Tandem solar cells, marrying perovskite semiconductors with conventional silicon cells, are revolutionizing solar energy efficiency, boasting a remarkable 33% efficiency compared to traditional silicon counterparts. However, challenges in stability and manufacturing hinder their swift market adoption.

Researchers from the Karlsruhe Institute of Technology (KIT), along with experts from Helmholtz Imaging and Helmholtz AI, have made significant strides by harnessing Machine Learning and Artificial Intelligence (AI). They've developed AI models that predict the quality of perovskite layers, crucial for enhancing solar cell efficiency. By analyzing photoluminescence variations in thin perovskite layers during manufacturing, AI can identify hidden signs of coating quality, providing a revolutionary blueprint for precision in materials science.

The interdisciplinary team's innovative approach uses Explainable AI (XAI) to systematically understand the factors influencing coating quality. Through training neural networks on vast datasets of video recordings showing photoluminescence during manufacturing, the AI can predict the efficiency levels of each solar cell. This breakthrough allows researchers to pinpoint parameters requiring adjustment, streamlining experiments and offering a roadmap for future research not only in solar energy but also across various domains in materials science and energy research.

Based on: https://www.sciencedaily.com

Story Source: Materials provided by Karlsruher Institut für Technologie (KIT)