5. Multiscale modeling 1.0
The acceptance of multiphysics simulation techniques has helped bridge
the gap between theory and experiment [62]. Electronic structure
(quantum level or ab initio ) simulations can reveal how specific
molecules assume stable geometrical configurations and charge
distributions when subject to a specific chemical environment. By
examining the charge distributions and structure, it is possible to
quantify and predict structural properties as well as chemical
reactivity pertaining to the molecule, which is particularly pertinent
when investigating novel materials. Although the quantum simulations
provide a wealth of information regarding structure and reactivity, it
is currently not possible to model much more than a few hundred atoms at
most. Molecular dynamics simulations based on classical (empirical)
force-fields can model hundreds of thousands of atoms for tens of
microseconds in time. Since MD simulations can be set up at atomic
resolution, they are uniquely suited to examine thermodynamic and
statistical properties of (bio)materials: such properties include (but
not limited to) Young’s modulus, surface hydration energies, and protein
adsorption to different surfaces [63]. Coarse-grained or mesoscale
simulations are used to bridge the gap between the atomistic scale of MD
simulations and continuum approaches such as elasticity theory or
hydrodynamics at the macroscale (i.e., milliseconds, millimeters and
beyond) [62].
The ultimate purpose of multiscale modeling is to predict the
macroscopic behavior from the first principles. Finding appropriate
protocols for multiscale simulations is also challenging as either
multiphysics simulations need to operate at multiple resolutions, or two
or more multiphysics simulations need to be combined. In general, these
are achieved via adaptive resolution schemes, coarse-graining,
sequential multiscale modeling, concurrent multiscale modeling, and
enhanced sampling schemes [18, 64], see Table 5.