Perovskite + silicon solar panels hit efficiencies of over 30%

In most industrialized countries, solar panels account for only a quarter to a third of the overall cost of building a solar farm. All the other expenses—additional hardware, financing, installation, permitting, etc—make up the bulk of the cost. To make the most of all these other costs, it makes sense to pay a bit more to install efficient panels that convert more of the incoming light into electricity.

Unfortunately, the cutting edge of silicon panels is already at about 25 percent efficiency, and there's no way to push the material past 29 percent. And there's an immense jump in price between those and the sorts of specialized, hyper-efficient photovoltaic hardware we use in space.

Those pricey panels have three layers of photovoltaic materials, each tuned to a different wavelength of light. So to hit something in between on the cost/efficiency scale, it makes sense to develop a two-layer device. This week saw some progress in that regard, with two separate reports of two-layer perovskite/silicon solar cells with efficiencies of well above 30 percent. Right now, they don't last long enough to be useful, but they may point the way toward developing better materials.

Wearing layers

The idea behind two-layer—called tandem—photovoltaic devices is very simple. The top layer should absorb high-energy photons and convert them to electricity while remaining transparent to other wavelengths. Then, the layer underneath it should absorb lower energy photons. Silicon, which tends to have peak absorption toward the red end of the spectrum, is a great candidate for the lower layer. That leaves the question of what might make sense to put on top of it.

Perovskites make an appealing candidate. They're an entire class of materials that are defined by the structure of the crystals they can form; they can be made from a huge variety of unrelated chemicals. That has some considerable advantages since it means you can potentially identify some very inexpensive source materials that can combine into a perovskite crystal. Many perovskites will also readily form from a solution of the raw materials, potentially allowing us to put a photovoltaic perovskite coating on a huge range of hardware.

The big problem has been that a lot of these crystals aren't especially stable and will break down into raw materials over time. And that time can be as little as weeks to months for some of the more promising materials. There has been some progress in extending their lifespan, but we're still not at the point where it makes sense to manufacture perovskite panels.

The other good thing about perovskites is that, by choosing the raw materials carefully, you can tune the peak wavelength absorbed by the resulting crystal. So you can pick a wavelength that pairs well with silicon. And there have been a few demonstrations that tandem perovskite/silicon cells work, but the efficiencies haven't been much above what silicon can achieve on its own.

So nice, people did it twice

The latest edition of Science features two papers reporting much higher efficiencies from perovskite/silicon tandems. The papers use very different methods to get there but inadvertently end up in similar places.

One of the two, from a large team based in Europe, focuses on the physical structure of the panel. Some high-performance silicon panels have surfaces etched with countless microscopic pyramids. These function to increase the total light absorbed, since any photons that happen to be reflected by one of the pyramids are likely to end up hitting a second, increasing the chances they'll eventually be absorbed. But coating these with a layer of perovskite tends to simply fill in the gaps between the pyramids and then produce a flat surface above that.

The goal of the work was to find a way to get the perovskite to conform to the silicon surface and thus form pyramids on top of the silicon ones. To do that, the researchers tested a variety of different additives to the solution containing the perovskite's raw ingredients. They eventually settled on something called 2,3,4,5,6-pentafluoro-benzylphosphonic acid—basically a benzene ring with one carbon linked to a phosphate, and the rest linked to fluorine. This slowed down the crystallization process, which allowed the perovskite to coat the silicon evenly, reproducing its sea of pyramids.

During this process, however, the chemical additive was squeezed out of the perovskite crystals and ended up coating their surface. And once it was there, it helped mitigate defects where electrons get trapped, allowing more of them to make their way usefully into the current collector instead of falling back into an orbital in the perovskite. The net result was an efficiency of over 31 percent.

The other work, also coming out of a European collaboration, was focused on optimizing the combination of silicon and perovskite. It's possible to do calculations that tell us what wavelength the perovskite's peak absorption should be in order to maximize the range of light that's converted into electricity. From there, we can figure out what chemical formula you need to get that.

With that information in hand, the researchers then optimized the interface between the perovskite and the current collector, intentionally trying to limit the loss of useful electrons—something the other group had accomplished inadvertently. To do so, they added an organic molecule that could accept or donate electrons and so could serve as a holding area as the electrons find their way to the current collector.

The net result was a perovskite that, on its own, had an efficiency of over 20 percent. When combined with silicon into a tandem device, the efficiency cleared 32 percent.

Work to be done

The good thing is that there's a lot of headroom left, as calculations based on these devices suggest that they can hit efficiencies in the neighborhood of 45 percent. In any case, they're already considerably more efficient than silicon alone, and perovskites retained their advantages of being cheap and easy to work with.

The big problem is that the devices are horrifically short-lived. Even the most stable device made by the first group had dropped to 80 percent of its original efficiency after just 66 hours of exposure to sunlight. The second was somewhat better, managing to reach 347 hours before dropping below 80 percent. Assuming 12 hours of sunlight a day, however, that translates to less than a month of use, which is terrible.

We do know how to make perovskites that last longer than this. But it's not clear whether those are compatible with decent efficiencies in a tandem configuration. So there's a lot of work left to do before we try to commercialize these things, and there's a chance that some other tandem tech will work out sooner. But the work is likely to go on, as higher-efficiency panels could go a long way toward getting renewables to expand at the rates we need them to.

Science, 2023. DOI: 10.1126/science.adf5872, 10.1126/science.adg0091  (About DOIs).