Challenges for the Quantification of Metal Induced Recombination Losses

One major loss mechanism for the currently highly relevant passivated emitter and rear cells (PERC) is locally enhanced recombination at the interface between semiconductor and front side metallization. For investigating these losses in detail, a reliable detection technique is crucial. A method we call calibrated photoluminescence imaging (PLI) method is a promising technique to extract the local dark saturation current density in the metallized area j0,loc. To investigate the sources of error of this method, metallized test samples are processed and the influence of sample specific parameters is considered in detail. Additionally, the difference in the resulting j0,loc between the calibrated PLI method and a simulative approach using numerical PLI simulations (Quokka3) is evaluated. We find that the results from the calibrated PLI method strongly depend on the accurate knowledge of the base resistivity B (including the impact of thermal donors) and less strongly on the reflectivity R of the illuminated side of the sample. In addition, metastable defects in the Si bulk can falsify the results, due to changes in the PLI intensity as a function of the illumination time. The difference in the resulting j0,loc between the calibrated PLI method and the simulative approach is within the error tolerances, which implies that the calibrated PLI method delivers accurate results despite the assumption of a uniform n throughout the sample. Here, it is important to mention that the non-uniformity of n is expected to be stronger for structures without a highly doped region at the investigated side (e.g. PERC rear contacts).