Multiscale Simulation Methods I
Multiscale
Simulations of Materials
Presented by -
Vishwas Manjunath , Matriculation no- 22061633
Tutor -
Dr. Stefan Sandfeld
Department für Werkstoffwissenschaften Lehrstuhl für
Werkstoffsimulation (WW8)
Introduction
• The most interesting problems in all fields of science involve multiple spatial or temporal
scales.
• The microscopic constituents of materials are atoms, and the interactions among them
at the microscopic level determine the behavior of the material at the macroscopic scale.
• The idea of performing simulations of materials across several characteristic length and
timescales has, a great effect on technological innovation.
Multiscale Simulation Methods I Fig[1]: Characteristic length scales 2
In the context of materials simulations, we have four characteristic length levels
1. Atomic scale(∼10−9/m or a few nanometers): The electrons are the players, and
their quantum-mechanical state dictates the interactions among the atoms.
2. Microscopic scale (∼10−6/m or a few micrometers): Atoms are the players
and their interactions can be described by classical interatomic potentials (CIP) that
has the effects of bonding between them, which is mediated by electrons.
3. Mesoscopic scale (∼10−4/m or hundreds of micrometers): Here lattice defects
such as dislocations, grain boundaries, and other microstructural elements are the
players.
4. The macroscopic scale (∼10−2/m or centimeters and beyond) : In the
macroscale, continuum fields such as density, velocity, temperature, displacement
and stress fields, and so forth are the players. The constitutive laws are usually
formulated so that they can capture the effects on materials properties from lattice
defects and microstructural elements.
Multiscale Simulation Methods I 3
Different computational approaches
QMC : To treat electrons
explicitly and accurately at the atomic scale,
QMC can be employed.
DFT: Applied to systems containing
several hundred atoms for static properties.
TBA : Can extend the reach of simulations to
a few nanometers in linear size and a few
nanoseconds in timescale for the dynamics.
CIP: provide insight into atomic processes
involving considerably
larger systems, reaching up to ∼10^9 atoms.
Fig:[2] Schematic representation of various simulation approaches
At the mesoscopic scale, a recent progress has been made on Dislocation Dynamics which
can study systems a few tens of microns in size and with a maximum strain ∼05% for a
strain rate of 10 sec−1 in bcc metals
For the macroscopic scale, finite-element (FE) methodsare routinely used to examine the
large-scale properties of materials considered an elastic continuum
Multiscale Simulation Methods I 4
Other Sequential Approaches
• Kinetic Monte Carlo (KMC) simulations, coupled with atomistically determined kinetic
energy barriers, represent a powerful class of sequential multiscale approaches.
• KMC employs the level-set method for the largest (macroscopic) scale.
• In the lateral direction, continuum equations representing the field variables can
be coupled to growth through island evolution by solving the appropriate boundary-value
problem for the field and using local values of this field to determine the velocity of the
island boundaries.
• For a given boundary velocity v, the equation for j is [2]
• Growth is naturally described by the smooth evolution of j, determined by this
differential equation.
Multiscale Simulation Methods I 5
Level-set multiscale approach of surface growth
The level-set method is a manifestly
multiscale approach, combining
information from three different regimes
(atomistic, mesoscopic, and continuum)
into a neatly integrated scheme.
• The atomistic processes that are
responsible for terrace or island-edge
diffusion are identified and their energetics Fig[2]:Macroscopic Mesoscopic Microscopic
analyzed using atomistic calculations.
• The energy barriers for the atomistic processes are incorporated in a KMC model that
provides the means for coarse-graining the atomistic degrees of freedom to a few
mesoscopic degrees of freedom, describing the evolution of surface features.
• Finally, the results of the KMC model are coarse-grained to provide the input to the
level-set equations—that is, they define the values of the boundary velocity v, which
depends on the local surface morphology.
Multiscale Simulation Methods I 6
References
[1] National Technical University of Athens ,School of Chemical Engineering
Department of Materials Science and Engineering: Computational Materials Science
and Engineering Group: Objectives
[2] Overview of Multiscale Simulations of Materials, Gang Lu, Efthimios Kaxiras,
ISBN: 1-58883-042-X, Volume X: Pages 1-4, 10-12
Multiscale Simulation Methods I 7
