Next disruption in space business: solar panels

Use of solar energy in space is as almost as old as the space age itself, going back to 4th satellite even launched, Vanguard 1. Indeed, space industry has been an early adopter of solar (PV) panels, providing power for almost all objects send to space during the last sixty years.

Meanwhile, price of PV panels has dropped dramatically, reaching less than fifty cents per watt. But for space applications, that same price has remained "sky-high". Leading manufacturer of space solar panels (Spectrolab, subsidiary of Boeing), estimates price of 400USD/Watt, three orders of magnitude more for high power satellites. For cubesats, prices go up to 1000USD/Watt (for example, using Azur space cells).

This cost is caused by a need to squeeze as much power in available mass budget (typically on the order of 300W/kg), need for high conversion efficiency (typically using triple-junction cells approaching 30%) due to limited "real estate" available to the panels, long operating times (15 years or more), hostile radiation environment and most importantly, by low production volume (that is less than 1MW per year). So combination of these factors leads to high cost factor, that reaches several percentage points of a typical satellite.

So although silicon based PV cells for terrestrial use can achieve 20% efficiency in terrestrial use, due to low volume of production, manufacturers were not particularly interested in providing low-cost space qualified PV panels based on low-cost technology used in terrestrial applications. But several factors might change that situation.

Cubesat market might be an early adopter of such cells, in order to reduce total mission cost and accepting risk of unproved cells. It can be also used as a testbed for innovative solar panel deployment mechanisms that would reduce limited surfaces available for solar panels. But the volume of sales in this market is very small. Geostationary satellites require long life and ever increasing power requirements (especially for HTS applications), so they are not likely adopter of low-cost panels. But upcoming mega-constellations of LEO satellites,  have every reason to look into lowering cost of solar panels. To compete with GEO satellites, they must be produced in much higher numbers than GEO satellites. For example, leading satellite communications provider SES operates 35 geostationary satellites. But its subsidiary O3b operates 12 satellites in MEO and will expand to 20 satellites in the near future. And these satellites have 1500W solar panels and mass of 700kg (roughly five times less than typical GEO satellite).

Upcoming OneWeb constellation will even more illustrate growing power requirements for satellites. If we take assumed 200kg of satellite weight and scale down power requirements to 400W, and multiply that with targeted price of the satellite of five million USD, Triple junction PV cells would represent 3% of the total satellite cost. Total value of all PV cells for the whole constellation would be over 100 million USD. That is not a small change even for multi-billion dollar investment.

The second force that could influence migration to low cost PV panels is dropping cost of satellite launches, primarily driven by SpaceX reusability. GTO launch used to cost in the range 100-200 million USD. Current medium sized GEO satellite can be launched on reused F9 booster for 50 million, and it is likely that this price will slowly drop further towards 30 million USD. Therefore satellite manufacturers will feel pressure to reduce their prices from 100-300 million USD downward, while power requirements will go upward.

The last blow might come from introduction of ITS SSTO as Falcon 9 replacement. Although such launch vehicle would be equal to F9 in LEO capacity and without GTO capability, it would bring another huge advantage - much more available space under its 12 meter wide body (compared to 5m standard fairings used today). That would enable to at least double size of fixed/rigid PV panels, easily compensating efficiency disadvantage.

In any case, manufacturers of space qualified PV cells will be under increasing price pressure, as their terrestrial competition will be "good enough" for more and more space applications. Similar to radiation hardened electronics, it will become more and more limited to special applications. Whether this will be triple junction Ga-As or crystalline silicon, it is not easy to predict. But it will not remain a thousand times more expensive than terrestrial applications.