Solar cells that actually work 10 meters underwater

I read about this in a study published in Joule last week. It's a nice example of solving a problem by changing the assumption instead of pushing harder on the same idea.

The problem is power. Underwater robots and sensors run out of it, and then you either haul them up, swap the batteries, or run a cable down to them. Solar looks hopeless at that depth because water absorbs sunlight. But it doesn't absorb all of it the same way. Red and infrared are gone in the first few meters. What's still there at 10 meters is mostly blue-green.

A normal silicon solar cell is designed for the full spectrum you get at the surface, so underwater most of that design does nothing. The researchers used a perovskite cell instead, with a wide bandgap of around 1.96 eV, tuned to the light that actually makes it that far down. In tests that simulated the light at 10 meters, it hit 34.71 percent efficiency. That's better than a good rooftop panel does in full sun.

Then they put it in real water, 10 meters down off the Weizhou Islands in the South China Sea, attached to underwater robots. The modules produced 324 mWh in two hours, which was enough to charge lithium-ion batteries and run LEDs. A sealed module barely degraded after about 1,160 hours underwater, and they estimate it would last roughly five and a half years.

Earlier work on this only got to about two meters. Ten meters is a depth where real equipment actually sits, so this is the version of the idea that could be useful.

Source: https://techxplore.com/news/2026-09-underwater-solar-cells-meters-ocean.html