Title: Atomistic simulations of contact physics Alejandro Strachan Materials Engineering strachan@purdue.edu
1Atomistic simulations of contact
physicsAlejandro StrachanMaterials Engineering
strachan_at_purdue.edu
2Atomistic materials simulations in PRISM
Develop first principles-based constitutive
relationships and provide atomic level insight
for coarse grain models
- Identify and quantify the molecular level
mechanisms that govern performance, reliability
and failure of PRISM device using - Ab initio simulations
- Large-scale MD simulations
3PRISM multi-physics integration
- Trapped charges in dielectric
Predictions
Electronic processes
Validation Experiments Microstructure evolution,
device performance reliability
PRISM Device simulation MPM FVM
- Elastic, plastic deformation, failure
Micromechanics
- Defect nucleation mobility in dielectric
Fluid dynamics
- Dislocation and vacancy nucleation mobility in
metal
Atomistics
Thermal and mass transport
- Thermal electrical conductivity
Input Experiments Surface roughness,
composition, defect densities, grain size and
texture
4Atomistic modeling of contact physics
How classical MD with ab initio-based
potentials Size 200 M to 1.5 B atoms Time
scales nanoseconds
Predictions Role of initial microstructure
surface roughness, moisture and impact velocity
on
Force-separation relationships (history dependent)
Generation of defects in metal roughness
evolution
Mechanical response
Generation of defects in dielectric (dielectric
charging)
Thermal role of electrons in metals Current
crowding and Joule Heating
Electronic properties
Surface chemical reactions
Chemistry
Main Challenges
5Atomistic modeling of contact physics II
Smaller scale (0.5 2 M atom) and longer time
(100 ns) simulations to uncover specific physics
- Mobility of dislocations in metal,
- Interactions with other defects (e.g. GBs)
- Link to phase fields
- Surface chemical reactions
- Reactive MD using ReaxFF
- Defects in semiconductor
- Mobility and recombination
- Role of electric charging
- Fluid-solid interaction
- Interaction of single gas molecule with surface
(accommodation coefficients) for rarefied gas
regime
6Obtaining surface separation-force relationships
- Contact closing and opening simulation
- 200 M to 1.5 billion atoms nanoseconds
- (1 billion atom for 1 nanosecond 1 day on a
petascale computer) - Characterize effect of
- Impact velocities (4 values)
- Moisture (4 values)
- Applied force and stress (2 values)
- Surface roughness
- Peak to peak distance (2) and RMS (2)
- Presence of a grain boundary (4 runs)
16 runs
4 runs
4 runs
4 runs
28 runs
7Upscaling MD to fluid dynamics
Given a distribution of incident momenta
characterize the distribution of reflected
momenta
Accommodation coefficients
pi
Fluid FVM models use accommodation coefficients
from MD and predict incident distribution
Role of temperature and surface moisture on
accommodation coefficients
8Upscaling MD to electronic processes
- Defect formation energies
- Equilibrium concentration
- Formation rates if temperature increases
- Impact generated defects
- Characterize their energy and mobility as a
function of temperature - Predict the distribution non-equilibrium defects
- Characterize energy level of defects
- SeqQuest
9Upscaling MD to micromechanics
- Elastic constants
- Vacancy formation energy and mobility
- Bulk and grain boundaries
- Dislocation core energies
- Screw and edge
- Dislocation nucleation energies
- At grain boundaries, metal/oxide interface
- Nucleation under non-equilibrium conditions
(impact) - Dislocation mobility and cross slip
- Interaction of dislocations with defects
- Solute atoms and grain boundaries
Upscaling MD to thermals
- Thermal conductivity of each component
- Interfacial thermal resistivity
- Role of closing force, moisture and temperature
10MD simulations challenges
- Accurate interatomic potentials
- Start with state-of-the-art
- Parameterize using ab initio calculations
(ReaxFF, MEAM) - Incorporate thermal and transport role of
electrons - Accurate description of thermal transport and
Joule heating - Extend new method for dynamics with implicit
degrees of freedom - Strachan and Holian, Phys.
Rev. Lett. (2005)