Build Nanocrystalline Materials and MD Simulations
Most metallic materials are polycrytalline, rather than a perfect single crystal. The properties of these materials have dependence on the size of the grains (polycrystals), such as yield strength. Conventionally, the Hall-Petch effect captures the relation between yield strength and grain sizes. However, at the nanoscale, materials behave differently compared to at scales larger than millimeters, showing inverse Hall-Petch effect in many nanocrystalline materials. In this tutorial, the generation of nanocrytalline Aluminum using Atomsk and its MD simulations using LAMMPS is demonstrated.
Initial Nanocrystalline Al System Buildup
Atomsk was used to build the system. The system (a cubic box) size is 5 nm wide. The number of grains is 4, 10, 30, and 100, giving rise to the average grain size: 2 nm, 1.5 nm, 1.0 nm, and 0.667nm, respectively. Here are the initial systems.
Here is a sample code to generate these systems:
#!/bin/bash
# generate the unit cell
atomsk --create fcc 4.046 Al aluminium.xsf
for i in 4 10 30 100
do
input=seed$i.txt
# generation
atomsk --polycrystal aluminium.xsf $input final$i.cfg
# convert to lammps data file
atomsk final$i.cfg lammps
done
where seed$i.txt is as follows
box 50 50 50
random i // i from 4/10/30/100
MD Simulation in Lammps
Lammps (version Aug-7-2019) was used as the MD simulation engine for the uniaxial tensile tests of the nanocrystalline Al systems and a single crystalline for comparison. The simulation protocol follows the classical tutorial given here. The tension test was carried out at 300 K and 1 bar using the EAM potential. Here is the result of the stress-strain curves for single crystalline Al and nanocrystalline Al (inset for zoom in of the nanocrystalline systems for comparison of size effect).