In this work, the synthesis of the mixed tin-lead MAPb0.75Sn0.25(I0.4Br0.6)3 perovskite is investigated. The presence of tin renders the understanding and control of the perovskite crystallization crucial. Spin-coating is the most common deposition technique employed to synthesize perovskite. This work aims to discriminate the process parameters, which are playing a key role in the properties of the perovskite layer.A statistical design of experiment (DoE) was used to avoid varying one factor at a time. This DoE allows an assessment of the process parameters influence and their possible interaction, and can predict further outputs and optimization.

Spin-coating process is broken down into two parts; first the perovskite ink deposition, then the perovskite nucleation and growth. Here, the speed and duration of the second step have been varied from 2000 to 8000 rpm and from 8 to 22s respectively. Simultaneously, anti-solvent washing has also been investigated; the delay of anti-solvent pouring has been varied from -3s before apparition of the first perovskite crystals to +3s after.

Crystallinity, morphology, composition and absorption of the perovskite layers were compared in the DoE. It was observed that, thickness aside, rotation speed of the substrate does not impact the final properties of the layer. No major change can either be imputed on a lengthening of the second step. DoE showed that anti-solvent washing determines all perovskite layer properties. Anti-solvent pouring after first perovskite crystals formation leads to incomplete conversion of intermediate into cubic perovskite, larger FWHM, and preferred grain growth along the {001} planes. Even though the elemental composition is identical, a slight band gap variation is observed for the substrates made from a late anti-solvent washing. This variation may come from perovskite intermediates still present in the perovskite layer. Thus, to synthesize a homogenous cubic perovskite with larger crystallite sizes, anti-solvent can be poured just before the vaporization of the last uncoordinated solvent. However, the crystallite will be less oriented along the same planes and the absorption will decrease slightly.

To investigate the incidence of film crystallinity on its opto-electrical properties, time resolved luminescence of perovskites exhibiting different crystallinities have been analyzed. Measurements show slower decay and more intense photoluminescence for fully converted films. Orientation of growth seems to take precedence over crystallites sizes and crystallinity. As a conclusion, to manufacture the optimal perovskite film, full perovskite conversion must be assured and attention should be paid to crystal orientation.