shrunk volume upon tension under NPT emsemble

LAMMPS LAMMPS Mailing List Mirror
1

Hello all,

I performed a tensile test on a slab within a box. The slab is periodic along x and z and is separated from its images along y- by a vacuum of 200-angstrom. It was noted that the volume of the box constantly decreased and after step 400000 no vacuum was left in the box. I was wondering why this happened as I set p_x and p_y to be zero. Thanks for your help.

Thanks,

Chunguang

-----------Input script:--------------------------------

units metal
boundary p p p
atom_style atomic
read_restart restart.*
reset_timestep 0
pair_style meam
pair_coeff * * \{home\}/code/lammps/potentials/Jelinek\_2012\_meamf FeS SiS {home}/code/lammps/potentials/Jelinek_2012_meam.alsimgcufe FeS SiS
dump dumptrj all atom 10000 trj/trj.*.lammpstrj
dump_modify dumptrj scale no sort id pad 6
fix fnpt all npt temp 300 300 0.1 x 0.0 0.0 1.0 y 0.0 0.0 1.0
fix ftensile all deform 10 z erate 0.0001 remap v # strain-rate=10e-4 per ps
compute truetemp all temp/deform
compute stress all pressure truetemp
thermo 1000
thermo_style custom step temp c_truetemp pe ke etotal vol lx ly lz press
timestep 0.002 # 1 fs per step
run 1500000

-------------step and box volume: ---------------------------

0 8070631.0000
100000 7733787.0000
200000 6662421.0000
300000 5278614.0000
400000 4355766.5000
500000 4364334.5000
600000 4368085.5000
700000 4367421.0000
800000 4365988.0000
900000 4365405.0000
1000000 4364855.5000
1100000 4364121.5000
1200000 4363222.5000
1300000 4362278.0000
1400000 4363028.0000
1500000 4363665.5000

2

Your boundary condition along y is “p” (periodic) and you are relaxing the pressure along y (via fix npt … y 0.0 0.0 1.0). There is no reason that your box dimension should not change along y. What you saw was an expected result of your input.

p.s. 200 angstrom of vacuum is unnecessarily large. Any vacuum that is sufficiently larger than your pair cutoff is enough.

3

Thanks Ray.

Yes, I understand that box dimension along y can change somewhat due to npt along y. But the constant decrease of y dimension consumed the vacuum such that the ’ free surface’ essentially disappeared. Probably the pressure along y was not modulated correctly by npt.??

4

Thanks Ray.

Yes, I understand that box dimension along y can change somewhat due to npt
along y. But the constant decrease of y dimension consumed the vacuum such
that the ' free surface' essentially disappeared. Probably the pressure
along y was not modulated correctly by npt.??

no, it was. what you are describing is the expected behavior for your
setup. if you don't want the vacuum to go away, you must not couple y
to the barostat.

axel.

5

Thanks Axel. Your suggestion sounds reasonable - y dimension should not change if decoupled from the barostat. The reason that I raised the question is in some cases the volume only changed slightly when y was coupled and so I thought the large change was abnormal.