Designing kinetics: from modeling to optimization and back again

In the past decade, the design of new materials and molecules with tailor-made properties has made huge strides with the help of molecular simulation. Major advances have been made by integrating in- and out-of-equilibrium simulation techniques and machine learning into inverse-design pipelines for the development of novel molecules, biomolecular materials, and soft matter systems. Nonetheless, these design approaches are usually  focused on static, thermodynamic and structural properties. In contrast, designing for kinetic or other time-dependent properties has been explored to a much lesser extent, even though it is clear that designing dynamical properties is useful for many different purposes, including optimizing chemical reactions, catalysis, ligand-binding kinetics, and self-assembling materials. 

The absence of kinetics in commonly used design protocols is most likely due to the enormous costs for computing a single prediction using standard molecular simulation methods. A case in point is the computation of timescales of reactive processes in complex molecular systems, which are often determined by the rate of crossing high free energy barriers. While “rare event” methods have been developed to investigate mechanisms and kinetic properties that could not otherwise be studied (e.g. Markov state modeling, Transition Path Sampling, Path Reweighting ), such techniques are only now beginning to be applied for the purpose of inverse design. Combined with the integration of machine learning and other related techniques with dynamical calculations, we sit at the precipice of a new era of inverse kinetic design.

In this workshop we will bring together theoretical, computational, and experimental scientists from several diverse communities, who are interested in designing kinetics in a variety of contexts such as: Path Ensemble methods, Markov state models, Differentiable Molecular Dynamics, Landscape engineering, Optimization methods, and Machine learning approaches. Important application areas include Equilibrium and nonequilibrium self-assembly, Design of nucleation pathways in soft matter systems, Colloidal catalysis,  DNA nanotechnology,  Vitrimers, Inverse design of chemical reaction networks,  and Protocol design for active matter.