Artificial Intelligence
Our laboratory is advancing the design of artificially intelligent (AI) micro-scale reactors for the study of multiphase physical and chemical rate processes. The design of such novel laboratory flow reactors has the potential to advance green chemistry, make chemical reactions safer, discovery new catalysts, minimize the building space and energy requirements, expedite information in weeks-to-months that would normally require years, and yield more accurate kinetic models and mechanisms. The approach has merit to revolutionize science and the engineering.
As an example, microreactors designed with in situ spectroscopic methods create the possibility to exploit artificial neural networks for catalytic polymerizations. It is estimated by the US Department of Energy that ~6% of all energy produced in the United States is used in the production of polymers. More importantly, ~37% of all global greenhouse gases (GHGs) are created directly or indirectly by polymer manufacturing. The design of AI microreactors could accelerate the discovery of metallocene catalysts for these highly complex and notoriously difficult to control olefin polymerizations, especially when little or no experimental kinetic data are available. One can explore the catalytic reaction space topology and the chemistry of co-catalysts to make decisions faster and with less energy and chemical waste generated. Unsupervised deep learning methods, in the other hand, can help to establish relationships or understudied combinations of monomer, co-catalyst, and catalysts using the literature and without the need for experiments.