Organic-inorganic metal halide perovskites have attracted a considerable interest in the photovoltaic scientific community demonstrating a rapid and unprecedented increase in conversion efficiency in the last decade. Besides the stunning progress in performance, the understanding of the physical mechanisms and limitations that govern perovskite solar cells are far to be completely unravelled. In this work, we study the origin of their hysteretic behaviour from the standpoint of fundamental semiconductor physics by means of technology computer aided design (TCAD) electrical simulations. More in detail, we explain the hysteresis in current density-voltage (J-V) characteristic of perovskite solar cells (PSCs) in terms of charge carrier accumulation due to slow charge dynamics occurring at energy states near the interfaces with the perovskite absorber.

Our simulations reveal that hysteretic behaviour can be caused by shallow energy states near the interface with ETL (or HTL) with particular properties as: (i) density N_t > 10¹⁸ cm^-3, (ii) capture cross-section σ ≈ 10^‑23 cm², and (iii) defect energy E_t ≈ 0.25 eV with respect to VBE (or E_t ≈ 0.2 eV with respect to CBE). Interestingly, such values are consistent with experimental findings reported in literature.

Furthermore, stronger hysteresis at higher defect densities near the ETL or HTL reveals that the hysteretic behaviour is due to recombination processes. We provide an explanation of the observed phenomena in terms of slow charge capture and emission times at interface energy states with capture cross-sections lower than 10^-22 cm². Moreover, the effect is most apparent when the energy states exhibit energies close to the quasi-Fermi levels.

Finally, J-V simulations at different scan rates demonstrate realistic behaviour in the analysis of hysteresis for different scan rates. In particular, we observe that there is a scan rate that maximizes hysteresis as observed in experimental studies. Furthermore, our simulation platform allow us to relate the profile of energy states (energy and position) to interface passivation and crystallinity of absorber material in PCSs. In this regard, we observe that independently improving perovskite material or quenching interface defects is not sufficient to eliminate the hysteresis. Rather, only concurrently pursuing high quality perovskite material and surface passivation yields highest efficiency with negligible hysteresis. Our model can be easily extended and adapted to different device architectures, defect distributions and device life-time for a custom optimization that addresses different fabrication sequence, absorber material formulations, choice of supporting layers and stability.

In conclusion, our results show that hysteresis can be explained using fundamental semiconductor physics alone with realistic defect distributions in both the spatial and energetic domains. This supports the claim that interface defects play a crucial role in the formation of hysteresis. This fact is of particular relevance, because it links defects dynamics with hysteresis and stability of PSCs. Therefore, a deep study of such defects should lead to new insights for solving the reliability issue of PCSs, because defects can be created or disappear in particular conditions such as temperature, illumination and bias potential.