Perovskite solar cells’ power conversion efficiency has reached an impressive 23.7% [1] after vast improvements in material and fabrication techniques, but their widespread application is hampered by challenges in upscaling lab-sized cells into large modules. One enabling step in the upscaling process is laser scribing, in which a multilayer solar module is divided into series-connected cells by 3 scribe lines using a pulsed laser beam. So far, in active area and efficiency calculations it has been assumed that this laser beam does not damage the perovskite layer beyond the scribe lines themselves.

In this work, we aim to investigate the validity of this assumption by comparing modules scribed with a pulsed laser beam to that mechanically scribed with a knife. Specifically, we focus on the P3 scribe which removes the top metal contact from a module stack. We used focused ion beam milling to cut a lamella immediately adjacent to a P3 scribe. Then, we studied the samples with cross-sectional high-angle annular dark field (HAADF) imaging and energy-dispersive X-ray spectroscopy (EDX) in a scanning transmission electron microscope (STEM) to directly visualise laser damage on this slice of the module. We subsequently applied three multivariate statistical analysis algorithms to denoise and decompose our data sets, namely principal component analysis, independent component analysis, and non-negative matrix factorisation. These algorithms enabled acquisition of meaningful data with high signal-to-noise ratio while minimising beam current and dwell time. Plotting the brightness of our HAADF images showed lower intensity in the perovskite layer of laser scribed modules, indicating a loss of heavy elements such as lead or iodine. Meanwhile, an analysis of the EDX data revealed a much higher prevalence of PbI₂ flakes in the perovskite layer of laser scribed modules, suggesting laser-triggered decomposition of perovskite into PbI₂. Furthermore, we also investigated the effects of laser power, perovskite deposition method, and ablation depth on the extent of damage. Finally, we conducted top-view EDX scans in a scanning electron microscope to examine the integrity of the indium tin oxide contact, which must not be damaged by the P3 scribe. We anticipate that this work will help the perovskite photovoltaics community to further optimise the laser scribing process, simultaneously minimising the width of dead area and perovskite damage.