Improved lab-to-fab solar cell performance assessment by statistical data analysis in an automated, high-throughput metrology line

The development of new solar cell technologies and processes requires a reliable electrical and electro-optical performance characterization. In most cases, laboratory measurements on a smaller number of cells serve as a starting point for research facilities. In our work, we present results showing that a statistical data analysis in the pre-production phase can lead to significant additional information relevant for a full-scale production implementation. Thus, the assessment of cell process steps is improved as more robust and statistically sound conclusions can be drawn. As a first case study, we analyze the contacting process during the current-voltage-measurement. We present a data analysis approach that relies on the investigation of shape parameters of the distribution of electrical performance parameters. We show that the skewness of the fill-factor distribution is a much more sensitive parameter regarding an unreliable measurement setup than the mean values of efficiency of fill-factor themselves. As a second case study, we consider two types of laser processes for labeling solar cells. Here, our analysis of a larger number of cells appears to be much more conclusive than a single cell consideration. This is related to the small changes in the electrical parameters imposed by the laser processing. Based on these results, novel strategies for improving metrology techniques, that are based on well-assorted reference solar cell batches and associated data sets, will be discussed.