AI Accelerates Search for Materials for Bright and Efficient Blue OLEDs

Scientists from Nagoya University and Kyushu University (Institute of Transformative Bio-Molecules, Institute for Advanced Study) applied machine learning combined with quantum chemical calculations to search for new materials for blue OLEDs. They created a virtual library of over 19,000 molecules based on carbon, hydrogen, and nitrogen, from which 1,000 were randomly selected for precise calculations. An AI was trained on this data and evaluated the remaining 17,000 molecules. Then, the top 50 candidates were recalculated using quantum chemistry, and two promising compounds — Cz-PAH-1 and Cz-PAH-2 — were selected for synthesis. They exhibited narrow-band blue emission with a spectrum width of 17–19 nm, photoluminescence quantum yield up to 99%, and external quantum efficiency up to 35.2%. Importantly, the materials contain no boron, simplifying synthesis. The developed material discovery cycle can be applied for accelerated discovery of other organic materials for batteries, catalysts, solar panels, and so on.