Lecture Martin Z. Bazant
E.G. Roos (1944) Professor of Chemical Engineering and Mathematics
Department of Chemical Engineering
Massachusetts Institute of Technology (MIR), Boston, MA
Tuesday, December 7, 2021
Lecture at 4pm
1610 Engineering Hall
Traditional methods of theoretical chemical engineering begin with human intelligence: Mathematical models encoding physical hypotheses are proposed, tested against experimental data and refined by fitting a few adjustable parameters. Recent advances in artificial intelligence appear to challenge this paradigm, since predictions can be made directly from data without the need for models, but such knowledge is often not transferrable to new situations. This talk will present a hybrid approach of solving PDE-constrained inverse problems to derive new models of electrochemical nonequilibrium thermodynamics, in the context of Li-ion batteries. Examples include inferring electro-autocatalytic reaction models from x-ray diffraction spectra for nickel-rich oxides, optical videos of lithium metal growth on graphite and scanning tunneling x-ray adsorption images of driven phase separation in lithium iron phosphate, as well as inversion of impedance spectra to infer microstructural heterogeneity and of acoustic emission spectra to reveal degradation processes during battery forming.