The rapid expansion of the Internet of Things (IoT) is driving demand for sustainable, low-power energy solutions for distributed electronic devices, which are currently mostly powered by batteries. Indoor photovoltaics (IPVs) represent a promising alternative, enabling self-powered systems by harvesting energy from ambient indoor illumination. [1] Halide perovskites have emerged as leading candidates for IPVs due to their high absorption coefficients and tunable bandgaps.

Indoor operating conditions differ significantly from standard 1-Sun conditions, with lower photon flux and narrower light spectra. This reduces photogenerated carrier densities and increases the relative influence of recombination, leakage currents, and interfacial processes on device performance. [1] These factors strongly influence hysteresis, a phenomenon commonly attributed to ion migration, interfacial charge accumulation, and trapping. From a physical perspective, hysteresis arises from the coupling between slow ionic redistribution and fast electronic transport, resulting in time-dependent internal electric fields. In addition, hysteresis is strongly affected by measurement conditions, such as scan rate. Under solar illumination, hysteresis is expected to diminish with decreasing light intensity, as the ion-migration barrier becomes more difficult to overcome, leading to reduced ionic movement. [2,3] Our previous work showed the opposite trend under indoor conditions: hysteresis increased as illumination decreases from 1000 to 50 lx [4], challenging the widespread assumption that device behavior under 1-Sun conditions can be directly extrapolated to indoor operation.

To further investigate this phenomenon, we conducted a comprehensive study using four perovskite-based n–i–p devices incorporating either TiO₂ or SnO₂ as the electron transport layer and doped or undoped Spiro-OMeTAD as the hole transport layer. Hysteresis behavior was probed through current–voltage measurements at three illumination levels (50, 200, and 1000 lx) using varying scan rates. Complementary device simulations were performed to evaluate hysteresis as a function of scan rate and light intensity, providing insight into the underlying mechanisms. To elucidate the observations, hysteresis was decomposed into contributions from short-circuit current, open-circuit voltage, and fill factor. This approach reveals distinct hysteresis mechanisms that depend on both device architecture and operating conditions, showing that IPV hysteresis cannot be defined by a single value. These results challenge conventional hysteresis assessment approaches and highlight the need for indoor-specific characterization protocols for the reliable evaluation of IPV devices.