A Method for Optimizing PERC Cells in Industrial Production Lines Using Final IV Parameters, Statistical Procedures and Numerical Device Modeling

PERC Si solar cells have many processing parameters that interact with each other in a complex manner. Hence, efficiency changes can often not be allocated to a single fabrication tool. This makes the mass fabrication of PERC cells difficult to optimize and improve. On top of this, the wafers usually go through batch processes with parallel equipment, and are not traced. Thus, we developed an optimization method that uses solely the final IV parameters, treats these data statistically, and combines them with device modeling to identify sets of fabrication tools that most likely cause suboptimum cell efficiencies, and we identify specific reasons for the best cells, so the baseline efficiency can be increased. Concretely, an automatic algorithm identifies which departures from the median cell efficiency are statistically significant. These departures are caused by changes of different sets of IV parameters. The different sets are classified with a clustering algorithm, and each cluster is traced back to deviations in a group of fabrication tools by means of numerical device modeling tailored to the situation. We applied this method in a 2-week period to 552' 000 cells in a line being ramped-up. We found 10 clusters (i.e. different ways of departures from median efficiency), which are then exploited to improve PERC cell efficiency more swiftly.