Proceedings of International Conference on Hybrid and Organic Photovoltaics (HOPV24)
Publication date: 6th February 2024
The optimization of the thickness of the perovskite material, electron transport layer (ETL), and hole transport layer (HTL) is essential for enhancing the efficiency and performance of perovskite solar cells (PSCs). However, experimental testing to determine the optimal thickness of the transport layers can be time-consuming and costly. Bayesian optimization (BO), a sequential machine learning algorithm, is used to determine the optimum parameters in problems with large parameter space[1],[2]. In this study, we compare the BO recommendation of the optimal SnO2 thickness for ETL with the results of repeated laboratory testing to decide if the BO results are a reliable and economic indicator for the optimum thickness of SnO2 used in the PSCs. The structure of the PSC used in the study was determined as ITO/SnO2/FAMAPbI3/MoO3/Ag. We tested spin coating rates ranging from 3500 rpm to 5000 rpm for casting of SnO2 layer. The highest efficiency achieved for the PSC was 12.2 % at a spin coating rate of 4500 rpm for SnO2 layer. The machine learning results are consistent with these experimental findings and have the advantage of being time saving and cost efficient.
We would like to acknowledge the Presidency of Turkey, the Department of Strategy and Budget for infrastructure and, some of consumables (project #: 2016 K121200 - 16DPT002). Basak Turgut thanks the project support of the Scientific and Technological Research Council of Turkey (TUBITAK) (2211-C) for the financial support and some of consumables for the device fabrications. Burak Kahraman also thanks the project support of the Scientific and Technological Research Council of Turkey (TUBITAK) (2211-C) for the financial support and some of consumables for the device fabrications.