Resolving the Effects of the Mixing Process on Hierarchical Structures in Polymer Nanocomposites by Bayesian Ultra-Small-Angle X-ray Scattering Analysis

Abstract

Bayesian inference was combined with Unified Fit to resolve the effects of the mixing history on polymer nanocomposites whose ultra-small-angle X-ray scattering profiles are nearly indistinguishable by conventional analysis. Bayesian Unified Fit was applied to comparing the aggregate size, internal mass-fractal structure, and surface roughness for three carbon black--filled ethylene propylene diene copolymer rubber samples prepared by different mixing processes. Whereas conventional Unified Fit reported essentially identical carbon-black dispersions, the radius of gyration, mass fractal dimension, and surface fractal dimension for the three samples all followed the same ordering, which coincides with the reported ordering of the tensile strength and elongation at break and suggests that the inferred aggregate hierarchy is physically meaningful for the macroscopic mechanical response. These results identify the aggregate size as a candidate structural descriptor of rubber reinforcement and provide an uncertainty-aware basis for examining processing--structure--property relationships in polymer nanocomposites.