Molecular dynamics using coarse-grained models

Molecular Dynamics Using Coarse-Grained Models, 206... 

Besides microscopic technologies, the molecular dynamics simulation using coarse grain models to speed up the simulations improved the understanding of the effects of tentacle surface modifications (Cavalloti, 2008). The tentacular structure of lEX materials is characterized by a macroscopic reorganization caused by... 

As discussed above, the large length- and time-scales involved in most experimental realizations of gelation and other aspects of aerogel preparatimi are not directly accessible by molecular simulations. An alternative approach is to use coarse-grained models, which alleviate these scale problems at the expense of the loss of atomistic detail. The construction of coarse-grained models affords considerable freedom in choice of the primary objects simulated, their interaction, and the associated dynamics so much so, in fact, that much work has focussed on the simplest possible models, both to expose the most fundamental physics involved and to avoid laborious and possibly underdetermined parametrization problems. 

In most cases, the DNA CG models presented here are used in Langevin-based molecular dynamics (MD) simulations that propagate the positions and velocities of the CG sites in time and space. As described in the examples, when the appropriate sampling strategies are used, these simulations provide detailed information about the conformation space, in particular, appropriately describing the thermodynamics associated with the different conformations. Moving beyond equilibrium structural distributions, much less attention has been paid to the dynamical behavior of molecular systems in coarse-grained simulations, which should be addressed in future studies. 

This chapter will focus on a simpler version of such a spatially coarse-grained model applied to micellization in binary (surfactant-solvent) systems and to phase behavior in three-component solutions containing an oil phase. The use of simulations for studying solubilization and phase separation in surfactant-oil-water systems is relatively recent, and only limited results are available in the literature. We consider a few major studies from among those available. Although the bulk of this chapter focuses on lattice Monte Carlo (MC) simulations, we begin with some observations based on molecular dynamics (MD) simulations of micellization. In the case of MC simulations, studies of both micellization and microemulsion phase behavior are presented. (Readers unfamiliar with details of Monte Carlo and molecular dynamics methods may consult standard references such as Refs. 5-8 for background.)... 

In contrast to Monte Carlo simulations, the molecular dynamics method simulates the motions of molecules however, this can only be extended up to very short time-scales because the time step per cycle is so small. Thus, the method cannot be used to model collective motions, which are important in diffusion and hydrodynamic flows, for example. For this it is necessary to turn to more coarse-grained models, where molecular details are ignored. Here the properties of collections of molecules are simulated, or the fluid structure is simulated at an even larger length-scale where the system behaves as a continuum. Some of these methods are now discussed. 

Figure 10 Molecular dynamics simulation of a surfactant system using a coarse-grained model. A surfactant molecule consists of four head (light grey) and four tail sites (white) a water molecule is represented by one site (dark grey). The inset on the lower right shows the random distribution of the molecules at the beginning of the simulation.

In the next section, we review in more detail the relevant theoretical concepts. In Section 4.3, we discuss the different simulation methods, both Monte Carlo and molecular dynamics, which have been used to study both the static and dynamic properties of a melt. We also show why it is essential to use a coarse-grained model, instead of a more realistic model, if one is to have any chance of addressing the issues related to whether reptation theory is correct or not. In Section 4.4, we review the progress which has been made recently in understanding the motion of long linear chains in a melt. In Section 4.5, we consider the properties of highly crosslinked networks or rubbers. Finally, in Section 4.6, we briefly summarize our main results and present our outlook for future simulations in this area. 

---