This study evaluates the noise resilience of multi-objective Bayesian optimization (MOBO) algorithms in chemical synthesis, an aspect critical for processes like telescoped reactions and heterogeneous catalysis but seldom systematically assessed. Through simulation experiments on amidation, acylation, and SNAr reactions under varying noise levels, we identify the qNEHVI acquisition function as notably proficient in handling noise. Subsequently, qNEHVI is employed to optimize a two-step heterogeneous catalysis for the continuous-flow synthesis of hexafluoroisopropanol. Achieving considerable optimization within just 20 experimental runs, we report an E-factor of 0.3744 and a conversion rate of 76.20%, with optimal conditions set at 5.00 sccm and 35.00℃ for the first step, and 80.00 sccm and 170℃ for the second. This research highlights qNEHVI’s potential in noisy multi-objective optimization and its practical utility in refining complex synthesis processes.

This paper describes a machine learning guided framework for screening the potential toxicity impact of amine chemistries used in the synthesis of hybrid organic-inorganic perovskites. Using a combination of a probabilistic molecular fingerprint technique that encodes bond connectivity (MinHash) coupled to non-linear data dimensionality reduction methods (UMAP), we develop an “Amine Atlas’. We show how the Amine Atlas can be used to rapidly screen the relative toxicity levels of amine molecules used in the synthesis of 2D and 3D perovskites and help identify safer alternatives. Our work also serves as a framework for rapidly identifying molecular similarity guided, structure-function relationships for safer materials chemistries that also incorporate sustainability/ toxicity concerns.