How measurement protocols influence the dynamic J-V characteristics of perovskite solar cells: Theory and experiment

• Published in: Solar energy 2018-10, Vol.173, p.976-983
• Main Authors: Nemnes, George Alexandru, Besleaga, Cristina, Tomulescu, Andrei Gabriel, Palici, Alexandra, Pintilie, Lucian, Manolescu, Andrei, Pintilie, Ioana
• Format: Article
• Language: English
• Subjects: Bias; Computer simulation; Deoxidizing; Energy conversion efficiency; Experimental data; Hysteresis; Hysteresis index; Hysteresis type; Measurement; Measurement protocol; Perovskite; Perovskite solar cell; Perovskites; Photovoltaic cells; Preconditioning; Reproducibility; Solar cells; Solar energy
• ISSN: 0038-092X; 1471-1257
• DOI: 10.1016/j.solener.2018.08.033

•The type of the hysteresis depends on both pre-poling voltage and scan procedure.•A new hysteresis index properly accounts for the hysteresis magnitude and type.•Correlated forward and reverse bias scans consistently describe the hysteresis.•A set of guidelines for a proper characterization of the hysteresis is discussed. The dynamic effects observed in the J-V measurements represent one important hallmark in the behavior of the perovskite solar cells. Proper measurement protocols (MPs) should be employed for the experimental data reproducibility, in particular for a reliable evaluation of the power conversion efficiency (PCE), as well as for a meaningful characterization of the type and magnitude of the hysteresis. We discuss here several MPs by comparing the experimental J-V characteristics with simulated ones using the dynamic electrical model (DEM). Pre-poling conditions and bias scan rate can have a dramatic influence not only on the apparent solar cell performance, but also on the hysteretic phenomena. Under certain measurement conditions, a hysteresis-free behavior with relatively high PCEs may be observed, although the J-V characteristics may be far away from the stationary case. Furthermore, forward-reverse and reverse-forward bias scans show qualitatively different behaviors regarding the type of the hysteresis, normal and inverted, depending on the bias pre-poling. We emphasize here that correlated double-scans, forward-reverse or reverse-forward, where the second scan is conducted in the opposite sweep direction and begins immediately after the first scan is complete, are essential for a correct assessment of the dynamic hysteresis. In this context, we define a hysteresis index which consistently assigns the hysteresis type and magnitude. Our DEM simulations, supported by experimental data, provide further guidance for an efficient and accurate determination of the stationary J-V characteristics, showing that the type and magnitude of the dynamic hysteresis may be affected by unintentional pre-conditioning in typical experiments.