Machine learning (ML)-assisted optimization doping of KI in MAPbI 3 solar cells
- PDF / 1,517,694 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 6 Downloads / 184 Views
RARE METALS
ORIGINAL ARTICLE
Machine learning (ML)-assisted optimization doping of KI in MAPbI3 solar cells Sheng Jiang, Cun-Cun Wu, Fan Li, Yu-Qing Zhang, Ze-Hao Zhang, Qiao-Hui Zhang, Zhi-Jian Chen, Bo Qu, Li-Xin Xiao*, Min-Lin Jiang*
Received: 1 June 2020 / Revised: 10 August 2020 / Accepted: 22 August 2020 Ó GRINM Bohan (Beijing) Publishing Co., Ltd 2020
Abstract Perovskite solar cells have drawn extensive attention in the photovoltaic (PV) field due to their rapidly increasing efficiency. Recently, additives have become necessary for the fabrication of highly efficient perovskite solar cells (PSCs). Additionally, alkali metal doping has been an effective method to decrease the defect density in the perovskite film. However, the traditional trial-and-error method to find the optimal doping concentration is timeconsuming and needs a significant amount of raw materials. In this work, in order to explore new ways of facilitating the process of finding the optimal doping concentration in perovskite solar cells, we applied a machine learning (ML) approach to assist the optimization of KI doping in MAPbI3 solar cells. With the aid of ML technique, we quickly found that 3% KI doping could further improve the efficiency of MAPbI3 solar cells. As a result, a highest efficiency of 20.91% has been obtained for MAPbI3 solar cells. Keywords Perovskite solar cell; Machine learning; KI; Doping
S. Jiang, F. Li, M.-L. Jiang* Institute for Advanced Study, Nanchang University, Jiangxi 330031, China e-mail: [email protected] S. Jiang School of Materials Science and Engineering, Nanchang University, Jiangxi 330031, China C.-C. Wu, Y.-Q. Zhang, Z.-H. Zhang, Q.-H. Zhang, Z.-J. Chen, B. Qu, L.-X. Xiao* State Key Laboratory for Artificial Microstructures and Mesoscopic Physics, Department of Physics, Peking University, Beijing 100871, China e-mail: [email protected]
1 Introduction Owing to the efforts of researchers and the superior properties of perovskite solar cells (PSCs) such as high absorption coefficient [1, 2], long carrier diffusion length [3, 4] and tunable bandgap [5, 6], the efficiency of PSCs has been rapidly improved from 3.8% to 25.2% [7, 8] in the past decade. Moreover, solution-based processes with cheap raw materials render PSC as a promising photovoltaic (PV) technology. Additives, which can improve the electronic properties of semiconductors, have been widely applied to the functional layers of PSCs including the absorption layers [9–11], electron transport layers [12–15] and hole transport layers [16–19]. The amount of additives added to PSCs has been usually optimized by a trial-anderror approach which is time-consuming and needs a significant amount of raw materials. Machine learning (ML), which is a branch of artificial intelligence (AI), enables computers to learn to perform a specific task. It has become a hot research area and has been applied to many areas due to the significant amount of data collected and tremendous development of computer hardware. Especially in material science, not just novel
Data Loading...
