Hole transport materials for QLEDs: a combined approach of machine learning and atomistic simulation

QLEDs have emerged as an alternative for optoelectronic applications. However, for widespread application of QLEDs, the device efficiency is required to be improved. There is a significant energy level mismatch between the valence band of commonly used quantum dots (QDs) and the HOMO level of traditional hole transport materials (HTMs). Given the small energy level mismatch between the conduction bands of the QDs and commercial electron transport materials, charge carriers in the light-emitting layer are imbalanced. Such a charge imbalance decreases the efficiency of QLED devices, and thus it is of great importance to design novel HTL materials with small energy mismatch with the QDs. Given the numerous potential molecules in the organic space, employing expensive and time-consuming approaches based on chemical intuition and trial-and-error experimentation is practically ineffective. Thus, realizing next-generation QLEDs technologies requires a paradigm change in materials design and development. Here, we combine active learning (AL) and high-throughput quantum mechanical calculations as a novel strategy to efficiently navigate the search space in a large materials library. The AL enables a systematic material screening by accounting multiple optoelectronic properties while minimizing the number of calculations. We further evaluated the top candidates using atomistic simulations and machine learning to investigate charge mobility and thermal stability in their amorphous films. This work offers guidelines for efficient computational screening of materials for QLEDs, reducing laborious, time-consuming, and expensive computer simulations, materials synthesis, and device fabrication.