Partial shading degradation mapping in monolithically integrated perovskite modules

Perovskite-based PV devices are increasingly attractive, both as a single junction technology and in tandem, combined with lower bandgap thin films or c-Si. However, they are still lacking maturity at the industrial scale, especially when it comes to reliability, such as partial shading tolerance. This particular hazard is critical for monolithically integrated devices, where bypass diodes cannot easily be implemented and the impact of the induced reverse bias on performance can be severe [1] and must therefore be studied in depth. This topic has been investigated to some extent, but most studies focus on individual cells [2-4]. These studies, however, might not extrapolate well to actual modules, where several variations might be occurring at any given location, such as variations in coating thicknesses, composition or laser scribe quality. To fill this gap, the present investigation focuses on a single junction perovskite module consisting of 30 monolithically interconnected cells (in series), where partial shading is applied in a controlled manner and the resulting degradation is characterized and mapped, to assess spatial homogeneity. The controlled degradation studied here was induced by applying a partial shading experiment on three adjacent cells: the cells were shaded for two minutes while the module was kept at short circuit. To study them, we developed a characterization protocol that would allow us to better understand the degradation. This sequence needed to be easy to apply and have a sufficiently high resolution. The method developed at that stage combines imaging photoluminescence (PL), both at ISC, which images charge extraction, and VOC, showing regions of high radiative recombination, with imaging electroluminescence (EL), where both current injection and recombination affect the luminescence map, and illuminated lock-in thermography (ILIT). The partial shading induced a visible degradation in PL (both at ISC and Voc) and EL, concentrated at one end of the cells. ILIT also shows local heating in this region. This non-uniformity in degradation is likely related to a locally lower breakdown voltage on this part of the module. This ties up with tests showing that the breakdown voltage can vary significantly across the module surface, for reasons that are still being investigated. To investigate the mechanism of the degradation, we combine the analysis of the three luminescence maps generated. We find that 1. The charge carriers are poorly extracted from the bottom area of the shaded cells (brighter PL region at ISC), 2. There is reduced recombination at the perovskite/charge transport layer interface (bright PL region at VOC) in this same area and 3. In that same area, most of the charge carriers injected in the electrodes do not reach the perovskite layer (darker EL region). Those three observation point to an energy barrier at one of the perovskite/charge transport layer interface. These results show that non-obvious heterogeneities (e.g. in breakdown voltage) can be present in modules. We find that precious information on reverse bias degradation can be gained by studying perovskite modules and combining luminescence and thermography imaging techniques to map the degradation. This work paves the way towards scaling up the perovskite technology in a more reliable fashion.