Lead halide perovskite solar cells afford high power conversion efficiencies, even though the photoactive layer is formed in a solution process. At the same time, solution processing may impose some severe dewetting issues, especially if organic, hydrophobic charge transport layers are considered. Ultimately, very narrow processing windows with a relatively large spread in device performance and a considerable lab-to-lab variation result. Here, we unambiguously identify dimethylsulfoxide (DMSO), which is commonly used as a co-solvent and complexing agent, to be the main reason for dewetting of the precursor solution on hydrophobic hole transport layers, such as polytriarylamine, in a gas-quenching-assisted deposition process. In striking contrast, we will show that N-methyl-2-pyrrolidon (NMP), which has a lower hydrophilic-lipophilic-balance, can be favorably used instead of DMSO to strongly mitigate these dewetting issues. The resulting high-quality perovskite layers are extremely tolerant with respect to the mixing ratio (NMP/dimethylformamide) and other process parameters. Thus, our findings afford an outstandingly robust, easy to use, and fail-safe deposition technique, yielding single (MAPbI3) and double (FA0.94Cs0.06PbI3) cation perovskite solar cells with high efficiencies (∼18.5%). Most notably, the statistical variation of the devices is significantly reduced, even if the deposition process is performed by different persons. We foresee that our results will further the reliable preparation of perovskite thin films and mitigate process-to-process variations that still hinder the prospects of upscaling perovskite solar technology.
Keywords: gas quenching; perovskite deposition; perovskite precursor wetting; perovskite solar cells; reproducibility.