Nuclear propulsion

Nuclear propulsion has been technically demonstrated for sixty years and has never flown. Understanding why requires separating the physics, which is favourable, from the engineering and political constraints, which are not.

Why it doubles specific impulse

From Chemical propulsion, exhaust velocity scales as

A chemical engine is stuck with both terms. Its temperature is whatever the reaction produces, and its exhaust molecular mass is whatever the products happen to be.

A nuclear thermal rocket breaks the coupling entirely. The reactor supplies heat from fission, so the working fluid can be chosen purely for low molecular mass. Pure hydrogen, $M = 2$, is roughly six times lighter than hydrolox combustion products.

The nuclear engine runs cooler and still more than doubles the figure of merit, because the molecular mass term dominates. That is the entire case for nuclear thermal propulsion, and it is sound: roughly 900 s against 450 s.

What was actually built

The Rover and NERVA programmes ran from 1955 to 1973 and were far more successful than their cancellation suggests. Around twenty reactors were built and tested at the Nevada Test Site.

Phoebus-2A, 1968

Operated at 4082 MW thermal, still the most powerful nuclear reactor of any kind ever built for propulsion.

NRX-A6, 1967

Ran for 60 minutes continuously at full power, demonstrating the endurance a real mission would need.

XE-Prime, 1969

Fired downward into a low-pressure chamber in a flight-representative configuration, and demonstrated 28 restarts.

Performance achieved

Specific impulse of 825 to 850 s, against roughly 450 s for the best chemical engines of the era.

The programme was cancelled in 1973 not because it failed but because the missions it existed for, crewed Mars flight in particular, were cancelled first. The technology was left at a high state of readiness with nothing to do.

The engineering constraints

Three physical problems bound nuclear thermal performance, and none has been fully solved.

Fuel element erosion. Hot hydrogen at 2500 K and above is chemically aggressive and attacks the fuel elements, carrying fissile material out through the nozzle. This limits both operating temperature and total burn duration, and it means the exhaust is not as clean as the simple physics suggests.

Hydrogen dissociation. Above about 3000 K molecular hydrogen begins to dissociate into atomic hydrogen. Dissociation absorbs energy that would otherwise become directed kinetic energy, so raising chamber temperature stops paying at exactly the point where the material limits become severe anyway.

Thrust-to-weight ratio. A reactor plus its shielding is heavy. A nuclear thermal stage has a thrust-to-weight ratio around 3 to 7, against 50 to 100 for a chemical engine. It cannot lift off from Earth, and it is only useful as an in-space stage. Shielding is not optional either, because crew and avionics need a shadow shield between them and the core.

Nuclear electric propulsion

A different architecture: use the reactor to generate electricity, then run electric thrusters from it.

The difficulty with nuclear electric is thermodynamic rather than nuclear. Converting reactor heat to electricity is at best 25 to 35% efficient, so most of the reactor’s output must be rejected as waste heat. In space the only rejection path is radiation, and radiator area scales as $T^{-4}$, so a megawatt-class system needs radiators of hundreds of square metres. Radiator mass, not reactor mass, is what dominates a nuclear electric vehicle and what has kept the concept on paper.

The KRUSTY experiment in 2018 demonstrated a compact 1 kWe fission system under the Kilopower programme, using a solid uranium-molybdenum core with Stirling convertors. It is a genuine milestone, but a kilowatt is three orders of magnitude below what a propulsion system would need. Its nearer application is fission surface power, where radiators can sit on the ground and mass constraints are looser.

Why it matters for crewed Mars flight

The case for nuclear propulsion is not primarily about propellant mass. It is about exposure time.

From Space radiation, the MSL RAD instrument measured 1.8 mSv per day during Mars cruise, or 662 mSv for a 360-day round trip. That alone consumes most of NASA’s 600 mSv career limit before any surface operations. Galactic cosmic rays cannot be shielded away with practical mass, so the only remaining variable is how long the crew is exposed.

Nuclear thermal propulsion shortens transit. That converts a radiation problem, which has no good engineering solution, into a propulsion problem, which does. It is the strongest argument for the technology and the reason it keeps returning.

Programmes and constraints

DRACO, a DARPA and NASA nuclear thermal demonstration, has been the most substantial recent effort, aiming at an in-space demonstration of a flight-scale engine. Programmes of this kind have repeatedly been announced, funded for a few years and then descoped, which is itself the most reliable pattern in the field.

The non-technical constraints are real and should not be dismissed as mere politics.

None of these is insurmountable, and all of them add years and cost to any programme. Combined with the fact that chemical propulsion is adequate for every mission actually being funded, the result is a technology that is perpetually ten years away.

References

• Dewar, J. A. To the End of the Solar System: The Story of the Nuclear Rocket, 2nd ed., Apogee, 2007. The history of Rover and NERVA.
• Finseth, J. L. “Rover Nuclear Rocket Engine Program: Overview of Rover Engine Tests”, NASA CR-184270, 1991. The primary test record.
• Gibson, M. A. et al. “The Kilopower Reactor Using Stirling Technology (KRUSTY) Nuclear Ground Test Results and Lessons Learned”, AIAA 2018-4973.
• Borowski, S. K., McCurdy, D. R. and Packard, T. W. “Nuclear Thermal Rocket (NTR) Propulsion: A Proven Game-Changing Technology for Future Human Exploration Missions”, NASA/TM-2012-217290.
• National Security Presidential Memorandum 20, Launch of Spacecraft Containing Space Nuclear Systems, 2019.