Distributed series resistance in a one-dimensional two-diode model revisited

Abstract

The lumped series resistance Rs of large-area silicon solar cells, obtained from current–voltage (I–U) data according to the two-light-level method, varies along the I–U characteristic. Such a variation can most simply be described by the linear-response series resistance model (LR-Rs), recently developed in connection with luminescence imaging. Here, independently obtained experimental data are used to test the applicability of the LR-Rs model to Rs data based on I–U characteristics. After subtracting a non-distributed part from the measured Rs data, the inverse of the remaining distributed part shows a scaling proportional to the inverse of the bias-dependent diode resistance; a slope value of 1 is used as predicted by the LR-Rs model applied to a laterally one-dimensional geometry. The same experimental data have previously been interpreted based on a mathematically rather complicated model published already many years ago; just recently it was found that in some cases this model may lead to unphysical results. The present LR-Rs model based proper interpretation of the variation of the lumped series resistance along the I–U characteristic leads to a roughly half-by-half splitting between the distributed and the non-distributed part of Rs. This share has been observed many years for “economically reasonable” solar cells investigated by the CELLO technique. The successful usage of the LR-Rs model for I–U based Rs data is a strong hint that its underlying physical concepts are of general validity.