An Automatic Condition Detection Approach for Quality Assurance in Solar Cell Manufacturing Processes

Solar conversion efficiency is one of the most important quality metrics in solar cell production. During the production process, many operating factors might potentially influence the process conditions, thereby leading to unstable solar conversion efficiency of solar cell products. However, most solar cell fabrication plants focus more on inspections of the solar conversion efficiency after the growth of semiconductor epitaxy layers and electrodes because of a lack of an efficient strategy for online detection of process condition changes. This study develops an automatic approach for change condition detection in the multistage solar cell manufacturing process. By fully analyzing the geometrical features of the multichannel epitaxy data, process condition information can be obtained from temperature and reflectance profiles during the epitaxy layer growth. A likelihood ratio test is used to address the extracted features for in-advance change detection in solar conversion efficiency at the semiconductor epitaxy layer growth stage, thereby facilitating timely process adjustment and remedies. A real case from a multistage solar cell manufacturing process is used to validate the proposed method.