Chem.Engg. R Kumar Distinguished Award Lecture Series. talk-2-Prof. Arthi Jayaraman

Linking Block Copolymer Design to Morphology and Thermal Conductivity using Machine Learning, Molecular Simulations, and Theory.

Alongside our experimental collaborators we are working to engineer polymer-based materials for high thermal conductivity applications. Tailoring polymers for desired properties in applications involves optimization over a large design parameter space including polymer chemistry, molar mass, sequence, and architecture, and linking each candidate to their assembled structures and in turn their properties. Having to do this optimization purely using experiments – synthesis and characterization – is time-consuming and expensive. To accelerate this process, there is a critical need for computational methods that can explore the design – structure – property space and identify a few promising candidates; this can be particularly challenging for polymers which exhibit multi-length scale structures assembled at long time scales, all of which together drive the properties. In our work, we are tackling this challenge for multiblock copolymers using a combination of self-consistent field theory, molecular dynamics simulations, and a new method called RAPSIDY – Rapid Analysis of Polymer Structure and Inverse Design strategY. [1-3] We demonstrate the above approach using an example of a symmetric, linear pentablock, AxByAzByAx, copolymer system for which we determine polymer sequences that exhibit stable a desired morphology and property (e.g., thermal conductivity). RAPSIDY reduces computational costs for design parameter exploration by up to two orders-of-magnitude compared to traditional MD methods, thus, accelerating design and engineering of novel polymer materials for target applications.