Large-Area In Situ Multichannel Imaging on Blade Coated Hybrid Perovskite Thin Films

Hybrid perovskite thin-film solar cells are considered for the next generation of photovoltaics. The reason is not only their fast increase in power conversion efficiency (PCE) from 3.5% in 2009 to over 25% to date, but also their high defect tolerance, their band-gap tunability and their compatibility with solution processing. However, the record PCEs are mostly demonstrated in laboratory-scale devices, while the process transfer to scalable solution processing methods such as slot-die coating remains a major challenge for commercializing the technology. The reason is that the involved drying and crystallization dynamics are poorly understood and most studies rely on characterization of readily crystallized perovskite films.<br/>In response, we developed a novel in situ imaging technique capable of recording the monochromatic reflectance and (spectral) photoluminescence (PL) response of forming perovskite thin-films with high spatial and temporal resolution, which we call in situ multichannel imaging (IMI). We designed IMI with the goal of surveilling the local perovskite morphology formation on large areas over 10×10 cm². In detail, IMI consists of a scientific CMOS camera, a high-power LED excitation at 470 nm and a 3D-printed filter wheel equipped with a neutral density (ND) filter, two long pass (LP) filters and one short pass (SP) filter, all with cut-offs near the PL emission wavelength of the investigated perovskite. The filter wheel is rotated at 180 rpm and the camera is triggered synchronously. The high resolution images are sorted according to the filter they were recorded through. From each four images per time step, we calculate images of three optoelectrical properties: First, the ND filter images are a representation of the reflectance of the film at the excitation wavelength. Second, adding up images of the SP and one LP filter yields the full photoluminescence intensity. Third, by dividing the two LP images by the SP image we estimate the PL emission wavelength based on a gaussian model of the PL spectrum.<br/>We perform two different methods of analyzing these extensive IMI image channels. On the one side, we investigate the transients of the signals over time by averaging over multiple spots on each substrate, comparing perovskite films that were processed at different environment parameters. As a result, we find that the reflectance channel yields critical information about the coalescence of the nucleating perovskite domains, the PL intensity channel shows significant differences (two orders of magnitude) caused by varying PL quantum efficiency and/or re-absorption effects due to distinct morphological features, while the PL emission wavelength channel probes the homogeneity and roughness of the film as well as the completeness of the perovskite formation. The PCEs of perovskite solar cells incorporating these perovskite thin-films correlate well with these observations. On the other side, instead of analyzing the signal transients, we analyze spatial inhomogeneities on only one sample at one instant during the formation of the perovskite. We find that the occurring irregularities can be classified into three categories depending on their correlative presence in the three image channels: Thickness/roughness variations, subsurface defects/semiconductor quality defects or pinholes/dust particles. The category of the defect determines the probability of a perovskite device to experience performance loss. However, both mentioned analysis methods critically reduce the information content of the present data. Therefore, we started working on more advanced data analysis methods to extract hidden patterns that might be present in the four channels image series.<br/>More detail on the described work can be found in the original, open access paper “Correlative In Situ Multichannel Imaging for Large-Area Monitoring of Morphology Formation in Solution-Processed Perovskite Layers” published in Solar RRL 2100353, September 15, 2021.