Hybrid perovskite photovoltaics is a very promising, emerging thin-film technology that has not only rapidly increased in power conversion efficiencies of perovskite solar cells featuring a large range of bandgaps, but also promises low-cost industrial-scale fabrication due to its compatibility with solution processing and the abundancy of the precursor materials. However, the fabrication of perovskite modules on the industrial scale still faces major challenges. Besides the material’s toxicity due to lead incorporation and its instability with respect to stress factors like humidity, temperature and light (which could be addressed in the future by employing suitable encapsulation and product cycle strategies), the scalability and reproducibility of large-scale solution processing does not yet meet industrial standards. For one thing, morphological defects are very likely to occur in large-area deposition techniques such as slot-die coating, spray coating and inkjet printing. For another thing, the opto-electronic functionality of solution processed perovskite thin-films varies from batch to batch due to their sensitivity to the processing parameters such as temperature, humidity, lighting, etc.

In response to these issues, we propose a combined toolkit of modeling and monitoring. First, we in situ characterize the perovskite formation process starting from the precursor solution thin-films with high spatial and temporal resolution, that is, we evaluate multidimensional series of reflectance and (spectral) photoluminescence signals. Second, we correlate these properties with sophisticated models of the perovskite thin-film drying and crystallization. That is to say, we vary the processing parameters over a wide range and evaluate the respective response signals from the in situ characterization techniques in reference to the predictions of our models. In this way, we show that the drying, as measured by reflectance oscillations, is mainly controlled by the temperature and the local mass transfer coefficient (as well as the choice of solvents). Likewise, we demonstrate that the crystallization dynamics, as correlated with the photoluminescence response, depend on the drying rate and temperature (as well as the precursor materials). To sum up, we manage to understand and surveil the morphology formation in perovskite thin-films both on the lateral and on the temporal scale. This methodology could enable direct process control in the future, meaning that both spatial and temporal non-homogeneities in solution processing can be directly identified and corrected instantly by feedback controlling the processing parameters.