Using artificial intelligence, researchers in Sweden have identified the elusive low-temperature structure of a key perovskite material, a breakthrough that could help stabilize next-generation solar cells.
Formamidinium lead iodide, a material that could make solar cells cheaper and more flexible, has stumped scientists for years, but a team at Chalmers University of Technology has now taken a closer look, according to a report on the university's website.
These materials can absorb and emit light with remarkable efficiency, making them promising for thin, lightweight solar panels and even LED devices. But they break down too quickly to be reliable.
By combining traditional computer simulations with machine learning, the researchers mapped a low-temperature phase of the material that had been impossible to pin down with experiments alone.
Their results, published in the Journal of the American Chemical Society, could help stabilize perovskites for wider use.
"The low-temperature phase of this material has long been a missing piece of the research puzzle," said Chalmers researcher Sangita Dutta. "We’ve now settled a fundamental question about the structure of this phase."
The models showed that as the compound cools, molecules become stuck in semi-stable states. To test the findings, the team worked with the University of Birmingham, cooling the material to minus 200 degrees Celsius. The experiments matched the simulations.
Machine learning made the breakthrough possible by enabling the team to run simulations that were significantly longer and model millions of atoms, rather than hundreds. That "brings them closer to the real world," Dutta said.
Julia Wiktor, the study's principal investigator and an associate professor at Chalmers, said the work could help prepare for a future where electricity use is expected to soar.
The International Energy Agency's Net Zero by 2050 scenario projects electricity's share of global final energy use will rise from about 20% today to around 50% by mid-century.
"Our findings are essential to engineer and control one of the most promising solar cell materials for optimal utilization," said Wiktor. "It’s very exciting that we now have simulation methods that can answer questions that were unresolved just a few years ago."
By Handan Kazanci
Anadolu Agency
energy@aa.com.tr
