An EPFL research project developed a machine learning-based method to quickly and accurately search large databases, leading to the discovery of 14 new materials for solar cells.
Image: Through the generation of an accurate bandgap dataset of perovskite materials and the use of machine learning techniques, several promising halide perovskites for photovoltaic applications have been identified. Credit: H. Wang (EPFL)
As we incorporate solar energy into our daily lives, it has become important to find materials that efficiently convert sunlight into electricity. Silicon has dominated solar power technology so far, but materials known as perovskites are steadily gaining attention due to their low cost and easy manufacturing process.
But the challenge was finding a perovskite with the right “bandgap.” A bandgap is a specific energy range that determines how efficiently a material can absorb sunlight and convert it into electricity without losing it as heat.
Now, the EPFL research project led by Haiyuan Wang and Alfredo Pasquarello, along with collaborators in Shanghai and Louvain-la-Neuve, will combine advanced computational techniques and machine learning to develop perovskite materials ideal for photovoltaic applications. We have developed a method to explore. This approach could lead to more efficient and cheaper solar panels and change standards in the solar power industry.
The researchers began by developing a comprehensive, high-quality dataset of bandgap values for 246 perovskite materials. The dataset was constructed using advanced computations based on hybrid functionals. Hybrid functionals are an advanced type of calculation that involves electron exchange and are an improvement over the more traditional density functional theory (DFT). DFT is a quantum mechanical modeling technique used to investigate the electronic structure of many-body systems such as atoms and molecules.
The hybrid functional used is “dielectric dependent”, meaning that it incorporates the electronic polarization properties of the material into the calculations. This significantly improved the accuracy of bandgap predictions compared to the standard His DFT. This is particularly important for materials such as perovskites, where electronic interactions and polarization effects are important for their electronic properties.
The resulting data set provides a robust foundation for identifying perovskite materials with optimal electronic properties for applications such as photovoltaics, where precise control of bandgap values is essential to maximize efficiency. provided.
The team then used bandgap calculations to develop a machine learning model trained on 246 perovskites and applied it to a database of approximately 15,000 solar cell candidate materials, selecting the most promising based on their predictions. We narrowed our search to perovskites. Bandgap and stability. The model identified 14 completely new perovskites. All of these have band gaps and sufficiently high energy stability to be excellent candidates for high-efficiency solar cells.
This research uses machine learning to streamline the discovery and validation of new photovoltaic materials, thereby reducing costs, significantly accelerating the deployment of solar energy, reducing dependence on fossil fuels, and reducing climate change. This shows that we can support global efforts to combat this.
other contributors
- shanghai university
- Catholic University of Leuven