Prediction of Thermal Conductivities of Rubbers by MD Simulations-New Insights

In this article, two main approaches to the prediction of thermal conductivities by molecular dynamics (MD) simulations are discussed, namely non-equilibrium molecular dynamics simulations (NEMD) and the application of the Green-Kubo formula, i.e., EMD. NEMD methods are more affected by size effects than EMD methods. The thermal conductivities of silicone rubbers in special were found as a function of the degree of crosslinking. Moreover, the thermal conductivities of thermoplastic polyurethane as function of the mass fraction of soft segments were obtained by those MD simulations. All results are in good agreement with data from the experimental literature. After the analysis of normalized heat flux autocorrelation functions, it has been revealed that heat in the polymers is mainly transferred by low-frequency phonons. Simulation details as well as advantages and disadvantages of the single methods are discussed in the article.

Automated summary

This article discusses two main approaches to predicting thermal conductivities through molecular dynamics simulations: non-equilibrium molecular dynamics (NEMD) and the application of the Green-Kubo formula, i.e., EMD. NEMD methods are more affected by size effects compared to EMD methods. The thermal conductivities of silicone rubbers and thermoplastic polyurethane were found to be functions of crosslinking degree and mass fraction of soft segments, respectively. The results obtained from the simulations are in good agreement with experimental data, and it is revealed that heat in polymers is mainly transferred by low-frequency phonons.