Machine learning accelerated prediction of self-trapped excitons in double halide perovskites

Abstract

Broadband emission induced by self-trapped excitons (STEs) in double halideperovskites (DHPs) has received continuous attention in recent years. However,the comprehensive understanding of the STEs formation mechanism is still in itsearly stage. The corresponding roles of different B-site cations also remainunclear in these advanced materials. The lack of an effective STEs database forDHPs hinders the efficient discovery of potential optoelectronic materials withstrong STEs. Herein, a systematic STEs database is built for DHPs throughdensity functional theory (DFT) calculations and proposed a highly efficientpredictive machine learning (ML) model of the Huang–Rhys factorS for the firsttime. Results reveal the different contributions of two B-site metal cations to theformation of STEs in DHPs, which helps to understand the in-depth nature ofSTEs. Based on the accurate predictions of the effective phonon frequencyωLO ,it is further realized that the prediction ofS without conducting the time-consuming phonon property calculations of DHPs offers new opportunities forexploring the STEs. Combining DFT calculations and ML techniques, this studysupplies an effective approach to efficiently discover the potential noveloptoelectronic materials, which provides important guidance for the futureexploration of promising solid-state phosphors.