Machine learning approaches to analyze the effect of reaction parameters on ZIF-8 synthesis

Abstract

The morphology of zeolitic imidazolate framework-8 (ZIF-8) determines its effectiveness in applications like targeted drug delivery and energy storage. However, the precise control over morphology during synthesis is challenging since many reaction parameters affect it. Among these, precursor concentrations, solvents, and temperature are important parameters. As an integrative approach to experimental studies, this work develops machine learning (ML) models to predict the effects of these parameters on ZIF-8 morphology. Additionally, this work compares the performance of these models to demonstrate their potential as predictive tools for guiding the synthesis of ZIF-8 with controllable morphology.