Molecular modeling, simulation, and machine learning of polymer nanocomposites containing nanorod fillers

Polymer nanocomposites (PNCs) with nanorods as fillers exhibit morphologies (e.g., percolation of nanorods) that are not easily accessible for spherical fillers, making them more desirable for applications requiring conductive and transport properties. Past studies have established the phase diagram of PNC morphologies obtained with homogeneous nanorods; in contrast studies on PNCs with nanorods with physically or chemically heterogeneous functionalization are lacking. In this poster we will present our recent work with coarse-grained molecular dynamics (MD) simulations to predict PNC morphologies with chemically patchy¹ and physically rough² nanorods. In this poster we will also present a machine learning (ML) approach for automatically classifying nanowire/nanorod morphologies (e.g., dispersed/aggregated/percolated nanowires) from transmission electron microscopy (TEM) images DOI: 10.1039/D2DD00066K. Our ML approach uses a small number of labeled microscopy images for training and performs as effectively as methods trained on significantly larger image datasets.



(1) Lu, S.; Wu, Z.; Jayaraman, A. Molecular Modeling and Simulation of Polymer Nanocomposites with Nanorod Fillers. The Journal of Physical Chemistry B 2021, 125 (9), 2435-2449.

(2) Lu, S.; Jayaraman, A. Effect of Nanorod Physical Roughness on the Aggregation and Percolation of Nanorods in Polymer Nanocomposites. ACS Macro Lett 2021, 10 (11), 1416-1422. DOI: 10.1021/acsmacrolett.1c00503.