(306a) Virtual Pareto-Front Mapping of Transition Metal-Catalyzed Gas-Liquid Reactions

Homogeneous catalytic reactions, facilitated by transition metal complexes, are pivotal in industrial processes and the specialty chemicals synthesis. Comprehensive mapping and optimization of homogeneous catalytic reactions present a complex challenge, traditionally approached through labor-intensive methods that often fail to unveil the full potential of candidate catalyst ligand systems.¹

In rhodium (Rh)-catalyzed hydroformylation of olefins, aiming to add a formyl group to a double bond leading to linear or branched aldehydes, the regioselectivity is paramount. Pareto-front Mapping of the reaction yield vs. regioselectivity offers critical insights into adjusting process parameters to favor the desired aldehyde which is essential for industrial applications to enhance product's efficiency, cost, and quality. Bayesian optimization (BO) integrated with an automated experimentation tool can effectively map the Pareto front of homogeneous catalytic reactions by intelligently navigating through the vast parameter space using a machine learning (ML) model as to predict the outcome of reactions.

This study, introduces a platform for autonomous reaction exploration³ which provided the experimental data needed to build an effective predictive ML² and uses a digital framework designed to rapidly explore and optimize transition metal-based homogeneous catalytic reactions, leveraging the combined power of autonomous experimentation and BO. Our investigation meticulously examines the influence of BO hyperparameters, such as the acquisition function and sampling size, on the efficiency and efficacy of mapping the reaction's Pareto front⁷. The findings emphasize the benefits of autonomous systems and ML in optimization and navigating catalytic processes, offering significant insights for chemical engineering and future innovations.

1. Y. Ibrahim, J. A. Bennett, D. Mason, J. Rodgers and M. Abolhasani, Journal of Catalysis, 2022, 409, 105-117.
2. Orouji, J. A. Bennett, M. Abolhasani and S. Sadeghi, Reaction Chemistry & Engineering, 2024.
3. Bennett, N. Orouji, M. Khan, S. Sadeghi, J. Rodgers and M. Abolhasani, Nature Chemical Engineering, 2024, 1-11.