Perovskite, the material being hailed as the future of solar cells and computer screens, contains previously unrecognized flaws. However, these can be fixed, suggesting perovskites have even more potential than previously suspected—which is saying something given how excited many engineers have become about it.

Perovskite is a mineral that occurs naturally in the Earth's mantle. In recent years, solar cell researchers have realized that compounds designed to mimic perovskite's structure can produce sunlight from electricity.

While initially much less efficient than silicon-based solar panels, perovskite has the potential to be made much more cheaply. Moreover, the efficiency has improved more rapidly than any solar material ever studied, creating so much buzz that Dane deQuilettes, a graduate student at the University of Washington and lead author on the paper, estimates there are a thousand teams worldwide studying its potential. Meanwhile, other potential uses have emerged, including for display screens and lasers.

The best perovskite-based cells are still not as efficient as silicon ones. For most purposes, this will not matter—if solar cells made from perovskite are 80% as efficient as silicon and cost 50% as much to make, they will dominate the market (provided issues such as longevity can be addressed). Still, there are plenty of people keen to see just how far perovskite cells can go.

In the paper published in Science, deQuilettes suggests there is plenty of further potential. Using a technique called confocal optical microscopy to filter out glare, deQuilletes and colleagues searched for imperfect spots in CH₃NH₃PbI₃(Cl) perovskite films. They wanted to see if the boundaries between grains within the films were interfering with the overall efficiency. They found effects larger than they expected.

"Surprisingly, this result shows that even what are being called good, or highly-efficient perovskite films today still are 'bad' compared to what they could be,” says co-author professor David Ginger. “This provides a clear target for future researchers seeking to improve and grow the materials."

The perovskite films are made up of grains, some of which looked dark when examined, indicating inefficient light conversion. The boundaries between grains also showed more interference than anticipated. The cells deQuilletes and Ginger used had efficiencies peaking at 14.5%, while the record for perovskites is 20.1%, but they think the lower efficiency is caused by different architecture rather than a higher concentration of flaws.

Ginger and deQuilletes went further than just identifying its potential. They showed defects were associated with chlorine deficiency and that deposition of pyridine caused dark grains to brighten and partially cured the interference at grain boundaries.

The major problem perovskite cells now face is their tendency to degrade, particularly when exposed to water or changes in temperature. Nevertheless, some companies are claiming that they will have perovskite panels on the market by 2017.

Advances based on the latest discovery may not be incorporated by then, but should be in use soon after. For solar researchers, perovskite is turning out to be an almost perfect offering. After all, the cracks are how the light gets in.