Identifying parasitic current pathways in CIGS solar cells by modelling dark J-V response

The non‐uniform presence of shunting defects is a significant cause of poor reproducibility across large‐area solar cells, or from batch‐to‐batch for small area cells, but the most commonly used value for shunt parameterisation (the shunt resistance) fails to identify the cause for shunting. Here, the use of equivalent circuit models to describe dark current–voltage characteristics of ZnO:Al/i‐ZnO/CdS/CIGS/Mo devices in order to understand shunting behaviour is evaluated. Simple models, with a single shunt pathway, were tested but failed to fit experimental data, whereas a more sophisticated model developed here, which includes three shunting pathways, yielded excellent agreement throughout the temperature range of 183–323 K. The temperature dependence of fitting parameters is consistent with known physical models. Activation energies and contact barriers are determined from the model, and extracted diode factors are unique across the voltage range. A case study is presented whereby the model is used to diagnose poor reproducibility for CIGS devices (efficiency ~3–14% across a 100 cm2 plate). It's shown that lower efficiencies correlated with greater prevalence of Ohmic and non‐Ohmic shunt currents, which may form due to pinholes in absorber and buffer layers respectively, whereas the quality of the main junction was constant for all cells (diode factor ~1.5–2). Electron microscopy confirmed the presence of ZnO:Al/i‐ZnO/Mo and ZnO:Al/CIGS/Mo regions, supporting the multi‐shunt pathway scheme disclosed by modelling. While the model is tested with CIGS cells here, this general model is a powerful diagnostic tool for process development for any type of thin‐film device. Copyright © 2015 John Wiley & Sons, Ltd. An equivalent circuit model, which allows for the presence of three types of shunting pathways, has been developed to describe the dark J–V characteristics in CIGS solar cells. Excellent agreement between the model and experimental data was apparent throughout a temperature range of 183–323K. Voltage‐independent ideality factors may be extracted, along with activation energies and barrier heights. The model may be used for rapid diagnosis of low, or inconsistent, device efficiencies for various thin‐film technologies.