(648f) Applying Protracted Colored Noise Dynamics to Dramatically Increase the Simulation Efficiency of Linear Polymer Systems

Ludovice, P. J., Georgia Institute of Technology
Peters, A., Louisiana Tech University
Nation, B., Georgia Institute of Technology
Henderson, C. L., Georgia Institute of Technology
Molecular dynamics simulations of polymers are significantly limited because of the large time and length scales required by the viscous nature of polymeric systems. A previous stochastic simulation method called Protracted Colored Noise Dynamics (PCND) increased the local diffusivity of highly viscous molecular systems by applying correlated stochastic forces (colored noise). This method dramatically increased the diffusivity of a molecular glass and resulted in a crystallization rate that was approximately three orders of magnitude faster than normal molecular dynamics systems. We have applied PCND to linear polymer systems by directing the stochastic colored noise along the backbone of the polymer contour. This increases the reptation mode of the polymer, and hence the local diffusivity. We tested the effectiveness of this new method using simulations of defect annealing in lamellae formed by the directed self-assembly of block copolymers. This polymer PCND increased defect annihilation by approximately four orders of magnitude compared to molecular dynamics. Application of molecular dynamics restores the correct domain size, indicating that the free energy landscape is not being significantly altered. The computational overhead associated with polymer PCND is minimal, and outweighed by the increased equilibration rate. We also studied the effective parameter space of the time relaxation of the colored noise and the magnitude of the colored noise. If these parameters are too large, the stochastic forces overwhelm the system and do not produce equilibrated structures. The effectiveness of the method is due, in part, to a strategic violation of the fluctuation dissipation theorem. This does perturb the dynamic trajectory of the simulation, but correct structures are produced by switching off the stochastic force at the end of the simulation. This makes this application of PCND very effective at accessing polymer conformation space that contains significant energy barriers.