Molecular Simulation Approach to the Prediction of Mechanical Properties of Silica Reinforced Rubbers

Thursday, October 13, 2016: 8:30 AM
Rm 306-7 (David L. Lawrence Convention Center )
Jan Meyer1, Jonathan Hager1, Reinhard Hentschke1, Nils Walter Hojdis2 and Hossein Ali Karimi Varzaneh2, (1)Physics, Wuppertal University, Wuppertal, Germany, (2)Material Development, Continental Reifen Deutschland GmbH, Hannover, Germany
Our goal is the calculation of macroscopic viscoelastic properties of filled rubber, in the context of the Payne-effect, using molecular dynamics simulation on the atomistic scale.

Filled rubbers acquire most of their mechanical strength through fillers forming spanning branched networks throughout the rubber matrix. Here we focus on one “breakable” aggregate-to-aggregate contact [1,2] within a network branch and its contribution to dissipative loss. We present a fully atomistic simulation study of a filler-to-filler contact embedded in polymer matrix. The modeled system consists of spherical silica particles embedded in a polyisoprene-matrix. Additional options are silanes attached to the particle surfaces, cross-links between the polymer itself and cross-links between polymer with silica-particle via silanes. The utilized rubber force field is taken from [3]. The polymer molecular weight in our work ranges from 272 to 54400 g/mol. The silica particles are cut from the β-cristobalite crystal structure. Cut valences are saturated with hydroxyl groups. The final systems consists of about 10⁵ atoms. For the simulations we use the program package Gromacs 5.0.5.

First we validate key properties of the pure polymer, including density, diffusion coefficient and characteristic ratio. Furthermore we investigate the dynamical behavior of the polymer in the vicinity of the particles. We also study the dynamical response of the polymer to a displacement of a single silica, from which we obtain the polymer shear modules in reasonable agreement with the experiment. We calculate force-vs-particle separation curves for cyclic loading, confirming a previously suggested loss mechanism contributing to the Payne-effect. Finally we discuss possible predictions obtainable from this model in relation to experimental tan-delta measurements.

[1] R. Hentschke et al., J. Appl. Pol. Sci 2014, 131(18), 40806

[2] R. Hentschke et al., Macromolecules 2015, 48(24), 9039

[3] V. A. Harmandaris and M. Doxastakis, J. Chem. Phys. 2002, 116(1), 436