissolve enriched uranium nitrate in the water until you observe incipient criticality (1/M method). At this point, jerk the PVC tubes out of the bucket in a microsecond and you will have something like the Tokaimura Criticality Accident of 1999. Do it right, and it might even vaporize the water, but most likely it will eject the water, since there will be a steep cosine gradient with the power peaked in the very center. That power peak, voiding, and ejecting the water will shut the reactor down. In the case of The Tokaimura accident, it may oscillate for a while until it loses enough material to no longer support chain reaction. You can make a bucket of water many times prompt critical and it will do this. It might blow up the room with steam, but the neutron generation lifetime is not fast enough to outrun the inertial effects. High explosives detonate – the chemical reaction propagates more quickly than the speed of sound in the explosive, so the material cannot “get out of its own way”, thus the block of solid/plastic material becomes hot gas very quickly, and then expands from the envelope of the solid. Nuclear detonations are the same, in that the material cannot get out of its way. The water in the bucket physically moves/expands/disrupts before 10,000,000 generations of neutrons can fission enough fuel to ionize the material in the bucket (or the NSWR). You could contain it in a vessel, or bind the hydrogen in the solid fuel (uranium hydride), and then it will undergo more generations, spike higher, but ultimately the pulse will turn from one or more inertial/thermal effects. Think of a detonation as [nearly] complete ionization. The 9kg of Plutonium becomes 9kg of million kelvin gas in 500 cc volume – atoms have been liberated of entire shells of electrons. In high explosives, molecules have been liberated of valence bonds. Sure, there is a spectrum in the amount of ionization that would occur in a fizzle all the way up to a full detonation, but one thing shown by the Tokaimura accident example is a bucket of water criticality accident will not detonate.
Even better. Freeze the bucket; that would confine it for a moment. That would probably splinter ice and bucket all over the place. Still not interstellar propulsion.
Sorry, I am probably not adding anything new. But removing the neutron moderating water entirely is probably going to be a bit better. Using a UCI3 nano-dust like material might be better. Seems that clumping might be a problem but the advantages in mechanically controlling the chamber density is worth solving that problem (probably many many other issues too). Any-way GROK3 had some details. I put that below – and I will try to resist from grokking everyone in the future.
——————————————————————————————-
Concept Overview
In this setup, we’re replacing the aqueous uranium salt solution of Zubrin’s NSWR with dry UCl3 nano-dust—finely divided uranium trichloride particles. The idea is to inject this dust into a reaction chamber where its density increases (via compression, confinement, or accumulation) to achieve prompt supercriticality, triggering a fast-spectrum fission reaction. The fission products and unreacted material then expand and exit through a nozzle, generating thrust. Without water, we avoid neutron over-moderation, aiming to preserve the high-energy, fast-neutron-driven “continuous detonation” that promises high ISP and thrust.
UCl3 Nano-Dust Properties
Composition: UCl3 is a uranium compound with a molecular weight of ~344 g/mol (assuming U-235 or U-238, plus three chlorine atoms). It’s a solid at room temperature, melting at ~837°C, and can be ground into nano-scale particles.
Neutron Moderation: Chlorine (primarily Cl-35 and Cl-37) has a neutron scattering cross-section of ~11 barns and a capture cross-section of ~33 barns for thermal neutrons, but these drop significantly for fast neutrons (>1 MeV). Unlike water’s hydrogen, chlorine doesn’t efficiently thermalize neutrons, making UCl3 a better candidate for a fast-spectrum reaction.
Uranium Enrichment: Let’s assume 20% U-235 enrichment (as in Zubrin’s baseline), though higher enrichment (e.g., 90%) could boost criticality and yield.
Mechanics in the Chamber
Density Increase: Nano-dust’s small particle size (e.g., 10-100 nm) gives it a high surface-area-to-volume ratio, allowing dense packing or rapid agglomeration in the chamber. You could achieve this via:
Injection and Compression: High-pressure gas (e.g., inert helium) forces the dust into a confined volume.
Magnetic Confinement: If UCl3 dust can be charged or magnetized, electromagnetic fields might compress it.
Geometric Design: A converging chamber geometry could naturally concentrate the dust.
Criticality: For a fast-spectrum reaction, you need a critical mass and density. Nano-dust’s dispersibility means it’s subcritical in storage (spread out), but once compacted—say, to densities approaching solid UCl3 (~5.6 g/cm³)—it could go prompt supercritical. The exact threshold depends on enrichment, chamber size, and neutron reflection (e.g., beryllium or graphite walls).
Fission Reaction: Once critical, fast neutrons (~2 MeV) from U-235 fission sustain a chain reaction. Without water, the neutron spectrum stays fast, maximizing energy release (up to ~200 MeV per fission event) before the material blows apart.
Performance Potential
Specific Impulse (ISP): ISP depends on exhaust velocity, which scales with the square root of temperature over molecular weight (v ≈ √(T/M)). Fission heats the UCl3 and its dissociation products (U, Cl atoms, or UClx fragments) to millions of degrees, but the heavy molecular weight (~344 g/mol for intact UCl3, ~35-238 g/mol for fragments) tempers velocity. Still, if 90% of the uranium fissions (Zubrin’s optimistic claim), exhaust velocities could hit 50-70 km/s, yielding an ISP of 5,000-7,000 seconds—comparable to or slightly below Zubrin’s design.
Thrust: Thrust scales with mass flow rate and exhaust velocity (F = ṁ * v_e). Nano-dust allows precise control of ṁ by adjusting injection rate. With a high-density reaction, you could push megawatt-to-gigawatt power levels, potentially delivering meganewtons of thrust (e.g., 1-10 MN), depending on scale.
Efficiency: Without water diluting the reaction, more energy goes into kinetic exhaust rather than thermal losses. However, incomplete fission (e.g., 50% yield) would lower performance.
Advantages Over Zubrin’s Design
No Moderation: Ditching water keeps the neutron spectrum fast, avoiding thermalization losses and potentially improving fission yield per gram of fuel.
Storage Safety: Nano-dust dispersed in a low-density state (e.g., suspended in gas) is inherently subcritical until compressed, reducing premature criticality risks compared to a liquid solution near criticality in pipes.
Flexibility: Dust injection could allow throttling—adjusting thrust by varying the dust feed rate—whereas Zubrin’s design is an all-or-nothing burn.
Challenges and Trade-Offs
Containment: The reaction chamber must withstand temperatures exceeding 10^6 K and intense neutron fluxes. Nano-dust might erode or ablate walls faster than a liquid, and chlorine’s corrosiveness (especially if dissociated) could degrade materials like boron carbide or graphite.
Density Control: Achieving and sustaining supercritical density long enough for a high-yield reaction—yet short enough to avoid a catastrophic explosion—is tricky. The dust must transition from dispersed to dense in microseconds, requiring precise engineering.
Exhaust Composition: UCl3 may dissociate into uranium and chlorine atoms or form lighter compounds under fission heat. This affects ISP (lighter exhaust is better) but complicates predictions. Unburned nano-dust in the exhaust could also pose a radiological hazard.
Heat Management: Without water’s heat capacity, all energy goes into the reaction products and chamber walls. Cooling or magnetic nozzle designs (to avoid direct contact) become critical.
Particle Agglomeration: Nano-dust tends to clump due to van der Waals forces or electrostatics, which could clog injectors or unevenly distribute fuel in the chamber.
Could It Work?
Yes, a UCl3 nano-dust NSWR is conceptually viable and sidesteps the water moderation issue. It could match or exceed Zubrin’s ISP (6,730 s) and thrust (e.g., 13 MN for a 300 MW design) if:
Enrichment is high (20-90% U-235).
Chamber density hits supercriticality with a fast neutron spectrum.
Nozzle and containment survive the heat and corrosion.
A back-of-the-envelope estimate: If 1 kg/s of 20% enriched UCl3 nano-dust achieves 50% fission (100 MeV/kg effective yield), that’s ~10^13 W of power. With an exhaust velocity of 60 km/s (ISP ~6,100 s), thrust could reach 60 kN per kg/s—scalable to meganewtons with higher flow rates. Not bad for a dust-powered nuke rocket!
Next Steps
To refine this, you’d need:
Neutron Transport Modeling: Simulate criticality and fission yield with tools like MCNP or OpenMC.
Material Testing: Assess UCl3 nano-dust behavior under high heat and pressure.
Chamber Design: Optimize for density increase without melting down (literally).
——————————————————–
This seems a lot like the NERVA project in the 70’s. Maybe check why it failed as I recall it was a temperature difference, any way it came apart in Nevada. Maybe material science has overcome those hurdles.
To be clear, the issue is that the lower the ‘temperature’ of the neutrons, the more effectively they initiate fission reactions, essentially because they’re more likely to ‘stick’ when they hit a nucleus.
Neutrons start out in the MeV range, then if they bounce off a hydrogen atom, of nearly identical mass, half the energy is lost to the hydrogen atom. After a few bounces, they have “thermalized”, equilibrating their temperature with the hydrogen.
Bounces off U235, OTOH, do not efficiently thermalize, because the U235 nucleus doesn’t recoil much, and so most of the energy remains with the neutron.
So, in the fuel feed, the neutrons rapidly end up with an energy equivalent to 2-300 degrees kelvin, which is very efficient for initiating fission reactions. So criticality is easy to achieve.
As the fuel heats up, the neutrons equilibrate to ever higher temperatures, at which they are less efficient at initiating fission, and so criticality becomes difficult to maintain. So your cold fuel ‘ignites’, and then as soon as it’s hot, the reaction stops, and you don’t get anywhere near a complete reaction, most of the available energy is thrown away with the exhaust.
Great for reactors, it’s why they can be run stablily. Not so great if you want complete burnup in a tiny fraction of a second.
This is why you want to suppress neutron thermalization in the cold fuel. So that initial criticality isn’t lost as the fuel heats up.
You can reduce neutron thermalization in the cold fuel by increasing the atomic weight of the propellant, but that compromises ISP. It also helps to use very highly enriched U235, because when U238 is hit by thermalized neutrons, you don’t get prompt fission, it acts as a fission poison in the very short run. Switching to U233 would help, as its neutron cross section doesn’t drop as fast with increasing neutron energy.
My proposal: Incorporate a ‘burnable’ neutron poison, which will suppress criticality until it has been burnt up. If you get things just right, the neutron poison being burnt up will compensate for the rising temperature and result in an approximately constant criticality. Boron has a nice low atomic weight, and as it’s hit by a neutron fissions exothermically itself producing lithium and helium, with even lower atomic weights. Gadolinium lacks this desirable property, but you’d need enormously less of it, so the impact on the atomic weight of the exhaust might be less anyway.
This is done in regular reactors, to maintain criticality despite increasing fuel burnup. It might work in the NSWR, too.
With feed the feed of cold material, and the [sonic?] flow of material out of the bell, there would not be a steady-state, multiplication factor K = 1 quasi-reactor at arbitrarily enormous power level in a zone in the bell as shown in the diagram. Instead, the cold material would be in a constant state of TRIGA pulse with a capped ceiling power level dictated by the particulars of the geometry/composition/configuration, which then loses reactivity (reaction falls off) quickly thereafter by expanding into the bell. You have the “disassembly” of expansion working to drive power to zero and the negative TRIGA effect in the zone of maximum reactivity (before expansion). You have this NEW material entering a zone with negligible neutron population, and it enters the reaction space, heats up and rises in a prompt manner to the ceiling power level (see TRIGA) as it begins to disassemble via expansion. Adding thermal poisons isn’t going to allow a hydrogenous reactor to function in the thermal spectrum, they will simply keep it suppressed/inactive. Assuming this prompt power ceiling could be well into vaporization or even decomposition of the feed materials, it still looks a lot like a wasteful NTP engine missing the ISP target. Plus, the highest pressure in the system is always the feed pump just like any rocket.
A solid fueled NTP rocket would have some maximum achievable power level dictated by the particulars of the configuration (material on hand, enrichment, geometry) and coolant/propellant feed rate. The coolant/propellent feed rate would be a positive reactivity – the more you flow, the denser the H2, the more cooling, the higher the power in the solid fuel elements UNTIL the power level in the solid fuel elements begins to outrun the convective/radiative heat removal capability of the coolant/propellent flow, at which time the solid fuel elements would disassemble: melt on the inside and lose physical integrity if they manage not to decompose in other manners (coolant erosion). It may be that the geometry/enrichment physical configuration “runs out of reactivity” prior to the other effects, but ultimately there will be a ceiling on the arbitrarily enormous power level of a solid fueled NTP rocket engine.
The physics isn’t necessarily intuitive, but it is well understood. I believe this is why you don’t see serious mention of the NSWR elsewhere, and perhaps why there is not a co-author with reactor physics background.
Correction: ” Adding thermal poisons isn’t going to allow a hydrogenous reactor to function in the FAST spectrum”
That’s the last I’ll add to this discussion. Don’t let it stop you all from using the science fiction to reach Alpha Centauri.
The point isn’t to allow it to function in the fast spectrum. The point is to increase the requirements for criticality in the thermal regime, but go away as you’re transitioning into the fast neutron regime, so that the criticality requirements are about the same.
Picture your fuel flow: It’s coming in through a pintle that heavily suppresses fission. It comes out of it, and suddenly it’s in conditions that would favor going critical, but the boron is suppressing the reaction. As the boron gets eaten up, it becomes easier for the reaction to go critical, but now you’re far enough from the pintle that you don’t melt it. The reaction can’t ‘flash back’ to the pintle because the closer you get to the pintle, the higher the boron concentration.
The fuel is heating rapidly, but it’s also burning up the boron rapidly, which compensates for the cross section dropping as the temperature goes up. The boron is finally gone about the time you reach full temperature.
You know, bottom line, I AM an engineer, and when I encounter a system, I can’t help but think about how to improve it. That’s what I’ve been doing here.
But, really? I don’t like the NSWR because it inherently can’t be built small, and be efficient. The ‘small’ version Zubrin analyzed in his paper,
https://path-2.narod.ru/design/base_e/nswr.pdf
operated in thermal regime, only had a burnup of 0.1%, and an ISP of under 7,000. Which makes chemical rockets look sick, sure, but it’s not the NSWR Brian is talking about.
Brian’s talking about, and we’ve been discussing, the hypothetical scaled up version with a high burnup and an ISP of about 500,000 seconds. It has to operate in the fast neutron regime to do this, and Zubrin proposed it for an interstellar colonization mission. THAT version has to be freaking huge, given that you need a fast neutron critical mass in the reaction zone, and the material is moving through that zone at nearly the speed of sound. I makes no sense for anything smaller than a flying city.
My preference if we’re talking fission rockets with high thrust, would be PuFF, which uses a z pinch to initiate a pulsed fission/fusion reaction. It doesn’t have to be nearly so big to work.
If the boron was in abundance enough to “suppress the reaction” in the thermal regime at the “pintle”, it wouldn’t get “eaten up” to allow operation in the fast regime, and the physical expansion would still drive the assembly sub-critical. I know your an engineer and I generally savor your analyses as other readers and Brian do, but in this case, you ARE reaching and fighting known physics. For fun, I put my bucket example into Grok and it told me the physics was fully consistent and warned me not to do it, as-if.
The comment from Astepintime, switching to the chloride (35Cl) fuel has the MOST validity, although the criticality controls would be more dependent on geometry and spacing in that material because absorbers are basically transparent to fast neutrons aside from elastic collision. Moderation effects tend to drop off after mass number 12 (carbon) and are becoming trivial by mass number 23 (sodium).
Admittedly, while I’m an engineer, I’m not a nuclear engineer.
Bottom line seems to be that the NSWR can beat chemical rockets all hollow, and is somewhat better than closed cycle nuclear rockets, but if you want interstellar performance you need actual explosions.
I doubt the water cooling scheme works well enough.
The absorbed heat still has to be radiated away, and the starship cannot easily have huge surfaces to radiate away heat while it’s accelerating.
I asked Grok 3 (on x):
Provide solutions to the following criticisms of the Nuclear Saltwater Rocket as described below. Would magnetic containment enable the maximum ISP performance without sacrificing nozzle integrity to meltdown from the extremely hot supercritical reaction?
I then loaded criticisms 1 & 2 into Grok 3 in full. Grok’s answer (does it work?):
Let’s address the criticisms of the Nuclear Salt Water Rocket (NSWR) and explore potential solutions, including whether magnetic containment could enable maximum specific impulse (Isp) performance without compromising nozzle integrity. The NSWR, as proposed by Robert Zubrin, relies on a continuous supercritical fission reaction in a mixture of uranium salts (e.g., uranium bromide) dissolved in water, expelled through a nozzle to generate thrust. The criticisms focus on neutron moderation and the feasibility of a fast-spectrum reaction, so I’ll tackle each point systematically and propose solutions.
Criticism 1: Thermal Neutrons and “Detonation”
The critic argues that hydrogen in the water moderates neutrons to thermal energies (~0.025 eV), preventing a fast-spectrum reaction and limiting the system to a self-regulating “fizz” rather than the high-energy, sustained supercritical process Zubrin describes. This moderation reduces efficiency and thrust, dropping the Isp far below the theoretical 10,000+ seconds.
Solution: Deuterium Substitution and Neutron Economy
To mitigate neutron moderation by hydrogen, replace light water (H₂O) with heavy water (D₂O, deuterium oxide). Deuterium has a much lower neutron capture cross-section (0.0005 barns vs. 0.33 barns for hydrogen) and scatters neutrons less effectively, allowing more neutrons to remain in the fast or epithermal energy range (above 0.1 MeV). This shift could sustain a faster fission reaction closer to Zubrin’s vision. While not as ideal as a pure fast-spectrum system (e.g., a metallic core with no moderator), deuterium reduces thermalization significantly.
Additionally, the uranium concentration could be optimized—higher fissile content (e.g., 90% enriched U-235 or even Pu-239) increases the criticality margin, compensating for some moderation losses. The reaction wouldn’t be a “detonation” in the weapons sense but a controlled, rapid fission process with a higher energy release than a thermal reactor. Adjusting the injection rate and mixture density could further tune the reaction to avoid self-limitation, maintaining supercriticality without fizzling out.
Isp Impact
With less moderation, the exhaust velocity (proportional to √(energy/mass)) increases, pushing the Isp closer to the NSWR’s theoretical range. While not fully fast-spectrum, this could achieve an Isp of 5,000-7,000 seconds, a significant improvement over nuclear thermal propulsion (NTP) at 800-1,000 seconds.
Criticism 2: Fluorine and UF6
The critic points out that fluorine (e.g., in UF₆) scatters neutrons in the 0.5-1 MeV range, thermalizing them and making it unsuitable for a fast-spectrum reaction, similar to hydrogen’s effect.
Solution: Non-Fluorinated Salts and Alternative Propellants
Avoid fluorine-based compounds like UF₆ entirely. Stick with Zubrin’s original proposal of uranium bromide (UBr₄) or explore uranium chloride (UCl₄), which have lower neutron scattering cross-sections than fluorine compounds. Chlorine, while still a moderator, is less severe than fluorine or hydrogen, and its isotopes (Cl-35, Cl-37) have manageable capture cross-sections. Better yet, minimize the propellant’s moderating elements by using a low-hydrogen carrier—e.g., a suspension of uranium nanoparticles in a light, non-moderating liquid like liquid methane (CH₄ has less hydrogen per molecule than H₂O, and carbon scatters neutrons weakly).
For a radical approach, consider a hybrid system: inject a small amount of water or methane to initiate fission, then transition to a pure uranium plasma exhaust once the reaction stabilizes. This reduces moderation over time, though it complicates the delivery mechanism.
Isp Impact
By reducing neutron thermalization, the fission energy stays higher, boosting exhaust velocity and Isp. A methane-based system might still lag behind the ideal NSWR but could exceed NTP by a factor of 3-5.
Magnetic Containment for Maximum Isp and Nozzle Integrity
The NSWR’s supercritical reaction produces temperatures exceeding 100,000 K—far beyond the melting point of any material nozzle (e.g., tungsten melts at ~3,700 K). This risks nozzle meltdown, reducing performance and releasing radioactive debris. Could magnetic containment solve this?
Proposed Design
Use a magnetic nozzle—superconducting coils generating a strong magnetic field (e.g., 10-20 Tesla)—to confine and direct the charged fission products and plasma exhaust. In the NSWR, fission fragments are highly ionized and move at relativistic speeds (up to 0.03c), making them responsive to magnetic fields. The field lines form a virtual nozzle, channeling the exhaust without physical contact, thus preserving integrity.
Containment Mechanism: The magnetic field compresses the plasma via the Lorentz force (F = q(v × B)), preventing it from touching the physical structure. A diverging field geometry accelerates the exhaust, converting thermal energy into directed kinetic energy.
Isp Boost: By avoiding energy losses to nozzle heating and ablation, more fission energy goes into exhaust velocity. The theoretical Isp could reach 20,000 seconds or higher, as exhaust speed approaches 3-5% of light speed, limited only by the fission fragment energy (~167 MeV per U-235 fission).
Neutron Management: Magnetic fields don’t affect neutrons, so the deuterium or low-moderator solutions above are still needed to maintain a high-energy reaction.
Feasibility
Magnetic confinement is proven in fusion research (e.g., tokamaks) and ion thrusters, but scaling it to the NSWR’s power level (gigawatts to terawatts) requires massive coils and cryogenic cooling—challenging but not impossible with advances in high-temperature superconductors. The field must also withstand radiation damage from neutrons and gamma rays, necessitating shielding or robust materials like niobium-tin.
Nozzle Integrity
With no physical nozzle exposed to the plasma, meltdown is avoided entirely. The physical structure can be a radiation-shielded chamber, with the magnetic field doing the heavy lifting. Erosion from stray neutrons remains a concern, but this is manageable with replaceable liners or regenerative cooling.
Integrated Solution
Combine these approaches:
Deuterium-based propellant (e.g., UBr₄ in D₂O) to reduce neutron moderation and sustain a rapid fission reaction.
High fissile enrichment (90% U-235 or Pu-239) to ensure supercriticality.
Magnetic nozzle to contain the plasma, protect hardware, and maximize Isp.
This hybrid NSWR achieves an Isp of 10,000-20,000 seconds, far exceeding NTP, while addressing the neutron moderation critique and nozzle meltdown risk. The reaction isn’t a “detonation” but a controlled, high-energy burn—consistent with Zubrin’s intent. Challenges remain (e.g., coil mass, radiation shielding), but this design leverages real physics to push the NSWR toward practicality.
What do you think—should I refine any part of this further?
I followed up on the conversation with the following question:
Are there engineering and safety advantages from using the Integrated Solution design in the vacuum and near zero temperature conditions of space, assuming some other shuttle rocket is used to transport astronauts to the spaceship out of Earth’s gravity well to the hybrid NSWR?
Grok replied:
Let’s evaluate the engineering and safety advantages of the Integrated Solution design for the Nuclear Salt Water Rocket (NSWR)—combining a deuterium-based propellant, high fissile enrichment, and a magnetic nozzle—in the context of space’s vacuum and near-zero temperature conditions. The scenario assumes astronauts are transported to orbit by a separate shuttle rocket, rendezvousing with the hybrid NSWR spaceship outside Earth’s gravity well. This setup leverages space’s unique environment to enhance performance and mitigate risks. I’ll break this down into engineering and safety advantages, considering the physics and practicalities.
Engineering Advantages
Vacuum-Optimized Exhaust Expansion
Context: In a vacuum, there’s no atmospheric backpressure to resist exhaust flow. This allows the magnetic nozzle to optimize expansion of the fission-driven plasma exhaust, converting thermal energy into directed kinetic energy more efficiently.
Advantage: The magnetic field can shape a wider, diverging exhaust plume without the constraints of a physical nozzle’s geometry or atmospheric drag. This boosts specific impulse (Isp) closer to the theoretical maximum (10,000-20,000 seconds), as exhaust velocity scales with √(2E/m), where E is the fission energy (~167 MeV per U-235 fission) and m is the propellant mass. In atmosphere, a physical nozzle would limit this; in space, the magnetic nozzle fully exploits the vacuum.
Implementation: The diverging magnetic field lines act as a virtual nozzle, requiring precise coil design but no material constraints, simplifying construction and reducing mass.
Near-Zero Temperature as a Cooling Asset
Context: Space’s ambient temperature is ~2.7 K (from cosmic background radiation), and the hybrid NSWR operates far from heat sources like Earth’s surface or atmosphere.
Advantage: The superconducting coils for the magnetic nozzle benefit immensely. High-temperature superconductors (e.g., YBCO or BSCCO) operate below ~90 K, and space’s cold provides a natural cryogenic environment, reducing the need for active cooling systems. Passive radiative cooling—dumping heat via large radiators—further stabilizes coil performance under the gigawatt-to-terawatt power loads of the NSWR.
Implementation: Smaller, lighter cryocoolers suffice to maintain superconductivity, cutting mass and power demands compared to terrestrial operation. The cold also reduces thermal stress on structural components, enhancing durability.
No Atmospheric Erosion or Contamination
Context: On Earth, the NSWR’s radioactive exhaust would erode a physical nozzle and contaminate the environment. In space, there’s no atmosphere to ablate or pollute.
Advantage: The magnetic nozzle faces no erosion from atmospheric particles, and exhaust disperses harmlessly into the vacuum (assuming the ship operates beyond low Earth orbit, LEO). This eliminates the need for heavy shielding or ablation-resistant materials in the propulsion system, reducing mass and complexity.
Implementation: The design can prioritize lightweight, radiation-tolerant materials (e.g., carbon composites) for the chamber, with the magnetic field handling containment.
Simplified Ignition and Control
Context: Starting the NSWR in space avoids gravitational and atmospheric interference, allowing finer control over the supercritical reaction.
Advantage: The deuterium-based propellant (e.g., UBr₄ in D₂O) and high fissile enrichment can be injected precisely without wind shear or buoyancy effects disrupting flow. The vacuum also prevents unwanted pre-ignition from atmospheric oxygen or pressure.
Implementation: A compact neutron source (e.g., a californium-252 initiator) triggers fission, and the magnetic field stabilizes the plasma from the outset, leveraging space’s stability for a smoother startup.
Safety Advantages
Isolation from Biosphere
Context: Launching from Earth risks radioactive fallout if the NSWR fails. Operating in space, post-shuttle rendezvous, confines risks to the vacuum.
Advantage: A failure (e.g., containment breach) releases fission products into space, not Earth’s atmosphere or surface. Beyond LEO (say, 1,000 km altitude), debris disperses without reentry risk for millennia, protecting astronauts on the shuttle and Earth’s population.
Implementation: The NSWR ship could be assembled and fueled in orbit (e.g., at a lunar gateway or space dock), ensuring no ground-based exposure. Astronauts board after the system is tested remotely.
Reduced Thermal Hazard
Context: The NSWR’s plasma exceeds 100,000 K, posing a meltdown risk to physical components on Earth. Space’s cold and vacuum mitigate this.
Advantage: The magnetic nozzle eliminates hot plasma contact with physical surfaces, and space’s low thermal conductivity prevents heat buildup in the ship’s structure. Radiative cooling handles residual heat, reducing fire or explosion risks compared to atmospheric operation.
Implementation: Radiators sized for gigawatt dissipation (e.g., 100 m² of high-emissivity panels) keep the ship’s habitat and systems below 300 K, safe for crew and electronics.
Radiation Management
Context: The NSWR produces neutrons and gamma rays, hazardous in a crewed vehicle.
Advantage: In space, the magnetic field deflects charged fission fragments away from the crew module, and the vacuum allows neutrons to escape without scattering back (unlike in air). The cold environment also aids shielding—liquid hydrogen or water tanks double as coolant and neutron absorbers.
Implementation: Position the crew module 50-100 meters from the reactor, behind a shadow shield (e.g., 10 cm of tungsten or boron carbide), cutting exposure to 1 with gigawatt power, ideal for interplanetary missions.
Crew Safety: Rendezvous beyond LEO (e.g., at a Lagrange point) keeps astronauts far from radiation until the NSWR is proven operational. A separate shuttle minimizes exposure during ascent.
Testing: Initial tests could occur in deep space (e.g., cis-lunar orbit), refining the design without Earth-based risks.
Conclusion
The Integrated Solution shines in space. Engineering-wise, the vacuum maximizes Isp and simplifies exhaust management, while the cold enhances superconductivity and cooling. Safety-wise, isolation from Earth, reduced thermal risks, and easier radiation control make it far preferable to ground launch. The hybrid NSWR becomes a high-performance, low-risk propulsion system for deep-space missions—say, Mars in 30 days—once paired with a shuttle to escape Earth’s gravity well. Any specific aspect you’d like me to dig deeper into?
This short dialog session – done in a few minutes – confirms that the combination of cutting edge AI, combined with sophisticated yet flawed leading technological proposals can lead to a Transformative Artificial Intelligence Discovery (TAID). AI accelerates discovery by orders of magnitude not yet fully realized, as well as rocket development in this example.
Countering this is the current administration’s commitment to rolling back all STEM research funded by the government, defunding education (by a president who says he “loves the poorly educated”), and kowtowing to a single over-worked, over-committed, conflicted of interest technologist – Elon Musk – who is pushing a stale chemical rocket design to reach Mars and the rest of the solar system.
I hope you don’t consider your glorified google search (Grok) to be a solid rebuttal of my estimation of the important physics involved here. These AI models are trained on Reddit posts and vlogs like this and generally regurgitate to some degree whatever they can dredge. They often switch tac as contradictions in the responses are highlighted, but always end in cheerful, positive, pilot affirming conclusion. IOW, your google search != to my MS in Nuclear Engineering and 25 years in the most active part of the field I could find.
And for what we know about ‘Zubrin’s intent’ regarding detonation in the original paper, in contrast to your grokbot answer, is that Zubrin was clever enough to realize that the concept would need to detonate (in the sense of a continuous fission bomb) to achieve the high fuel utilization, temperature, and ISP called for in the orignal paper. His design is literally a force-fed bomb. Anything else is just a variant of NTP with poor fuel urtication.
Fascinating article. Different ways to exploit nuclear fission IMO, offer both understood science, and engineering doable “now”. The best way to find the best way to adapt this for space travel is likely trial and error. But this is an engineering problem, not a science one. (By contrast don’t hold your breath waiting for controlled nuclear fusion, you’ll turn blue…). For any technology to be effective, don’t strive for “perfection”, but for effective, efficient, and robust (meaning it doesn’t have to be perfect, to be good enough to do the job well enough).
IMO, this has been a problem with controlled fusion. Any one part of the reaction NOT perfect, no fusion. Fission offers tremendous potential for space propulsion, no doubt using different engineering architectures. As to this approach? Need to study it further. But looks quite interesting.
Comment from Scaryjello:
Thank you for your work.
Normally, the heat input occurs in a [combustion] chamber, right? And there is a transition to supersonic flow somewhere down the nozzle or at the outlet, right? Forgive me, I’m no rocket scientist. With this concept, the heat input (super prompt criticality) occurs in the diverging nozzle, AFTER EXPANSION WORKS TO DECREASE criticality/buckling? Your graphic is showing a tube of fuel fluid with “no neutron absorbers” for several pipe diameters leading up to the bell – that is where it would be most “reactive”, when it is in the liquid state outside of the “neutron absorbing plumbing”. Your graphic would blow up in that pipe leading up to the nozzle of the bell. Perhaps you would have better control of that if you had radial injection from neutron absorbing pipes into a traditional “combustion” chamber, BUT NOW YOU’RE LIMITED TO THE SAME MATERIAL SCIENCE as the NERVA or DRACO. I don’t think you can expect the sweet spot for this fizzle to be downstream in the bell, and you can’t run a 700GW reactor in an Inconel combustion chamber, likely no matter the cooling.
We want to accelerate light atoms, like H2 because they give good ISP. Unfortunately, those light atoms will fight the “runaway chain reaction” that would make this concept work.
Brian, I’m happy you’ve corrected the basic physics here and given me a shout out. I’m often wrong and often right. I kinda like you using the AI for quick turnaround on these articles – great tool. I look forward to the day when AI can design a reactor.