Accelerated Discovery of Targeted Environmentally Friendly A(II)B(I)X 3 -Type Three-Dimensional Hybrid Organic-Inorganic Perovskites for Potential Light Harvesting via Machine Learning.

The engineered hybrid organic-inorganic perovskites (HOIPs) with outstanding multifunctionalities have realized overarching targeted-driven applications and thus aroused intense research interest. The emergence of three-dimensional (3D) A(II)B(I)X-type HOIPs in 2018 brought a breakthrough to extend the 3D perovskite family and successfully realized prominent ferroelectricity at the same time. Here, we focus on these new-type HOIPs to perform machine-learning (ML)-based molecular design to screen promising candidates for versatile light harvesting, involving photovoltaics (77 ones), water splitting (216 ones), and photodetection (178 ones), out of 3180 A(II)B(I)X perovskites in total. These candidates await future experimental synthesis and characterization. Our high-throughput ML-based screening of 3D A(II)B(I)X HOIPs would enrich the material inventory by successfully introducing a class of new 3D HOIPs to realize property-oriented light harvesting and additional versatile energy harvesting due to their potential multifunctionalities such as ferroelectricity and electrocaloricity.