Fission surface power unit
Reactor core
Makes the heat. HALEU fuel pins in a hydride moderator, surrounded by a reflector that controls the chain reaction.
- HALEU uranium nitride fuel pins
- Yttrium hydride moderator
- Beryllium oxide reflector
- Control drums or sliding reflector with actuators
- Launch shutdown lock (plan design choice)
Power conversion
Turns reactor heat into electricity. Free-piston Stirling engines or closed Brayton turbines, in four independent strings.
- 4 × ~12.6 kW Stirling converters (or 4 × ~11.7 kWe Brayton units)
- Linear alternators (Stirling) or 2 kHz three-phase alternators (Brayton)
- Hot-end heat exchangers
- Cold-end coolant interface
- Gas bearings
Heat transport
Moves heat from the core to the converters without pumps, using sealed sodium heat pipes.
- Sodium heat pipes
- NaK or gas-plenum primary heat exchanger
- Heat-pipe wicks (Na–Mo, steel)
- Insulation and thermal shield
Radiators
Throws away the waste heat to space. Two vertical wings of panels with water heat pipes.
- Composite panels, 2.5 × 2 m each
- Embedded titanium–water heat pipes
- Pumped water manifold and pumps
- Scissor deployment mechanism
- Parasitic load (shunt) radiator
Radiation shield
A one-sided shadow shield that protects the crew zone, plus distance.
- Lithium hydride neutron layer
- Tungsten gamma layer
- Shield structure and cooling
- Keep-out zone markers and dosimeters
Control, instrumentation and power management
Runs the reactor and converters by itself, matches load, and steps voltage up for the long cable.
- Neutron and temperature sensors
- Reactor control computer and Ka-band radio
- Stirling or Brayton controllers
- Start-up Li-ion battery (~4 kWh)
- Step-up converters or transformers
- User interface pallet with 120 VDC outputs
Deployment and lander interface
Gets the unit off the cargo lander and the user interface out to the grid tie point.
- Reactor pallet adapter on lander
- User interface pallet on a sled
- 1 km high-voltage cable spool
- Utility rover for towing
- Levelling legs
What it does
The fission unit is the base’s night-time anchor. Polar sites can be dark for up to about 100 hours at a time1, and the US reactor draft asks for operation through at least 354 hours of darkness5. A reactor of about 40 kWe covers the oxygen pilot plant or the habitat through any shadow, and lets the solar towers handle the daytime peaks.
How it works
- Core. NASA’s 40 kWe concept uses about 20% enriched uranium nitride pellets, sodium–molybdenum heat pipes and a yttrium hydride moderator, at about 250 kWth2. KRUSTY proved the basic idea in 2018. It ran a heat-pipe reactor with Stirling converters for 28 hours, at up to 5 kWt and 880 °C fuel temperature, and it followed the converters’ load demand on its own4.
- Heat pipes. KRUSTY was the first reactor cooled mainly by heat pipes. Its sodium pipes began carrying real power once the core reached about 500 °C4.
- Stirling or Brayton. NASA’s 2025 reference design compared the two3:
| Free-piston Stirling | Closed Brayton | |
|---|---|---|
| Units | 4 × 12.6 kW | 4 × 11.7 kWe |
| Converter efficiency | 29% | 20% |
| Mass per converter | 250 kg | 300 kg |
| Radiator area | 140 m² | 180 m² |
| Power electronics efficiency | 83% | 90% |
| Total with 20% growth | 6.8 t | 7.5 t |
Stirling is about 10% lighter. Brayton gives simpler power electronics3. NASA’s 2025 request and the 2026 Lunar Reactor-1 draft both name closed Brayton65. NASA rated 10 kWe-class Stirling at TRL 5 and Brayton at TRL 43. Open Moon accepts either.
- Radiators. These are polymer composite panels with titanium–water heat pipes, 2.5 × 2 m each, arranged in two wings3. An earlier ISS-style design needed 133.4 m² at the pole and 62% more at the equator2.
- Shield. A directional shield with a 36° included angle keeps the dose below 5 rem/yr at 1 km3. In the 2022 concept, the lithium hydride and tungsten shield weighed 1,250 kg2.
- Control and power. Stirling engines need a constant electrical load. The controllers therefore dump unused power into a resistor bank or radiator2. A 4 kWh Li-ion battery handles start-up2.
- Deployment. In NASA’s reference option, the reactor pallet (about 6 t) stays on its lander. A rover tows the user interface pallet (about 1 t) to the end of a 1 km cable. Start-up takes about 8 hours3.
Sizing
NASA 2025 reference masses in kg3:
| Item | Brayton | Stirling |
|---|---|---|
| Reactor | 1,413 | 1,132 |
| Shield | 1,500 | 1,370 |
| Power conversion | 1,182 | 995 |
| Heat rejection | 884 | 701 |
| Local power electronics | 401 | 613 |
| User interface pallet (incl. 325 kg cable and spool) | 872 | 872 |
| Total with 20% growth | 7,502 | 6,820 |
Worked numbers:
- Specific mass (this plan’s arithmetic). 6,820 kg ÷ 40 kWe ≈ 170 kg/kWe. NASA’s 2022 figure was 250–270 kg/kW1.
- Dose at greater distance (ESTIMATE). Dose falls with 1/r²3. At 2 km, the 1 km limit of 5 rem/yr becomes about 1.25 rem/yr. This ignores scatter from the ground.
- Cable. At 3 km, NASA’s transmission study prefers about ±4 kVDC7; see C04.4.
Testing and qualification
- Components. Irradiate fuel, moderator, alternator organics and sensors. Run long vacuum tests on heat pipes and superalloys3.
- Engineering test unit. Build a full non-nuclear system with electric heaters in place of the core, as the Lunar Reactor-1 draft requires5. Test it in thermal vacuum with radiators and the full cable.
- Nuclear ground test. Repeat the KRUSTY method at flight scale: zero-power criticals, then a powered run with load-following and loss-of-cooling transients4.
- Dust and deployment. Load radiators with regolith simulant. Practise towing the user interface pallet on slopes up to 10°5.
- On the Moon. Commission it step by step: radiation mapping, power setpoint changes, string shutdown and restart, and day–night transients3. M04 exit requires 40 kWe for six months.
Open questions
- Which converter will the first flight unit actually use, and will its grid tie be 3 kV AC (as in the draft)5 or DC?
- How much does dust really reduce radiator performance after years near active landing pads?
- Can a regolith berm replace part of the 1.4–1.5 t shield, and at what rover effort?
- What is the best spare strategy: a fifth converter string, or a second smaller reactor later?
- How is the unit made safe and stored at end of life?
| Severity | Problem | Solution |
|---|---|---|
| Critical | The US requirement keeps changing (40 kWe Stirling in 2022, ≥100 kWe Brayton in 2025, ≥20 kWe Brayton in the 2026 draft) and no reactor contract exists. | Specify only the grid interface (power range, voltage, fault behaviour) and buy whichever unit flies first; keep the solar-plus-storage fallback in C04.2 and C04.3. |
| Critical | Radiation dose to crew if the unit lands or is placed wrongly. | Land the reactor ≥1 km from the crew zone, inside the shield's 36° shadow; add terrain or berms; monitor dose all the time at the user point. |
| High | A converter string fails; the 2022 concept carried no spares, so each lost pair cost about a quarter of output. | Use four 25% strings with separate radiator loops and controllers, and size the base's essential load below 75% of output. |
| High | Lunar dust on hundreds of square metres of radiator could cut heat rejection; the effect is not yet measured. | Fund NASA's radiator dust tests to completion; place the unit far from landing pads; add 20% radiator area margin. |
| High | Mass growth: the 2022 design hit ~10 t against a 6 t goal. | Keep the reactor on its lander (the lightest NASA option) and move only the ~1 t user interface pallet. |
| High | Radiation-hard power switches are rated far below the transmission voltage. | Use transformers (AC) or stacked low-voltage modules (DC) for step-up; see C04.4. |
| Medium | Organic parts in Stirling alternators and superalloys in vacuum may degrade over 10 years. | Complete the irradiation and metal-vaporisation tests now running at NASA and DOE before flight design freeze. |
| Medium | Nuclear launch approval and fuel supply can slip the schedule. | Start approval paperwork at contract award; size the first two years of the base to run on solar plus storage. |
Specification
| Output | 40 kWe net at the user (plan baseline); grid accepts any 20–100 kWe unit |
| Reactor heat | ~175 kWt (Stirling) to ~230 kWt (Brayton), NASA reference designs |
| Mass | ~6.8 t (Stirling) to ~7.5 t (Brayton) incl. 20% growth; older deployable concept ~10 t |
| Conversion | 4 × ~12 kWe Stirling or closed Brayton strings; ≥75% power after first failure |
| Radiators | ~140 m² (Stirling) to ~180 m² (Brayton) |
| Crew distance | ≥1 km behind a 36° shadow shield; ≤5 rem/yr at 1 km |
| Design life | 10 years (NASA reference); ≥5 years (LR-1 draft) |
| Start-up | Remote, about 8 h to full power |
| Delivery | 1 heavy cargo lander; reactor pallet stays on the lander |
Sources
- Csank & Scott, NASA GRC: Power and Energy for the Lunar Surface (Apr 2022)
- Oleson et al., NASA GRC: A Deployable 40 kWe Lunar Fission Surface Power Concept (NETS 2022) (May 2022)
- INL/NASA: Fission Surface Power Technical Maturation Webinar (government reference design) (Sep 2025)
- Poston et al.: Results of the KRUSTY nuclear system test (2020)
- NewSpace Economy: What NASA's SPARC procurement reveals about Lunar Reactor-1 (secondary) (Sep 2026)
- NASA: NASA seeks industry feedback on fission surface power (Aug 2025)
- Barth & Pike, NASA GRC: Lunar power transmission for fission surface power (2022)