(332f) Optimized Generation of Initial Conformations for the Simulation of Amorphous Polymer Systems to Reduce Required Simulation Resources

Significant breakthroughs have been derived from the simulation of non-crystalline polymer phases including optimization of such polymers for mechanical, membrane, surface, biomedical and microelectronics applications. Unfortunately, these simulations require long simulation times when the molecular weight of interest is at or above the entanglement length of the polymer chain. Some reduction in the required simulation time is attainable through the use of simulation algorithms that efficiently sample conformation space, but an additional reduction can be obtained by producing an accurate initial conformation of the polymer system that is close to the equilibrated structure. This work investigates the effect of density on the initial conformation and the importance of long-range corrections in both cohesive energy and density. We utilize the Suter-Theodorou initial guess algorithm that employs a Rotational Isomeric States model for short range conformational choices coupled with long range self-avoiding corrections. However, we carry this out as a function of the density of the periodic cell. Depending on the nature of the molecular interactions (dispersion, polar, and hydrogen bonding), there appears to be an optimal density at which the initial conformation is generated. Generation of the initial guess at the experimental density can produce unrealistically high energy conformations. While generation at lower density can alleviate these high energy conformations, it also reduces the chain entanglement. This entanglement is significant because chain diffusion is slow, and insufficient entanglement requires extensive simulation time to recover. By manipulating density, we can optimize both energetics and entanglement to produce an initial conformation close to the equilibrium structure of the amorphous polymer. Additionally, we utilize an updated protocol for calculating the long-range energy and pressure corrections that extrapolates a series of reproduced periodic structures. This approach avoids the conditional convergence of the many of the commonly used methods for electrostatic energy correction.