Hi,
I have a very basic question about the way “INF” command works in region definition.
I have set up a simple single crystal aluminum with the lattice parameter of 4.05 Angstroms and I am performing lattice static minimizations (0 K). The simulation box is periodic in all 3 directions.
My question is about how exactly INF command works in conjunction with periodic boundary conditions.
If I define:
inter-planar spacing: i_pl_s= a / sqrt(h^2+k^2+l^2)
and,
same plane spacing: s_pl_s = a * sqrt(h^2+k^2+l^2)
The simulation box size in X and Z directions are integer multiples of i_pl_s. In Y direction however the simulation box size is an integer multiple of s_pl_s.
I define the box sizes in X, Y, Z directions as (51.6275725751, 204.915592379, 50.2340675086) in box named “whole” and I define a region with “INF” in all directions allowing LAMMPS to fill up the space in “whole” with aluminum atoms.
lattice fcc ${LatParam}
region whole block 0.0 51.6275725751 -204.915592379 0.0 0.0 50.2340675086 units box
create_box 2 whole
region lower block INF INF INF INF INF INF units box
lattice fcc ${LatParam} orient x -1 3 -4 orient y 3 1 0 orient z 2 -6 -5
create_atoms 1 region lower
When I minimize the structure I see defects on surfaces normal to Z axis, suggesting that the length of the box in Z direction is not a periodic length hence the crystal does not converge to a perfect configuration. The question is; why isn’t this happening on normal to X surfaces? The simulation box in X direction is not a multiple of s_pl_s either.
To isolate the effect of minimization I store the pre-minimization structure of the crystal into a file named “dump.before”. The same trend can be observed in pre-minimized structure. Only the shape of defects are different which is expected.
I have attached the input file (in.1) as well as the EAM potential that I am using (Al99.eam.alloy).
Any insights on how exactly the INF command is developed in LAMMPS is highly appreciated.
LAMMPS Version: (Aug 8 2014)
Al99.eam.alloy (762 KB)
in.1 (2.09 KB)