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.