Can you use solar panels in space? Short answer: yes, and they work better up there than on your roof. Orbit delivers 1,361 W/m2 of unfiltered sun versus 1,000 at sea level, about 36% more, and the ISS's eight array wings generate 84-120 kW peak. The catch isn't power; it's the hardware. Spacecraft fly multi-junction gallium arsenide cells at $200-400/W, not $0.10-0.30/W rooftop silicon. No satellite, rover, or station has run more than a few weeks without panels since Vanguard 1 in 1958.
I've never built hardware for orbit (yet), but I spent a summer at university running radiation testing on cover glass samples, and the lesson stuck: every assumption about panel longevity from terrestrial work goes out the window once you're outside Earth's magnetosphere. Cells that last 30 years on a roof can degrade past usability in five years in geostationary orbit.
Do Solar Panels Actually Work in Space?
Yes, and they perform significantly better in space than on Earth for incoming light. The solar constant at 1 AU is 1,361 W/m2 (NASA SORCE TIM, uncertainty under 0.1%), versus only 1,000 W/m2 at Earth's surface after atmospheric absorption. So a panel in low Earth orbit (LEO) receives roughly 36% more direct beam radiation than the same panel on a roof, with no clouds, rain, or soiling until you reach Mars regolith. LEO day-night cycles run about 90 minutes (45 sun, 45 shadow), so spacecraft need battery backup, but panels work flat-out whenever the sun is in view. The ISS arrays cycle this way 16 times a day.
The ISS's eight solar array wings cover roughly 2,500 m2 and generate 84-120 kW peak depending on sun angle (NASA); new iROSA roll-out panels installed 2021-2023 pushed peak capacity higher. The catch isn't power production. It's everything else: thermal cycling between -160 deg C in shadow and +120 deg C in sun, atomic oxygen erosion in LEO, micrometeoroid impacts, and radiation damage that compounds over years.
What Kind of Solar Cells Does NASA Actually Use?
NASA and commercial operators almost universally use multi-junction gallium arsenide cells, not silicon. NREL's chart records 47.6% for six-junction GaAs under concentrated sunlight and 32-34% for production triple-junction cells, versus 27.3% (lab) and 22-23% (modules) for the best silicon. Three reasons drive the choice: higher efficiency means less mass and area per watt (launch mass runs $1,500-$5,000 per kg to LEO); multi-junction stacks (typically GaInP / GaAs / Ge) absorb a wider band of the unfiltered spectrum than single-bandgap silicon; and GaAs tolerates displacement damage from high-energy protons better than silicon.
The cost is brutal: space-rated cells run $200-400/W versus $0.10-0.30/W for residential silicon, so a 50 m2 satellite array can cost $5-10 million in cells alone, before structure or testing. Some CubeSats and short missions still use silicon where cost beats efficiency (the Soviet Salyut stations did, as did most early-2000s LEO missions), but anything modern with a serious power budget defaults to GaAs.
How Do Mars Rovers and Outer-Planet Probes Use Solar?
Mars rovers split between solar and nuclear by mission profile. Spirit and Opportunity (2003), Phoenix (2007), and InSight (2018) ran on solar; Curiosity (2012) and Perseverance (2020) use radioisotope thermoelectric generators (RTGs) for long durations, high latitudes, and dust-storm survivability.
Solar on Mars is harder. The planet sits 1.5 AU out, so top-of-atmosphere irradiance is roughly 590 W/m2 versus Earth's 1,361. Dust cuts surface insolation further, and during global storms the sky turns orange and irradiance can drop below 100 W/m2 for weeks. Opportunity died from dust accumulation in the 2018 global storm when power fell below the level needed to keep its heaters running and it froze; NASA's last contact was June 10, 2018, after 14 years against a 90-day design life (NASA JPL). Spirit met a slower end, stuck in soft regolith and unable to keep a sun-facing tilt.
Juno is the outlier that broke conventional thinking. Launched 2011, it reached Jupiter in 2016 at 5.2 AU where sunlight is just 50 W/m2, about 4% of Earth's. Wisdom said missions beyond Mars needed RTGs, but three 9 m arrays totalling 60 m2 powered a complex spacecraft on the equivalent of dim twilight, generating roughly 14 kW at Earth's orbit but only 400-500W at Jupiter. The trade-off was a structural challenge: the arrays had to deploy reliably and survive Jovian radiation, among the harshest environments in the solar system.
Why Does Radiation Damage Matter in Orbit?
High-energy particles from the sun and cosmic rays continuously bombard space hardware. Damage splits two ways: total ionizing dose shifts threshold voltages and raises dark current, while displacement damage knocks atoms out of crystal lattices, dropping short-circuit current. Geostationary satellites (35,786 km) sit inside the outer Van Allen belt and accumulate damage faster: a 15-year GEO mission loses 15-20% of array output by end-of-life, most of it in the first 3-5 years. Space cells fight back with thicker cover glass (cerium-doped, 100-150 micrometers) to absorb low-energy protons, plus junction depths and base doping that tolerate displacement damage. LEO is friendlier because Earth's magnetosphere shields most cosmic rays, but beyond geosynchronous altitude shielding becomes the dominant design constraint.
Could You Just Bolt a Silicon Panel to a Satellite?
Technically yes, practically no. A residential silicon panel would produce roughly 36% more power in orbit thanks to higher irradiance. The problem is everything else. Residential panels run 11-13 kg per m2 versus 3-5 for space-rated GaAs. They're rated to -40 to +85 deg C at 200 cycles a year; ISS panels see -160 to +120 deg C at 5,840 cycles a year. Residential encapsulants (EVA, POE) outgas organic compounds in vacuum that contaminate spacecraft surfaces, and unfiltered space UV browns standard EVA within months. The cells are the cheap part of a mission; launch mass, testing, and deployment mechanisms are the expensive parts, so flying unqualified hardware is the costliest way to save a few thousand dollars.
For the silicon manufacturing that makes cheap residential panels possible, see what solar panels are made of. On the ground that hardware pairs with mass-market electronics like the SMA Sunny Boy 6.0-US, the opposite end of the cost curve.
What About Solar Sails and Space-Based Solar Power?
Two often-confused concepts. Solar sails aren't solar panels: they use radiation pressure (photon momentum) for propulsion, not electricity, demonstrated by NASA's NEA Scout (2022) and ACS3 (2024). Space-based solar power (SBSP) collects sunlight in orbit and beams it to Earth via microwaves or lasers; JAXA and Caltech showed small-scale wireless transmission in 2023-2024. The economics stay weak: terrestrial solar runs $0.025-0.04/kWh installed while plausible SBSP targets $0.10-0.30/kWh delivered. Until launch costs collapse another order of magnitude, SBSP stays a science problem. For context on surface sunlight, see how much solar energy hits the Earth.
Summary
Solar panels work in space and power most spacecraft from the ISS to Juno at Jupiter. The cells aren't rooftop silicon; they're multi-junction gallium arsenide at $200-400 per watt, built for high efficiency, broad spectrum response, and radiation tolerance. Mars rovers Spirit, Opportunity, Phoenix, and InSight ran on solar until dust or mechanical failure ended them, while Curiosity and Perseverance moved to RTGs. Geostationary satellites lose 15-20% of array output across 15-year missions despite hardened design. Space solar is a different engineering problem from rooftop solar: pricier cells, a much harsher environment, but far more sunlight. For where the field is heading, see our solar technology advancements review.