‘Machine learning accelerates the discovery of perovskites on solar cells’
By generating a dataset of accurate band gaps for perovskite materials and using machine learning, several promising halide perovskites have been identified for photovoltaic applications (Illustration: Haiyuan Wang, Swiss Federal Institute of Technology Lausanne)
May 20, 2024 – Researchers from the Swiss EPFL (École Polytechnique Fédérale de Lausanne) have developed a method to quickly discover/develop new perovskite materials for photovoltaic applications. The method has now led to the discovery of fourteen new, promising materials for solar cells.
The model is based on advanced computational techniques and machine learning, and enables accurate and targeted searches of large data sets. According to EPFL, the research shows that machine learning can be used excellently to develop new photovoltaic materials, reducing costs and significantly improving the acceptance of solar energy. According to the Swiss research institute, this in turn contributes to reducing dependence on fossil fuels and helps to combat climate change.
