4 Minutes
Rising global electricity demand is intensifying the search for next-generation solar materials that are efficient, lightweight and low-cost. Researchers at Chalmers University of Technology in Sweden have combined machine learning with large-scale atomistic simulations to clarify a previously unresolved low-temperature structure of formamidinium lead iodide, a leading halide perovskite candidate for advanced solar cells.
Formamidinium lead iodide is considered one of the best-performing materials in the halide perovskite group, since it has promising properties for future solar cell technologies. New findings from Chalmers can now shed light on its structure; this is crucial if we are to engineer and control the material. Credit: Chalmers
Scientific background
Halide perovskites are a family of crystalline semiconductors that have transformed photovoltaic research over the past decade because of their strong light absorption, tunable bandgaps and potential for low-cost manufacturing. Formamidinium lead iodide (often abbreviated FAPbI3) stands out for excellent optoelectronic properties but has been limited by structural instability and degradation under operating conditions. Understanding the atomic-scale phases and molecular behavior inside the crystal lattice is essential to design stable, high-efficiency perovskite solar cells.
AI-driven simulations and key findings
Chalmers researchers used validated computational models augmented with machine learning potentials to extend simulation times and length scales by several orders of magnitude. These hybrid methods let the team model systems containing millions of atoms and explore phase transitions that were previously inaccessible to conventional first-principles calculations.
Why machine learning matters
Machine-learned interatomic potentials dramatically reduce the computational cost of molecular dynamics while retaining chemical accuracy. That enabled the team to follow FAPbI3 as it cooled and to observe how formamidinium molecules adopt semi-stable orientations within the lattice — a behavior linked to the material's low-temperature phase and to mechanisms that can influence stability and electronic properties.
Resolved structural question
The simulations identified the detailed arrangement of atoms in the elusive low-temperature phase and showed that organic cations can become trapped in a semi-stable configuration during cooling. This structural picture closes a long-standing gap in the fundamental understanding of FAPbI3 and provides parameters that experimentalists and device engineers can use to improve material mixtures and processing strategies.
"We now have simulation tools that can answer questions which were out of reach just a few years ago," says Julia Wiktor, associate professor and principal investigator at Chalmers. Sangita Dutta, a Chalmers researcher on the project, adds that resolving the low-temperature phase removes a critical unknown for material design.
Experimental validation and implications
To confirm the models, collaborators at the University of Birmingham cooled samples to around –200°C and compared lab measurements with simulated signatures. The experimental data matched the predicted structural motifs, strengthening confidence in the combined computational–experimental approach.
These insights affect how researchers tune compositions (for example, mixing different halide perovskites) to reduce degradation and improve stability in photovoltaic modules and optoelectronic devices like LEDs. By unlocking microscopic behavior, the study paves the way for more predictable materials engineering and for scaling perovskite technologies into flexible, thin-film solar panels that could be integrated into electronics and buildings.
Conclusion
The Chalmers work demonstrates how machine learning and large-scale simulation can resolve complex phase behavior in halide perovskites. With validated atomic-scale models for formamidinium lead iodide, researchers gain a clearer path to controlling stability and performance — a crucial advance for next-generation solar cells and related optoelectronic technologies.
Source: scitechdaily
Comments