Design and Discovery of Novel OLED Materials via Active Learning

Virtual high-throughput screening (vHTS) of materials has received significant attention as a critical tool for developing optoelectronic devices. However, exhaustive screening of an extensive materials library using first-principles methods could often be prohibitively expensive. With many potential molecules in the organic space, vHTS assisted by machine learning requires managing a high volume of data to properly assess the complexity of materials chemistry. In this regard, active learning has emerged as a promising strategy to efficiently explore an exhaustive library by prioritizing the decision-making process with unexplored data. Here, we present a robust and validated workflow that combines active learning with high-throughput physics-based simulations powered by first-principles calculations to identify promising candidates for organic light-emitting diodes (OLEDs). This workflow enables a systematic mechanism to account for multiple optoelectronic parameters while minimizing the number of computations required to explore an extensive library. Results of this work pave the way for efficient screening of materials for next-generation OLEDs and can be extended to other fields of organic electronic applications.