Data-Driven Controlled Synthesis of Oriented Quasi-Spherical CsPbBr Perovskite Materials.

Controlled synthesis of lead-halide perovskite crystals is challenging yet attractive because of the pivotal role played by the crystal structure and growth conditions in regulating their properties. This study introduces data-driven strategies for the controlled synthesis of oriented quasi-spherical CsPbBr, alongside an investigation into the synthesis mechanism. High-throughput rapid characterization of absorption spectra and color under ultraviolet illumination was conducted using 23 possible ligands for the synthesis of CsPbBr crystals. The links between the absorption spectra slope (difference in the absorbance at 400 nm and 450 nm divided by a wavelength interval of 50 nm) and crystal size were determined through statistical analysis of more than 100 related publications. Big data analysis and machine learning were employed to investigate a total of 688 absorption spectra and 652 color values, revealing correlations between synthesis parameters and properties. Ex situ characterization confirmed successful synthesis of oriented quasi-spherical CsPbBr perovskites using polyvinylpyrrolidone and Acacia. Density functional theory calculations highlighted strong adsorption of Acacia on the (110) facet of CsPbBr. Optical properties of the oriented quasi-spherical perovskites prepared with these data-driven strategies were significantly improved. This study demonstrates that data-driven controlled synthesis facilitates morphology-controlled perovskites with excellent optical properties.