Generation-relaxation algorithms to construct representative atomistic models of amorphous polymers: Influence of the relaxation method

Two alternative methods to relax the atomistic models of amorphous polymers generated by structure simulation (SuSi), a procedure recently developed that combines a generation algorithm and a minimization algorithm to localize independent energy minima (J. Chem. Phys. 2003, 119, 2915), are presented. These relaxation methods, which are based on the geometric aspects of the conventional configurational bias (CB) and concerted rotation (ConRot) Monte Carlo techniques, have been compared with the minimization algorithm originally implemented in SuSi. For this purpose, we considered two polymeric systems: four polyethylene chains with each one containing 400 CH 2 groups and eight polyethylene chains with each one containing 800 CH 2 groups. The procedure based on ConRot has been demonstrated to be the most efficient of the three relaxation algorithms, while the minimization one is the least efficient. On the other hand, the method based on CB is limited by the difficulties to relax the interior segments of chain molecules.