Investigation of Recombination Mechanisms in Electronic Devices Using Bias-dependent Admittance Spectroscopy Applied to CIGS Solar Cells

The main cause for the power conversion efficiency limitations in Cu(In,Ga)(S,Se)₂ (CIGS) solar cells is still heavily debated in literature. Possible culprits for the limitation of the open circuit voltage of CIGS devices are conduction barriers, recombination in the bulk of the absorber, at grain boundaries, at the back contact or at the interface between the p-type absorber and the n-type buffer layer. In the present work we perform a large amount of bias-dependent admittance spectroscopy measurements on CIGS solar cells. We represent the data using CVf loss maps, comparing the measurement results to simulations, allowing us to draw conclusions about the recombination processes observed in the devices. Analyzing a range of devices consisting of state-of-the-art absorber layers with varying buffer layers and power conversion efficiencies, we could draw conclusions on the presence of an interface defect at the absorber-buffer interface. In fact, all devices, independent of power conversion efficiency, showed the presence of an admittance trace that could be related to a defect at the CIGS-buffer interface. A correlation could be found between the bias voltage position of the admittance trace with the open circuit voltage of the devices, indicating that the defect is limiting the photocurrent and open circuit voltage. A digital twin model involving only an interface defect at the CIGS-buffer interface was able to reproduce current voltage and admittance measurements of the best performing cell, proving the viability of the findings. We conclude that future improvements to the power conversion efficiency of these CIGS solar cells must come from interface engineering at the CIGS-buffer interface. Variations in doping of the absorber and buffer layer, the nature of the interface and buffer layer as well as the number of fixed charges at the interface all have the potential to drastically influence the significance and bias range of the interface recombination.