Switching to modern silicon solar cells could help the cost of powering satellites fall by as much as 90 per cent based on beginning-of-life (BOL) performance, according to a review led by the University of Surrey. The switch to silicon cells could also halve the weight of the solar cells needed on the spacecraft, freeing mass for fuel or instruments, or cutting launch costs.

Silicon was the industry-standard solar cell material for spacecraft from 1958 until 1977, when gallium arsenide cells displaced it, offering better efficiency and radiation resistance. Triple-junction cells built from gallium, indium and germanium have been the standard ever since.

Today, silicon heterostructure cells hold a record efficiency level of 27.8 per cent and perovskite/silicon tandem cells hold a record of 34.85 per cent.

In a review published in Acta Astronautica, researchers from Surrey show that triple-junction space cells cost between $250 and $450 per watt, whereas silicon costs tens of cents. As of November 2025, the three main silicon designs (PERC, TOPCon and heterojunction) averaged $0.275, $0.285 and $0.39 per watt respectively.

The raw materials are cheaper too: silicon costs a few dollars per kilogram, while gallium and germanium cost thousands per kilogram.

To make this more tangible for real-world usage, the team from Surrey modelled two formats – one face of a 3U CubeSat (a satellite roughly the size of a large loaf) and a Micro Sat built by Surrey Satellite Technology Limited (SSTL), which co-funds the lead author's PhD and supplied the spacecraft data used in the study. Including the space-qualified glass that protects the cells, the saving was as large as 85 to 90 per cent.

Coverglass, not the cell, dominates the cost of a silicon array.

Tommy Richards, lead author and postgraduate researcher at the University of Surrey's Advanced Technology Institute, said:

"The interesting finding for us was not that silicon is cheaper but where the remaining cost sits. Once you put silicon cells behind space-qualified glass, the glass is what you are paying for.

"That changes what we should be working on. If we can make the cell itself tougher against radiation, we can use thinner glass or substrates, and we cut cost and weight at the same time. It reframes the problem from a materials contest into an engineering one we know how to attack."

In the review, silicon heterojunction cells delivered roughly twice the specific power of the triple-junction option, at around 920 to 1,000 watts per kilogram against 455 to 505. In simple terms, a mission could carry half the solar cell mass for the same power output.

Silicon does produce less power for a given area, and the review puts the reduction at around 28 per cent at the start of a mission. But when the researchers scaled the panels to match triple-junction output, including the cost of the extra panel structure, silicon remained several times cheaper on BOL performance.

However, silicon is not the winner on every measure. For a given coverglass thickness, silicon is around 2.6 times less resistant to radiation than triple-junction cells. Triple-junction cells are also expected to retain more of their performance after five years in orbit, which highlights the importance of improving the radiation resistance of silicon devices.

The number of objects launched each year has risen from around 120 in 2010 to more than 2,800 in 2024. An average satellite needs close to a kilowatt. Proposed space-based solar power stations, which would collect sunlight in orbit and beam it to Earth as microwaves, could require structures kilometres across, delivering gigawatts. The review argues that demand on that scale would consume the entire triple-junction market but barely register against silicon output.

Professor Ravi Silva CBE FREng, Director of the Advanced Technology Institute at the University of Surrey, said:

"We really need to start looking at satellites as infrastructure now. Weather forecasts, navigation, broadband, crop monitoring and disaster response all run through them. Anything that makes a satellite cheaper to build and lighter to launch widens who gets to use that infrastructure, and that includes smaller nations, universities and start-ups rather than only the largest agencies and operators.

"And if we are serious about collecting solar power in orbit and sending it home, we will need solar cells on a scale the specialist space industry simply cannot produce. Silicon is the only material with factories already running at that size. The technology on rooftops and in solar farms could turn out to be the technology that powers the next generation of spacecraft."

The review identifies ultraviolet (UV) light as the least understood risk for bare devices. Modern silicon architectures degrade under UV exposure, and heterojunction cells, the best performers on efficiency, degrade more than PERC cells. Almost all testing so far has used UVA, the lower-energy ultraviolet that reaches the Earth's surface, rather than the more energetic vacuum ultraviolet found in space.

Dr Jae Sung Yun, corresponding author from the University of Surrey's Advanced Technology Institute, said:

"If we are going to collect solar power in orbit and send it back to Earth, we will need solar cells by the square kilometre. Silicon is the one material we already know how to make on that scale, and it is the one we can afford to make that much of."

The review sets out the areas the authors believe need attention: testing modern cells under the high-energy ultraviolet found in space, testing them across the temperature swings of a real orbit and testing them for radiation while they generate power. It also revisits older ideas worth another look, including lithium doping, an approach first reported in 1966 that allowed cells to repair their own radiation damage, but which was later set aside.