Radioisotope Power Systems¶
Created: session 23, 2026-04-05
Summary¶
NASA's radioisotope power landscape is in transition. The MMRTG (thermoelectric Pu-238) has been the workhorse for rover-class missions. Two structural shifts are underway: (1) Stirling conversion is replacing thermoelectrics — roughly 4× more efficient, dramatically reducing required isotope mass; (2) Americium-241 (Am-241) is emerging as a commercially-sourceable alternative to DoE-controlled Pu-238, opening paths to private-sector RPS development. Three active TechPort projects define the current frontier: Harmonia (GCD/commercial Am-241 Stirling), APPLE (NIAC modular Pu-238), and Small RTGs for Mars (MEP Pu-238/Am-241 impactor probes). Fission power (Kilopower → FSP) sits in a separate category but at the same destination layer.
Technology Baseline: MMRTG¶
The Multi-Mission RTG (MMRTG) is NASA's current operational standard, flying on Curiosity (2012) and Perseverance (2021). Key specifications: - Power: ~110 We beginning-of-mission; ~100 We after 14 years - Heat source: 18 General Purpose Heat Source (GPHS) modules, each ~250 Wt - Total thermal power: ~2,000 Wt → ~110 We (5.5% efficiency — thermoelectric) - Mass: ~45 kg - Specific power: ~2.4 We/kg
Limitation: Low efficiency (thermoelectric SiGe, ~5%), high Pu-238 consumption per watt of electricity. Pu-238 production at DoE was restarted in 2013 at Oak Ridge (~1.5 kg/yr target). A single MMRTG requires ~4.8 kg Pu-238. Supply is the binding constraint.
Active RPS Development Projects¶
Harmonia: Am-241 Stirling RPS (147008)¶
| Field | Value |
|---|---|
| Project ID | 147008 |
| Program | GCD (Tipping Point) |
| Lead | Zeno Power Systems, St. Louis, MO |
| Period | 2023-07-20 to 2028-07-20 |
| TRL | 4 (began 4, Am-241 heat source target 5; ESG target 6) |
| Destinations | Moon and Cislunar, Others Inside the Solar System |
| Views | 1,939 |
What it is: Am-241 Stirling Radioisotope Power System. The project has two separable milestones: 1. Am-241 heat source: Novel heat source design (currently TRL 4); target TRL 5 by 2028. Key challenge is that Am-241 as a heat source material has never been qualified for space. 2. Electrically heated Stirling Generator (ESG): Surrogate testing using electrical heaters in lieu of nuclear material → validates Stirling cycle in lunar thermal vacuum. ESG target TRL 6 by 2028.
Team structure — notable breadth:
| Organization | Role |
|---|---|
| Zeno Power Systems | Lead; Am-241 heat source design and commercialization |
| Blue Origin | Power system design; Stirling convertor integration into RPS |
| Intuitive Machines | Lander interface requirements; lander simulator design |
| Sandia National Laboratories | Nuclear material/safety expertise |
| Sunpower, Inc. | Stirling convertor — same supplier as Kilopower |
| University of Dayton Research Institute | Analysis support |
States: NM (Sandia), OH (Sunpower/Dayton), WA (Blue Origin) — three of the four nuclear-adjacent states in the FSP ecosystem.
Why Am-241? - Half-life: 433 years vs. Pu-238's 87 years → slower power decay over mission - Supply: Produced as decay product of Pu-241 in nuclear weapon stockpile waste; UK's National Nuclear Laboratory has been extracting from civil plutonium. Increasingly commercial. Not DoE-rate-limited. - Trade-off: Lower power density than Pu-238 (~3.4 W/g vs. ~0.56 W/g for Am-241, but relevant comparison is per gram of usable output given Stirling efficiency gains) - Cost: Potentially much lower than DoE-sourced Pu-238 if commercial scale achieved - Recycling angle: "long-lived nature of Am-241, lunar surface heat sources can be recycled via interoperability for decades" — a lunar RPS economy concept
Commercial significance: Harmonia explicitly targets a commercial radioisotope capability. If successful, it would be the first non-DoE, non-government-monopoly space nuclear heat source. Space Policy Directive-6 explicitly supports commercial nuclear power development.
Planetary Science Decadal Survey note: Recommended near-term demonstration of Stirling conversion for RPS — Harmonia is the GCD response.
Open thread: No TechPort documents. The technology readiness assessment (establishing Am-241 heat source baseline TRL) is the first deliverable — results not yet visible in TechPort.
Small RTGs for Distributed Mars Instruments (182844)¶
| Field | Value |
|---|---|
| Project ID | 182844 |
| Program | MEP (Mars Exploration Program) |
| Lead | Jet Propulsion Laboratory |
| Period | 2025-06-01 to 2028-09-30 |
| TRL | 2 (began 1, target 5) |
| Destinations | Mars, Moon, Others |
| Views | 2,753 |
What it is: Small-scale RTG (~60 Wt / ~3 We) for hard-landing impactor probes on Mars. Note: significantly smaller than MMRTG. Designed to be: - Dual-isotope compatible: Pu-238 or Am-241 (whichever is available commercially) - Hard-landing tolerant: Additively manufactured crushable absorber at landing — no parachute architecture needed - Long-lived: Up to 10 years operation
Mission concept — Mars Environment MAP:
Image source: TechPort fileId 387731 — JPL MEP concept art for distributed sensor network.
The concept deploys multiple probe nodes across Mars via atmospheric entry + hard landing. Each node sits in an impact crater-like zone, with an antenna mast. A mesh network relays "Daily Reports" (temperature, wind, seismology, radiation, dust) to orbiter relay. This would be the first geophysical sensor network on Mars (vs. single-node InSight).
NASA context: This addresses a key Artemis preparation gap — characterizing Mars surface conditions (radiation, dust, temperature extremes, seismic activity) at multiple locations before committing to human landing site selection. Sensors measure "temperature, radiation levels, dust, and wind velocity — all vital data for ensuring astronaut safety."
Isotope supply angle: By accepting Am-241, JPL creates a hedge against Pu-238 supply constraints and simultaneously validates the commercial isotope supply chain that Harmonia is trying to establish.
Granularity gap: Human-assigned TX03 (top-level), ML correctly identifies TX03.1.2 (Heat Sources) — the primary technology challenge is the Am-241/Pu-238 heat source design, not the power conversion.
APPLE (NIAC) — Modular Radioisotope Tiles (117025)¶
Document read: 2023 NIAC Symposium poster (file 381105) + SGL bus image (file 381104). Session 96 (2026-04-08).
| Field | Value |
|---|---|
| Program | NIAC Phase II (completed) |
| Lead | Aerospace Corporation (El Segundo, CA) |
| Team | Aerospace Corp + JPL (Bux, Chun-yip Li) + Oak Ridge NL (Veith) |
| Period | 2022-2024 |
| TRL | 3 (began 2, target 3) |
| Specific power (Pu-238) | 23 g/We (~15× MMRTG) |
| Specific power (Am-241) | 110 g/We (~3× MMRTG) |
| Thermoelectric material | Skutterudite (SKD) — newer than MMRTG SiGe |
| Battery | ORNL solid-state 7Li — radiation-hard |
| Tile dimensions | 10×10×1.7 cm, 39g (Pu-238 variant) |
| Electrical output/tile | 1.7 We |
| Thermal available/tile | 16 W (for spacecraft heating) |
APPLE tiles (green) covering an SGL sailcraft bus. 24 tiles/ring × 6 rings = ~160 We. Each ring is an independently-powered, deployable vehicle — the string-of-pearls SGL architecture.
Architecture — two isotope variants:
The poster explicitly compares two designs that produce identical 1.7 We output: - Pu-238 concept: 31g of ²³⁸PuO₂ (clad in Pt) + Silica Aerogel MLI + SKD thermoelectrics + radiator/battery. Total tile mass: 39g → 23 g/We (15× better than MMRTG) - Am-241 concept: 154g of ²⁴¹Am₂O₃ (clad in Pt) + same conversion chain. Total tile mass: 187g → 110 g/We (3× better than MMRTG). Am-241 has lower specific activity but potentially commercial supply.
Both variants use the same tile geometry, enabling a single hardware design that accepts either isotope depending on availability.
ORNL solid-state radiation-hard battery: Each tile integrates a solid-state battery cell to buffer power for variable loads and survive transient radiation events: - Isotopically pure ⁷Li (neutron-transparent — resists neutron activation from isotope decay) - No polymer separators, no organic electrolytes — radiation-stable chemistry - Electrolyte: Li₃PO₄·N₆Li₃ (LISICON family) - Cathode: NMC333 (nickel-manganese-cobalt) - Radiation qualification test running at ORNL using Am-Be-3215 source; 1 month running ≈ 3 years of Earth-to-Jupiter cumulative dose
Why APPLE is a more radical redesign than Harmonia: - MMRTG uses GPHS modules (bricks of Pu-238 clad) in a fixed cylindrical canister. 4π of heat, but only 2π of radiator area utilized. - APPLE tiles are placed on any surface of the spacecraft — every exterior panel becomes a power source. Eliminates the isolated canister architecture entirely. - Thermal waste heat serves dual purpose: spacecraft heating (no separate heater circuits needed; 16 W thermal per tile) - Scales from 10s of watts to kilowatts by adding tiles — no redesign required
The SGL design case (Phase II primary mission study): The poster's design case uses APPLE to power the Solar Gravitational Lens sailcraft: - 24 tiles/ring × 6 rings = 144 tiles total → ~160 We bus power - Each ring of the string-of-pearls SGL architecture is an independently powered vehicle — APPLE tiles enable this separation architecture without requiring inter-vehicle power transfer - Also present on bus: EP thruster (electric propulsion) + telescope primary mirror
Enabled mission classes: - Lunar rover night operations + permanently shadowed regions (2-week night is no barrier) - Mars helicopter: 20-50× more flights per sol (vs. solar-powered Ingenuity baseline) - Small landed platforms: 2 tiles (3.4 We) → 2 km traverse, 1.3 kg science payload - Distributed telescope array: tiles power individual elements of a formation that separates at target - Outer solar system small spacecraft (SGL, Neptune, interstellar precursor)
Open threads: - GCD pickup check done (session 97, 2026-04-08): Searched all active TechPort projects for Aerospace Corp + Nemanick/APPLE. No GCD follow-on found. APPLE remains at NIAC Phase II, TRL 3, no program champion as of April 2026. - Am-241 variant is a hedge against Pu-238 supply constraints. If Harmonia [147008] qualifies an Am-241 heat source, APPLE Am-241 becomes more attractive. - ORNL battery qualification status: 1-month radiation test ≈ 3yr equivalent. Full mission qualification (Saturn, ~7yr cruise) requires longer test campaign. - Harmonia vs APPLE positioning: Harmonia (funded GCD, Am-241+Stirling, IM/Blue Origin team) represents the funded path to small RPS. APPLE's unique value-add (tile form factor, Pu-238 efficiency) is complementary but lacks the commercial-end-user pull that got Harmonia funded. The Harmonia model (startup Zeno Power with a defined customer Intuitive Machines) is instructive for what APPLE would need.
Thermoradiative Cell (TRC) RPS (158671)¶
| Field | Value |
|---|---|
| Project ID | 158671 |
| Program | NIAC Phase II |
| Lead | Rochester Institute of Technology (NanoPower Research Labs) |
| Co-investigator | Geoffrey Landis (NASA GRC) |
| Period | 2024-05-15 to 2026-05-15 |
| TRL | (not populated — NIAC concept stage) |
| Views | 417 |
What it is: A thermoradiative cell converts heat from a radioisotope source into infrared radiation emitted to the cold universe (3 K), generating electricity in the process. It is the thermodynamic inverse of a solar cell: instead of absorbing photons from a hot source, it emits photons to a cold sink. Heat from the radioisotope keeps the cell warm; the cold universe acts as the thermal ground.
Key physics: TRC devices emit photons by thermal recombination of electrons and holes. Because the cold universe is the sink, the cell operates at negative voltage — current flows at positive terminal from the thermal emission asymmetry. A voltage bias applied externally extracts power.
Phase I result (demonstrated): 0.6 mW/cm² using a commercial InGaAs photodetector (0.74 eV bandgap), operated in a modified cryostat with a black absorber surface simulating cold space. This was an unoptimized, off-the-shelf detector — proof-of-concept only.
Phase II theory (targeted with Eg = 0.28 eV InAsSb, Tc = 600 K, T_ambient = 15 K): - Power density: 6 mW/cm² - Conversion efficiency: 12.3% (vs. MMRTG ~5.5%) - A 5W system needs less than 2/3 of one GPHS pellet (62.5 Wt) and <840 cm² TRC area — could fit on fewer than 9 faces of a 1U cubesat - Compared to MMRTG: 2× efficiency, potentially 3 orders of magnitude improvement in mass-specific power density (theoretical claim, not demonstrated)
Phase II work in progress: - Growing optimized InAsSb material (0.28 eV bandgap) at RIT via MOVPE - Advancing loss mechanism modeling - Uranus mission Compass study — NASA Compass team at GRC developing Uranus mission context showing TRC enables micro-satellites for outer solar system exploration
Why it matters: At outer planet distances (5–40 AU), solar power is impractical (flux drops as 1/r²) and RTGs are the only option. If TRC achieves the theoretical 12.3% efficiency with no moving parts (unlike Stirling), it would be: (1) more efficient than MMRTG, (2) simpler/more reliable than Stirling, (3) potentially smaller/lighter. Small outer-planet micro-satellites (currently impossible with MMRTG mass) would become feasible.
Confidence: speculative. Phase I demonstrated 0.6 mW/cm² on unoptimized material. Phase II material development is ongoing. The 12.3% efficiency figure is theoretical. No flight-ready device exists.
Document read: Poster file 317451 (session 54). PI: Stephen J. Polly; Geoffrey Landis (GRC) as co-investigator.
Open thread: TRC efficiency advantage over MMRTG is ~2× theoretically. But MMRTG has flight heritage (TRL 9) and a defined supply chain. TRC adoption path likely requires: (1) MOVPE-grown InAsSb demonstration, (2) Compass study validating mission applicability, (3) GCD or MEP follow-on. Currently NIAC-only — no GCD project yet.
Comparison Table¶
| System | Status | Isotope | Conversion | Efficiency | TRL | Program |
|---|---|---|---|---|---|---|
| MMRTG | Operational | Pu-238 | Thermoelectric (SiGe) | ~5.5% | 9 | SMD/baseline |
| Harmonia | Active (GCD) | Am-241 | Stirling | ~20% (target) | 4 | GCD/commercial |
| Small RTGs | Active (MEP) | Pu-238 or Am-241 | Thermoelectric | ~5% | 2 | MEP |
| APPLE | NIAC Phase II (done) | Pu-238 or Am-241 | SKD Thermoelectric (modular tiles) | ~5% (same physics, better geometry) | 3 | NIAC |
| TRC | NIAC Phase II | Pu-238/any | Thermoradiative cell | 12.3% (theory) | concept | NIAC |
| FSP | Active (TDM) | U-235 (fission) | Stirling | ~22% | 4→8 | TDM |
Harmonia We/kg estimate: assume ~15-kg ESG + heat source unit → 40-100 We Stirling output = 2.7–6.7 We/kg; actual spec TBD.
Key insight: Stirling conversion (Harmonia) offers ~4× efficiency improvement over MMRTG thermoelectrics (~20% vs. ~5%), meaning the same thermal power produces 4× the electrical output. Combined with Am-241's potentially commercial supply, Harmonia aims to make RPS economics look very different from the current DoE-controlled MMRTG paradigm.
Am-241 Commercial Supply Chain¶
Am-241 is produced when Pu-241 (from reactor fuel or weapons production) decays (half-life: 14.4 years). The UK has a civil plutonium stockpile from reprocessing (~130 tonnes), which contains significant Am-241 as a decay product. The UK's National Nuclear Laboratory (NNL) has been extracting Am-241 under an ESA contract for a proposed European radioisotope power system. This is the commercial supply chain that Harmonia and the Small RTGs project are betting on.
US path: The US DOE manages Am-241 from weapons complex waste but has not established a dedicated Am-241 production/purification program for space applications. Zeno Power's commercial angle implies they see a path to non-DOE Am-241 sourcing (likely UK/European production or future commercial extraction from spent nuclear fuel).
Supply risk: Am-241 production rates and availability for US commercial space are not established — this is genuinely uncertain. Harmonia's TRL 4→5 milestone for the heat source design is partly about characterizing what isotope purity and form factor is workable before committing to a supply chain.
What Is NOT in TechPort¶
| Domain | Status | Why |
|---|---|---|
| MMRTG program management | Not in TechPort | SMD/RPS program — operational, not STMD |
| Pu-238 production at ORNL | Not in TechPort | DoE program |
| GPHS module qualification | Not in TechPort | DoE/LANL — nuclear program |
| European Am-241 (NNL/ESA) | Not in TechPort | Foreign program |
The MMRTG baseline and the DoE production programs are structurally invisible in TechPort. TechPort only shows the STMD/MEP development layer (Harmonia, Small RTGs, APPLE) and the fission layer (FSP).
Key People¶
- Thomas D. Demichael (GRC/NASA): Harmonia PM
- Mary J. Werkheiser: GCD Program Director (oversees Harmonia)
- Erik J. Brandon (JPL): Small RTGs PI
- Marc A. Gibson (GRC): Kilopower PI — Stirling conversion champion at GRC
Open Threads¶
- Harmonia Am-241 TRL assessment result — The baseline TRL assessment is a first deliverable. If Am-241 heat source comes in below TRL 4, the timeline is at risk. Not yet visible in TechPort.
- Blue Origin's RPS role — Unusual to see Blue Origin in nuclear power integration. They are handling Stirling convertor integration into the RPS — is this a strategic capability build for their lunar lander program?
- Am-241 commercial supply path — US commercial sourcing of Am-241 is the key dependency. Harmonia's success is contingent on this path being real.
- Sunpower Inc. dual role — Same Stirling convertor supplier for both Kilopower (TRL 5, DoE-fuel fission) and Harmonia (TRL 4, Am-241 commercial). Suggests Sunpower is positioning as the standard Stirling supplier for space nuclear.
- Small RTG impactor probes vs. soft landers — Hard landing eliminates EDL cost but limits sensor suite. Worth comparing to InSight as a system-level design choice.
- APPLE Phase II completed (TRL 3) — Document read confirmed: Pu-238 tiles (23 g/We, 15× MMRTG), Am-241 alternative (3× MMRTG), ORNL solid-state battery under radiation qualification, JPL+ORNL team. TRL 3 without a GCD follow-on. No mission pickup visible yet — APPLE remains a Phase II output awaiting a program champion. See full APPLE section above.
- TRC GCD follow-on — TRC NIAC Phase II ends May 2026. If the Uranus Compass study + InAsSb MOVPE demonstration succeeds, watch for a GCD or MEP follow-on proposal. Geoffrey Landis (GRC) as co-investigator is the likely bridge to a program proposal. TRC at TRL 3+ would be the right time for a GCD pickup.
Cross-references¶
- topics/fission-surface-power.md — FSP (fission, 10 kWe, TRL 4→8) — the high-power end of the nuclear power spectrum
- topics/outer-planet-access.md — APPLE in depth; power desert analysis for outer solar system missions
- topics/propulsion-theme.md — NTP/NEP termination; nuclear propulsion context
- topics/cold-atom-quantum-sensing.md — quantum sensors that could be instrument payloads on Small RTG probes
- organizations/usnc-tech.md — EmberCore dynamic RPS (Co-60 Brayton cycle) — another alternative isotope approach
Confidence: confirmed for MMRTG baseline (public knowledge) and Harmonia/Small RTG project details (live TechPort API). Specific power estimates for Harmonia marked as estimates — official specs not yet in TechPort. Am-241 supply chain details from project descriptions + general domain knowledge (NNL/ESA program context).