Solar cells that learn to work underwater

A tuned perovskite crystal harvests blue-green light 10 meters down in the South China Sea

Can solar panels work underwater? For a long time the practical answer was “barely, and only near the surface.” A team at Yunnan University in China has now pushed that limit: their perovskite solar cells were tested 10 meters (32 feet) deep in the South China Sea, where they generated enough electricity to charge lithium-ion batteries and light an LED sign. The results, published on September 11, 2026 in the journal “Joule”, are described by the lead researchers as the first functional validation of submerged solar cells at that depth. Earlier studies, they note, had stayed within about two meters of the surface.

Why is solar power so hard to harvest underwater

Sunlight loses strength quickly in water. Part of it bounces off the surface, and the rest is absorbed and scattered over much shorter distances than in air. The loss is not even across the spectrum: red light is absorbed within the first few meters, while blue-green wavelengths, roughly 400 to 600 nanometers, travel deepest in clear seawater. That is a problem for conventional silicon cells. Silicon has a band gap of about 1.1 electronvolts, which suits the red and near-infrared light that water strips out first. A standard panel lowered into the sea is tuned for the wrong part of the spectrum.

How do perovskite cells capture blue-green light

The Yunnan team used lead halide perovskite, a family of semiconductor crystals whose band gap can be adjusted by changing their chemical composition. The band gap is the minimum energy a photon needs to free an electron and produce current. Here it was set to approximately 1.96 electronvolts, which corresponds to a cutoff around 630 nanometers. The cells therefore absorb the blue-to-orange light that survives underwater, and they waste less energy as heat than silicon would when absorbing those higher-energy photons. While the perovskite was forming, the researchers also mixed in a polymer, polyhexamethylene guanidine hydrochloride, which helps trap more electrons in the cell.

To test the design before going to sea, they built a laboratory system with optical filters that reproduces underwater illumination at different depths. Under simulated light at 10 meters, the cells reached a power conversion efficiency of almost 35 percent. That figure needs careful reading: efficiency depends on the light source, so it cannot be compared directly with the efficiency of a rooftop panel in full sunlight. It says the cell converts a large share of the *available* light into electricity, not that it produces a lot of power. What did the sea test show? For the field trial, remotely operated robots carried 115 cm² cells to depths of 2, 6 and 10 meters off the Weizhou Islands, in the South China Sea. The harvested energy fell steeply with depth, as expected:

• At 2 meters: 1,416 milliwatt-hours (mWh)
• At 6 meters: 752 mWh
• At 10 meters: 324 mWh

The 324 mWh figure refers to a two-hour submersion at 10 meters, enough to charge lithium-ion batteries. Public summaries of the study do not detail the duration of the other runs, so the numbers are best read as an indication of the trend, not as a power rating. At every depth the batteries stored enough energy to light an LED sign. The team also reports having scaled from small laboratory cells to large-area modules.

What could underwater solar cells power

Much ocean technology runs on batteries that must be replaced or on cables that are expensive to lay and maintain. Underwater photovoltaics could supply stationary sensors and detectors, cameras, communication systems and acoustic monitoring stations, and possibly reef-monitoring networks or even underwater habitats. The authors present the technology as a complement to other power sources, not a replacement for all of them. The researchers estimate that a cell could operate continuously at 10 meters for about five and a half years. If that holds up, remote installations would need far fewer maintenance visits.

The step from a promising experiment to a deployable product is large. On durability, the team reports almost no degradation after 1,160 hours under simulated 10-meter conditions, and cells stored for 300 days in a nitrogen-filled glovebox kept about 96 percent of their initial efficiency. The five-and-a-half-year figure is an estimate, and the documented sea trial at 10 meters lasted two hours, so long-term field data are still needed. The team says it plans to test deeper waters, to find the limit at which underwater solar cells can operate, and to establish standardized testing protocols for underwater photovoltaics. Beyond those stated goals, some challenges are general to perovskite technology in the sea, not findings of this study: