Anisotropic coarse-grained models for the structural characterization of unentangled linear polytetrafluoroethylene (PTFE)*

Polytetrafluoroethylene (PTFE) is widely employed in corrosive or high-temperature applications due to its excellent chemical inertness and thermal stability. However, melt processing is commonly problematic owing to its high viscosity in the melt state. The symmetric carbon-fluorine chemical structure is the primary factor that makes PTFE melts have strong van der Waals interactions in high molecular weight. Molecular dynamics (MD) methods have been successfully applied to gain physical insights into all-atom (AA) and united-atom (UA) scales. However, the resolution of details requires expensive computational costs that prohibit their applicability to more extensive time and length scales. Given the inherent rigidity nature of PTFE, we propose the anisotropic coarse-grained (CG) models utilizing RE-squared potential in two CG levels, equivalent to 6 and 8 CF₂ groups. The bottom-up CG models are optimized through the mapping of critical structural properties based on atomistic simulations of C₉₆F₁₉₄ at 600K. These properties include segmental geometry, angular flexibility, pair correlations, and overall chain dimensions. Additionally, the impact of CG levels on dynamic characteristics, such as self-diffusivity and zero shear viscosity, is studied. Both CG models demonstrate adequate predictive capabilities regarding the structural characteristics of PTFE melts, encompassing molecular weights from 48 to 192 carbons per chain. Furthermore, the zero-shear viscosity presents a similar trend compared with AA systems across different molecular weights in both models. Benchmark evaluations for the C₉₆F₁₉₄ system reveal that the computational efficiency can be improved approximately 60 times than their AA counterparts. Therefore, our CG models offer accurate and efficient predictions of PTFE properties across different molecular weights, and are potentially applicable to study other PTFE-based polymers.