Implementation and performance of the needle chain-algorithm for Brownian dynamics simulation

摘要

The first algorithm for Brownian dynamics (BD) simulation of needle chains (the NC-algorithm) was developed recently by Nyland et al. (J. Chem. Phys. 1996;105:1198). Here we report on an implementation of the NC-algorithm using N_n rigid segments composed of N_(bn) spherical beads, and its efficiency relative to the Ermak-McCammon BD algorithm (J. Chem. Phys. 1978;69:1352) used to simulate bead-spring polymer chains consisting of N_b beads arranged into N_n stiff segments each composed of N_bn beads. For segmented chains the NC-algorithm can use a dimensionless time-step that is about four orders of magnitude larger than that of the bead-spring BD algorithm. In addition, the computer time required to perform a single dimensionless time-step scales as N_b~2 and N_b~3 for the NC-algorithm and the bead-spring algorithms, respectively. This demonstrates that, for polymer chains consisting of needles or rigid segments with aspect ratios larger than three, the NC-algorithm is vastly more efficient than the Ermak-McCammon BD algorithm.