Overlap of atoms in modeling


I’m trying to model a substrate like structure with Si.

I modeled the substrate with two layers by defining regions separately.

One is a fixed region and the other one is a movable region.

However, through LAMMPS log file I could see that there are some overlap between fixed and movable region.

(group ‘substrate’ = group ‘afix’ + group ‘amov’)

(# of atoms in group ‘substrate’ = 2,600)

(# of atoms in group ‘afix’ = 1,000 and # of atoms in group ‘afix’ = 1,800)

Does this indicate 200 atoms are overlapped? How does it affect the simulation?

Furthermore, if this affects the simulation how can I avoid this overlapping?

Here’s the part of the input I’ve used and the LAMMPS log file.

Best regards

Hyun-seok Kim.


lattice diamond 5.43

region boxsim block 0 10 0 10 0 15 units lattice

region boxf block 0 10 0 10 0 1 units lattice

region boxm block 0 10 0 10 1 3 units lattice

region clus sphere 5 5 7 3 units lattice

region sub union 2 boxf boxm

create_box 1 boxsim

mass 1 28.0855

create_atoms 1 region sub

create_atoms 1 region clus

group afix region boxf

group amov region boxm

group substrate region sub

group cluster region clus

group runnve region boxsim

(LAMMPS log)

group afix region boxf

1000 atoms in group afix

group amov region boxm

1800 atoms in group amov

group substrate region sub

2600 atoms in group substrate

group cluster region clus

915 atoms in group cluster

group runnve region boxsim

3515 atoms in group runnve

It is almost always a bad idea to put region or box boundaries exactly on lattice points. That makes selecting atoms ambigous or dependent on rounding. Shift them by a small margin and you will see.