Atomistic Modelling of Liquid Crystal Phases

Abstract

The last few years have seen a large number of simulation studies of liquid crystal systems. Two general methods have been popular. In the first approach, simplified molecular potentials have been used. These seek to approximate the gross features of mesogenic interactions (e.g molecular shape anisotropy, molecular dipole moments etc.) by single site potentials [1–6]. Such models have been highly successful in predicting the formation of a number of liquid crystal phases, and allowing the development of techniques for the calculation of key bulk properties, such as transport coefficients [7] and elastic constants [8]. In a second simulation approach, atomistic models have been used [9–26]. These seek to represent molecular structure as faithfully as possible by representing nonbonded interactions at the atomic level and accurately representing internal molecular motion. This approach aims to be able to accurately predict mesophase behaviour for real systems based only on a knowledge of molecular structure. The drawback of this method is that it is computationally expensive, and large quantities of computer time must be expended in order to compute the numerous pair interactions in such systems.