Spacecraft Power Systems: Solar Arrays, RTGs, and the Physics of Energy in Space

A spacecraft without electrical power is a tumbling piece of metal. Every subsystem — communications, attitude control, science instruments, thermal heaters, onboard computers — requires electricity. Designing the power system that reliably provides that electricity across mission durations of years to decades, in environments ranging from Mercury’s dayside to the outer solar system, is one of the foundational challenges of spacecraft engineering.

Two technologies dominate: photovoltaic solar arrays and radioisotope thermoelectric generators (RTGs). Each has physical constraints that make it the correct choice for certain mission profiles and unworkable for others. A third category — nuclear fission reactors — is emerging for high-power missions where neither solar nor RTG is adequate.

Key parameters

Technology	Power density	Best missions	Limitation
Silicon solar cells (1 AU)	150–180 W/kg (BOL)	Inner solar system	Drops as r² beyond ~3 AU
Triple-junction GaAs (1 AU)	250–300 W/kg (BOL)	GEO, LEO	Cost; same r² limitation
MMRTG (Pu-238)	~2.8 W/kg (beginning)	Outer solar system, Mars	Decays ~1.5%/year
GPHS-RTG (Pu-238)	~5.1 W/kg (beginning)	Outer solar system	Discontinued Pu-238 supply
Kilopower (fission)	~40 W/kg target	Planetary surfaces, deep space	TRL 5–6 (2026)
ISS solar arrays (total)	84–120 kW	LEO reference	Degraded from ~160 kW at launch

---

Solar Cells: Beginning-of-Life vs End-of-Life Power

Spacecraft solar arrays are specified at two conditions: beginning-of-life (BOL) power, measured immediately after launch, and end-of-life (EOL) power, after mission duration exposure to the space environment.

The gap between BOL and EOL is driven by four degradation mechanisms:

Particle radiation damage: Electrons and protons in the Van Allen belts and galactic cosmic rays displace atoms in the semiconductor crystal lattice, creating defects that reduce charge carrier mobility and minority carrier lifetime. The effect is quantified by equivalent 1 MeV electron fluence. A GEO satellite in a 15-year mission accumulates fluences of 10¹⁵ electrons/cm² — sufficient to degrade triple-junction GaAs solar cells to approximately 70–80% of BOL performance.

Ultraviolet darkening: UV photons break chemical bonds in the transparent encapsulant and coverglass coatings over the cells, increasing absorption and reducing transmitted light. This is a secondary effect compared to particle damage for most missions.

Thermal cycling: Each orbit through eclipse and sunlight creates a temperature swing that fatigues the solder bonds and interconnects between cells. A LEO spacecraft cycles 5,800 times per year. After 15 years in LEO, thermal fatigue can contribute 3–5% additional power loss beyond radiation damage.

Coverglass micrometeorite erosion: Incremental surface roughening from submicrometre debris impacts, relevant primarily for missions exceeding 10 years in debris-dense environments.

Mission power system sizing therefore begins with EOL requirements — the power needed at the worst mission point — and works backward: EOL power divided by the accumulated degradation factor gives the BOL array size, which determines area and mass.

The International Space Station solar arrays illustrate the magnitude: originally 262,400 individual silicon solar cells across eight wings, rated at ~160 kW BOL. After more than two decades in LEO, output had degraded to ~84 kW before the 2021 iROSA (ISS Roll-Out Solar Array) upgrades added supplemental GaAs arrays restoring total capacity to approximately 120 kW.

---

The 1/r² Problem: Solar Power at Distance

Solar irradiance follows an inverse-square law with distance from the Sun. At Earth (1 AU), the solar constant is 1,361 W/m². Beyond Earth, power availability collapses:

Distance	Solar constant	Relative to Earth
Venus (0.72 AU)	2,600 W/m²	×1.91
Earth (1.00 AU)	1,361 W/m²	×1.00
Mars (1.52 AU)	589 W/m²	×0.43
Jupiter (5.2 AU)	50 W/m²	×0.037
Saturn (9.5 AU)	15 W/m²	×0.011
Pluto (39 AU)	0.9 W/m²	×0.00066

This is why solar power works at Mercury, Venus, Earth, and Mars — and becomes impractical beyond the asteroid belt without array areas that cannot be practically launched. Juno, in Jupiter orbit, has the largest solar arrays of any deep-space spacecraft: three 2.9 m × 8.9 m wings producing ~14 kW at Jupiter’s average distance. Europa Clipper, launched in 2024, carries arrays producing ~700 W at Europa — the practical limit of solar power in the Jovian system.

BepiColombo demonstrates the opposite problem: at Mercury (0.31–0.47 AU from the Sun), solar irradiance reaches 5,000–14,000 W/m². Arrays sized for Mercury cannot face the Sun directly; BepiColombo operates with its arrays tilted and thermally controlled to prevent damage from overheating.

---

RTGs: Nuclear Power for the Outer Solar System

A radioisotope thermoelectric generator converts the heat from radioactive decay directly into electricity using thermoelectric junctions (Seebeck effect). No moving parts. No pointing requirements. Power production determined by the decay rate of the fuel, not distance from the Sun.

The fuel of choice is plutonium-238, with a half-life of 87.7 years. It decays by alpha emission (4.27 MeV per decay), producing approximately 0.56 W of thermal power per gram of ²³⁸Pu at beginning-of-life. The long half-life means power declines by approximately 0.787% per year — slow enough that missions of 10–20 years are feasible.

The General Purpose Heat Source (GPHS): The standard modular unit of RTG fuel, each containing 151 grams of ²³⁸PuO₂ in a four-layer impact-resistant container (iridium, graphitic carbon, aeroshell). Each GPHS module produces ~250 W thermal at BOL. Multiple modules are assembled into RTGs.

The Multi-Mission RTG (MMRTG): The current US RTG design, used on Curiosity and Perseverance. Eight GPHS modules, ~2,000 W thermal BOL, 110 W electrical BOL. By Perseverance’s launch (2020), anticipated output was ~110 W. After a 14-year design life, projected output ~100 W. The thermoelectric conversion efficiency is approximately 6% — low, but thermocouples have no moving parts and can operate indefinitely.

The GPHS-RTG: The older, higher-power design used on Cassini (3 units, 885 W total), New Horizons (1 unit, 245 W at launch, ~170 W at Pluto in 2015), and Ulysses. No longer manufactured due to the discontinuation of US Pu-238 production in 1988; the Department of Energy restarted production in 2013 at Oak Ridge National Laboratory, reaching approximately 1.5 kg/year by 2023 with a target of 1.5 kg/year sustained.

The Voyager spacecraft, launched in 1977 with GPHS-RTG predecessors, still operated at approximately 40 W each in 2024 — generating power from fuel that has now decayed for 47 years.

---

Thermoelectric Efficiency and Waste Heat

The low conversion efficiency (~6%) means the majority of RTG thermal power is rejected as waste heat. For the MMRTG at ~2,000 W thermal and 110 W electrical, approximately 1,890 W is dumped as heat — which must be managed by the spacecraft thermal control system. On Mars, Curiosity and Perseverance benefit from this “waste” heat warming critical electronics during the frigid Martian nights (−70°C surface temperature). On outer planet missions, thermal management of RTG heat output is part of the spacecraft thermal design.

The theoretical efficiency limit for a thermoelectric device operating between temperature extremes is set by the Carnot efficiency:

$\eta_\text{Carnot} = 1 - \frac{T_\text{cold}}{T_\text{hot}}$

For an MMRTG with T_hot ≈ 800 K (fuel surface) and T_cold ≈ 500 K (cold-side junction), the Carnot limit is ~37.5%. The actual 6% reflects the poor thermoelectric figure of merit (ZT ≈ 0.6–0.8) of the silicon-germanium and lead-telluride thermocouple materials used. Advanced thermoelectric materials with ZT > 2 are being developed; if integrated into future RTGs, conversion efficiency could approach 15–20%.

---

Kilopower: Fission for High-Power Missions

RTGs are limited to tens or hundreds of watts. Missions requiring kilowatts — electric propulsion spacecraft, crewed planetary bases, high-power radar instruments — require nuclear fission reactors.

NASA and the Department of Energy’s Kilopower project demonstrated the Kilopower Reactor Using Stirling Technology (KRUSTY) experiment in 2018, producing 1 kW of electrical power from a uranium-235 reactor using Stirling cycle heat engines. The Stirling cycle achieves ~30% thermal-to-electric efficiency — far higher than thermoelectrics.

The current development target is a 10 kW system suitable for long-duration surface power on the Moon or Mars, or for powering high-power Hall thrusters on deep-space spacecraft. DARPA’s DRACO programme targets a nuclear thermal rocket using a similar reactor as a heat source rather than a power source; the physics overlap is substantial.

At 10 kW, a Kilopower-class system could power an electric propulsion stage producing continuous thrust sufficient for interplanetary trajectories. This is the architecture under consideration for future outer solar system missions where neither solar power nor RTG provides adequate energy at reasonable mass.

For the electric propulsion systems that high-power spacecraft power enables, see ion drives and Hall thrusters. For the radiation environment that threatens solar array cells and RTG electronics alike, see space radiation and the Van Allen belts.