Stacking Fault Energy calculation for hexagonal system, prismatic plane

Hi,

I try to calculate the stacking fault energy for Mg. Although my input
script works for the calculation of γ surface for the basal plane (with
good results) it fails to calculate the γ surface for the prismatic plane.
In the attached script I perform an initial minimisation, then a
displacement and then a minimisation again. For the prismatic plane, my
system has periodic boundaries along the y and z axis and non-periodic
and shrink-wrapped along the x axis. The box is triclinic with tilt
factors along xy.

When I apply displacement along the y axis for some specific atoms (half
of the atoms along the x axis) then the atoms move according to my
displacement.
When I apply displacement along the z axis for the same atoms, then all
the atoms of the system move (not only the chosen atoms) and not
according to my displacement but for a different value.
For both cases of displacement along either y or z axis the results are
not the expected from theory results (not even qualitatively).

I believe there is something related to the triclinic box but I can not
trace it.. Any help?

Thanks
Alex

Mg8x.pos (814 Bytes)

Mg1.eam (727 KB)

mgSFExyz.in (1.52 KB)

Hi,

I try to calculate the stacking fault energy for Mg. Although my input
script works for the calculation of γ surface for the basal plane (with
good results) it fails to calculate the γ surface for the prismatic plane.
In the attached script I perform an initial minimisation, then a
displacement and then a minimisation again. For the prismatic plane, my
system has periodic boundaries along the y and z axis and non-periodic
and shrink-wrapped along the x axis. The box is triclinic with tilt
factors along xy.

When I apply displacement along the y axis for some specific atoms (half
of the atoms along the x axis) then the atoms move according to my
displacement.
When I apply displacement along the z axis for the same atoms, then all
the atoms of the system move (not only the chosen atoms) and not
according to my displacement but for a different value.
For both cases of displacement along either y or z axis the results are
not the expected from theory results (not even qualitatively).

I believe there is something related to the triclinic box but I can not
trace it.. Any help?

i just took your inputs and "decorated" it with a couple of write_dump
image commands to write out a configuration before and after a
displacement and they work just as expected.

most likely something else in your input doesn't work as expect it,
possibly due to misunderstanding how they work. difficult to say from
the outside. how to debug your input is ultimately your
responsibility. i don't see anything where LAMMPS doesn't work as it
should.

axel.