Inline Characterization of Stacking Faults in EpiWafers

Manufacturing monocrystalline silicon wafers for solar cell applications by epitaxial growth on seed wafers has proven itself as a new and promising technology. Compared to the well-established process of ingot growth and wafering (based on the Czochralski method), it has the benefits of less energy consumption, less kerf loss, in-situ doping control and lower oxygen concentrations. Transferring such a technology into mass production requires a fast, appropriate, and robust inline quality control, which allows the significant defects to be monitored during production. One of the main defects in epitaxial wafers which has considerable influence on the electrical quality of the final solar cells are stacking faults. Their inspection with conventional techniques is time consuming using either scanner or microscope images. In this work, an inline approach for the detection of stacking faults and the classification of EpiWafers is presented based on optical inline inspection. Despite a mean absolute percentage error (MAPE) between the offline and the new inline method of 19%, spatial correlation is high and allows binning and sorting of wafers. The method has been integrated into pilot production and tested on more than 1500 EpiWafers.