(219c) Artificial Intelligence with Microreactors for the Activity and Mechanism of a Zirconocene-Catalyzed Olefin Polymerization

The design of artificially intelligent micro-scale reactors has the potential to revolutionize catalysis. When combined with in situ spectroscopy, they can generate high-fidelity, transient information faster, safer, and with less environmental impacts compared to conventional methods. This seminar will highlight our recent findings that motivate the use of artificial neural networks (ANNs) for a supervised, yet more thorough investigation of catalytic reaction kinetics and mechanisms [1-3]. In the first example, the temperature control of microfluidics by computer-vision based on ANNs will be introduced. Their application can accelerate, in the second example, the discovery of metallocene catalysts for highly complex and notoriously difficult to control exothermic olefin polymerizations when limited experimental kinetic data are available. Finally, ANNs combined with experimental automation can also help to validate such complex polymerization mechanisms or to discover the optimal catalytic activity. The future of supervised machine learning in chemical reaction engineering is promising, offering new methods to better understand catalytic reactions.

[1] Rizkin, B.A., Shkolnik, A.S., Ferraro, N.J., Hartman, R.L. Nat. Mach. Intell. (2020) DOI: 10.1038/s42256-020-0166-5.

[2] Rizkin, B.A., Hartman, R.L. Chem. Eng. Sci., 210, 115224 (2019).

[3] Rizkin, B.A., Popovich, K., Hartman, R.L. Comput. Chem. Eng., 121(2), 584-593 (2019).