Continuous flow synthesis of pyridinium salts accelerated by multi-objective Bayesian optimization with active learning

Abstract

We report a human-in-the-loop implementation of the multi-objective experimental design via a Bayesian optimization platform (EDBO+) towards the optimization of butylpyridinium bromide synthesis under continuous flow conditions. The algorithm simultaneously optimized reaction yield and production rate (or space-time yield) and generated a well defined Pareto front. The versatility of EDBO+ was demonstrated by expanding the reaction space mid-campaign by increasing the upper temperature limit. Incorporation of continuous flow techniques enabled improved control over reaction parameters compared to common batch chemistry processes, while providing a route towards future automated syntheses and improved scalability. To that end, we applied the open-source Python module, nmrglue, for semi-automated nuclear magnetic resonance (NMR) spectroscopy analysis, and compared the acquired outputs against those obtained through manual processing methods from spectra collected on both low-field (60 MHz) and high-field (400 MHz) NMR spectrometers. The EDBO+ based model was retrained with these four different datasets and the resulting Pareto front predictions provided insight into the effect of data analysis on model predictions. Finally, quaternization of poly(4-vinylpyridine) with bromobutane illustrated the extension of continuous flow chemistry to synthesize functional materials.

Scheme 1. Continuous flow synthesis setup and EDBO+ workflow. Initial seed reactions were conducted within the predefined input constraints. Subsequent rounds of experiments were performed in batch sizes of three unique reactions. The outputs were used to update EDBO+ and provide the next round of suggested experiments. Initially 10 rounds of experiments were perform followed by expansion of the upper temperature constraint to 168 °C and another 5 rounds.

Fig. 1. Monitoring metrics for the initial EDBO+ reaction optimization campaign. (A) The Pareto front solution of the multi-objective optimization (blue) and dominated solutions (grey). (B) Expansion of the hypervolume of all solutions to the Pareto front. (C) Maximum EI in production rate. (D) Maximum EI in reaction yield. Note that the EI for each round contains data from all previous experiments.

Fig. 2. Monitoring metrics for the expanded EDBO+ reaction optimization campaign. (A) The Pareto front solution of the multi-objective optimization (red) and dominated solutions (grey). (B) Expansion of the hypervolume of all solutions to the Pareto front. (C) Maximum EI in production rate. (D) Maximum EI in reaction yield. The EI for each round contains data from all previous experiments. Data from the initial and expanded reaction campaigns are shown in blue and red, respectively.

Fig. 3. Predicted Pareto fronts from low- and high-field NMR analysis outputs of manually and semi-automated processed data.

Fig. 4. Characterization of the polymer product synthesized under continuous flow. (A) ¹H NMR spectra of P4VP (top, red) and f-P4VP-1 (bottom, blue) in DMSO-d₆. (B). XPS N 1s spectrum of P4VP. (C). XPS N 1s spectrum of f-P4VP-1. Analysis reveals two distinct N species at 398.71 eV and 401.64 eV, corresponding to free pyridine and quaternized pyridinium on the polymer, respectively. Samples were isolated from solutions in DMAc prior to NMR and XPS analysis.