Knowledge-infused process monitoring for quality improvement in solar cell manufacturing processes

Abstract

Solar conversion efficiency (SCE), an important quality metric in solar cell manufacturing processes, is related to chemical vapor deposition in the epitaxy stage based on the photoelectric effect. A large number of solar cell fabrication plants still lack online process monitoring strategies at the epitaxy stage and instead use offline inspections after fabrication is completed. Consequently, production efficiency is reduced due to offline inspections and the quality of wafers in downstream manufacturing stages is uncertain because only a small portion of wafers can be inspected due to random sampling within a single batch. A knowledge-infused monitoring strategy in the epitaxy stage of solar cell manufacturing processes that enables the direct link of online process monitoring to quality SCE is proposed in this study. A customized nonlinear model based on light interference with parameters that can be physically interpreted and largely accepted by practitioners is proposed to capture key information of reflectance signals. Multiple process features are extracted and a general correlation-based variable ranking procedure is adopted in this nonlinear model to rank SCE-correlated key process features. This model enables online process monitoring of key features at the epitaxy stage and allows practitioners to apply timely remedies in case of unexpected conditions. The proposed knowledge-infused process monitoring approach fully considers the physical knowledge from light interference and interpretability of parameters in the established nonlinear model correlated with the quality metric SCE to facilitate the online process monitoring at the epitaxy stage. A real solar cell manufacturing case shows the effectiveness of the proposed monitoring strategy.

Keywords:

• solar cell manufacturing
• epitaxy stage
• feature extraction
• solar conversion efficiency
• nonlinear regression

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was funded by National Natural Science Foundation of China under grants No. 71932006, 72001139 and 71771004.

Notes on contributors

Juan Du

Juan Du is an assistant professor in the Smart Manufacturing Thrust, Systems Hub, The Hong Kong University of Science and Technology, Guangzhou, China. She is also an affiliate assistant professor in the Department of Mechanical and Aerospace Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong. Email address: [email protected].

Xi Zhang

Xi Zhang is an associate professor in the Department of Industrial Engineering & Management, Peking University, Beijing, China. Email address: [email protected].

Wei Ou

Wei Ou is a senior engineer in the 18th Research Institute of China Electronics Technology Group Corporation, Tianjin, China. Email address: [email protected].