Chemistry Reference
In-Depth Information
accurate transport properties; more work needs to be done to test the accu-
racy obtained with fixed-charge versus polarizable models. Finally, when
care is taken, transport properties can be modeled accurately using atomistic
simulations. This is encouraging because it suggests that ionic liquids may not
be so ''special'' after all and that the simulation techniques developed and
applied to the study of conventional liquids can be used successfully to study
ionic liquids as well.
COARSE-GRAINED MODELS
Fully atomistic simulation of an ionic liquid is computationally demand-
ing. Using state-of-the-art computing clusters and advanced codes, a
reasonable MD simulation at a single liquid-state point may take several
days. Multiple-state points can of course be run in parallel if processors are
available. Moreover, highly parallelized MD codes such as NAMD,
66
LAMMPS,
122
and DL_POLY
35
can speed up these calculations significantly.
MC calculations are embarrassingly parallel and can also be conducted using
a ''job farm'' approach.
123
Running long enough simulations to derive reliable
thermodynamic and transport properties, especially at lower temperatures, is
still quite challenging, Although simulations of systems containing more than
10 million atoms have been carried out, ''ordinary'' calculations usually
involve no more than tens of thousands of atoms and length scales of a few
nanometers. If one is interested in computing properties that emerge only
over very long length scales (say micron scale) are greater, performing a
detailed atomistic simulation quickly becomes intractable. If having absolute
accuracy is less important than deriving a qualitative insight, then performing
detailed simulations may not be the best selection.
One approach that can be used to speed up calculations and enable lar-
ger systems to be examined is to develop a
coarse-grained
model of the ionic
liquid. A classical atomistic simulation can already be thought of as a
coarse-
grained
quantum calculation in which the detailed treatment of electronic
degrees of freedom are replaced by semiempirical analytic functions. The
type of coarse graining referred to in this section goes even further, by coarse
graining the classical atomistic model. In essence, coarse graining seeks to
eliminate nonessential degrees of freedom to arrive at a simplified representa-
tion of the system. Because there are fewer degrees of freedom when using a
coarse-grained model and, often, the interactions between degrees of freedom
are simple, calculations can be fast and very large systems can be studied for
long periods.
The literature on coarse graining is vast and cannot be treated in any
detail here. There are many examples of coarse-graining strategies for bio-
molecules
124
and polymers,
125
but there have not been as many studies for
ionic liquids. A tutorial on coarse graining has been published in this topic
