Utilizing real and statistically reconstructed microstructures for the viscoelastic modeling of polymer nanocomposites

In this paper, we present a new approach to finite element modeling of a nanoparticle filled polymer system that utilizes the actual and statistically reconstructed microstructures of the material. Typically, description of polymer nanocomposites for microstructure generation is difficult given the high degrees of freedom inherent in the location of each nanoparticle. The lack of true microstructure utilization hinders our ability to understand the interaction between the nanoparticle and polymer, which cannot easily be deconvoluted from experiments alone. We consider here a material system of carbon black particle fillers dispersed in synthetic natural rubber. Scanning Electron Microscope (SEM) images are first taken of these carbon black-rubber composites samples and then transformed into binary images. The binary images from either a microscope image of original specimens or microstructure reconstruction according to the material statistical description are used as geometric inputs for the finite element model along with experimentally determined viscoelastic properties of pure rubber. Simulations on the viscoelastic properties of the rubber composites are performed through ABAQUS. The simulated results are then compared with composite viscoelastic data in both frequency and temperature domains. The comparison shows that for the specific rubber/CB composite discussed in this paper, the thickness being 25nm and relaxation time being 32 times that of matrix polymer provide the best approximations for the properties of interfacial polymer.