High-Performance Perovskite Quantum Dot Synthesis Investigated Through Exploratory Data Analysis

Abstract

Herein we introduce an exploratory data analysis (EDA) methodology to streamline the synthesis of metal halide perovskite quantum dots (PQDs) with high photoluminescence quantum yield (PLQY) and stability. Our approach integrates domain knowledge with a data-driven exploration of the chemical synthesis space. We assembled a targeted dataset and evaluated feature correlations, employing regression models and permutation importance to identify critical synthesis parameters. The OA/OLA ligand pair was pinpointed as a key factor, with the ligand ratio being finely tuned to 0.9, resulting in optimal PQD performance. This EDA-guided process, requiring minimal experimental resources, led to the discovery of significant synthesis parameters and their ideal values, enhancing PLQY notably. Our findings validate the efficacy of combining categorical and continuous features in PQD synthesis, underscoring the value of domain expertise in data preprocessing, feature selection, and model interpretation. This strategy elucidates the pathway to accelerate the development of high-performance PQDs, paving the way for their application in cutting-edge optoelectronic devices.