Numerical Simulations of Photoluminescence for the Precise Determination of Emitter Contact Recombination Parameters

Abstract

One major loss mechanism for currently relevant solar cells [e.g., passivated emitter and rear cells (PERC)] is locally enhanced recombination at the interface between semiconductor and metalization. For investigating these losses in detail, a reliable detection technique is crucial. The photovoltaics community frequently applies an area-weighted model to extract the local dark saturation current density in the metalized area of the emitter j <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0,met</sub> -e.g., from photoluminescence imaging (PLI) data. However, this model does not account for the nonuniformity of the excess carrier density Δn within the sample during the measurement. Therefore, we compare numerical PLI simulations using Quokka3 to PLI measurements, to quantitatively reveal the impact of the nonuniformity of Δn. Test structures with locally laser-ablated passivation on one side-serving as ideal test structures-are used for the experimental verification. Additionally, we show results on metalized samples. We find that the results (j <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0,met*</sub> ) using the simulative approach for the laser-ablated samples exceed the results using the conventional area-weighted approach by more than 20%. For the metalized samples, we see a similar trend where the area-weighted approach again underestimates the results (j <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0,met</sub> ), in this case by up to 20%. Based on our investigations, we show that the deviation between the results arises due to the assumption of uniform Δn applied by the area-weighted approach.