Optimize the quantum yield of G-quartet-based circularly polarized luminescence materials via active learning strategy-BgoFace

Abstract

G-quartet (G4)-based circularly polarized luminescence (CPL) materials within CPL engineering have attracted substantial attention in optoelectronics and photonics owing to their excellent chiral properties and promising applications in advanced optical devices. However, their practical use is limited by relatively low quantum yield (QY), which reduces emission efficiency. Addressing this challenge, we present BgoFace, an integrated active learning (AL)-based software, to optimize G4-based CPL materials with high QY. Starting with an initial dataset of 54 experimentally validated samples, the system executed six AL cycles encompassing 24 targeted experimental groups. Through this closed-loop workflow, BgoFace successfully identified G4 complexes exhibiting a near doubling of QY (37.25%). This achievement significantly advances the previously low QY values typically reported for G4-based CPL materials. The optimized materials demonstrate enhanced stability and processability, attributable to the AL algorithm's simultaneous consideration of multiple physicochemical parameters. This study not only advances the field of G4-based CPL materials for optical and photonic applications, but also establishes a generalizable AL framework suitable for optimizing functional nanomaterials in optoelectronic device design. By bridging data-driven design and experimental validation, BgoFace offers a transformative strategy for accelerating the development of functional nanomaterial engineering.