Bayesian Optimized Crystallization of a Hydroxamate‐Functionalized Covalent Organic Framework for Enhanced Uranyl Uptake

Abstract To address the synthetic challenge of covalent organic frameworks (COFs), especially those with interfering functional groups, a Bayesian optimization (BO) centered approach is developed and implemented. Specifically, the crystallinity index for a well‐known TAPB‐PDA COF is improved by ≈80% via a one‐round proof‐of‐concept BO. For a more complicated task toward the preparation of hydroxamate‐functionalized TpPa COF, where improvement of both crystallinity and selectivity (against a crystalline byproduct) is needed, an efficient protocol comprising 6 BO iterations (with 5 experiments each) from an initial 64‐experiment dataset is successfully developed. The functional COF, namely SUM‐99 (SUM = Sichuan University Materials), with enhanced crystallinity, is subsequently demonstrated to be an effective, reversible, and selective sorbent for aquatic uranyl uptake. The importance of improved crystallinity, reflecting the power of BO, is showcased by a 23.7% increase in uranyl adsorption capacity. Therefore, the BO protocol and toolkit is presented for the efficient evolution of COF synthetic conditions, toward higher crystallinity and enhanced performances for downstream applications