Chemistry Reference
In-Depth Information
of resolution and, consequently, on several time and length scales. The easiest way
to combine different simulation models on different scales is to treat them sepa-
rately and sequentially by simply passing information (structures, parameters,
energies etc.) from one level of resolution to the next. A step beyond these
sequential schemes is involved in those approaches where the scales are coupled
in a concurrent fashion within a unified computational scheme. In these approaches,
two or more levels of resolution are used at the same time in the simulation. A dual-
scale approach has already been used to study the interaction between bisphenol-A
polycarbonate and a nickel surface [
102
,
103
]. In this method, the regions with
different resolutions are fixed and the exchange of particles among the different
regions is not allowed. While this may not be a crucial point for hard matter, it is
certainly a strong limitation for soft matter (i.e., complex fluids) since relevant
density fluctuations are arbitrary. An even more sophisticated multiscale approach
allows adaptive switching between resolution levels for individual molecules on the
fly, e.g., depending on their spatial coordinates. Recently, such an adaptive resolu-
tion scheme (AdResS) has been developed in which molecules can freely exchange
between a high-resolution (CG) and a low-resolution (atomistic) region by chang-
ing the molecular degree of freedoms [
104
-
110
]. In this case, the atomistic and the
CG scales can be coupled via a position-dependent interpolation formula on the
atomistic and CG force in such a way that allows a smooth transition from atomistic
to CG trajectories without altering the equilibrium of the system [
111
]. The method
has been already used for liquid water [
105
] and for a polymer-solvent system in
which the water molecules within few solvation shells around the polymer chains
are considered atomistically while outside the water is treated on a rather coarse
level [
106
]. It has even been augmented by a continuum region, and a methane-like
liquid has been simulated using this triple-resolution scheme [
112
].
3.3 Coarse-Graining in Time
Although CG models have been successfully used to simulate large systems for
very long time and length scales, the lack of detailed atomistic information in CG
simulations still limits the systems and the properties that can be studied using these
models. As an alternative to the spatial coarse-graining techniques, Violi [
113
]
proposed a novel method to describe the evolution of reactive systems (diffusion
processes and chemical reactions) over long time scales while preserving an all-
atom description of the system by coarse-graining in time. The method combines
the MD methodology with kinetic Monte Carlo (KMC) to allow the extension of the
accessible time scales compared to the direct MD simulation [
114
]. In the KMC
step, the structure of the growing species is modified during the reaction and then
the newly formed structure is relaxed towards thermal equilibration using an MD
run. The MD describes the local phase space changes and rearrangement reactions
and allows for relaxation as well as processes very far from equilibrium. The KMC
method is responsible for the conformational changes that jump to a completely
Search WWH ::
Custom Search