Problem in incorporating screw dislocation into lammps

Dear experts,

I have created a screw dislocation in a W crystal using Atomsk (w_dislo.lmp).

I want to use this as the initial position of the dislocation in one side of the Peierls valley. (And my final aim is to calculate the Peierls barrier through NEB, where the final position of the dislocation is on the other side of the barrier.)

I have to relax the dislocation in the initial position as in Atomsk it is created using the isotropic elastic displacement solution.

Please find attached the in.neb file which I used to relax the initial dislocation using the Conjugate Gradient algorithm.

Now the initial output file dump.cg-minimize-initial_0 is not showing the atomic structure that I have input as w_dislo.lmp.

Is it due to the reason that in case of introduction of dislocation, the origin have to the shifted to the x, y position of the dislocation line?

From Atomsk,

Inserting a screw dislocation with line along z,
b=[0.000 0.000 2.741] at (116.290,70.124)

The dislocation was introduced at x=116.290, and y=70.124

Please help me with this anomaly,

Thanking you,

Sincerely,

Krishnendu

w_dislo.lmp (359 KB)

in.neb (661 Bytes)

dump.cg-minimize-initial_0 (157 KB)

Dear experts,

I have created a screw dislocation in a W crystal using Atomsk (w_dislo.lmp).

I want to use this as the initial position of the dislocation in one side of the Peierls valley. (And my final aim is to calculate the Peierls barrier through NEB, where the final position of the dislocation is on the other side of the barrier.)

I have to relax the dislocation in the initial position as in Atomsk it is created using the isotropic elastic displacement solution.

Please find attached the in.neb file which I used to relax the initial dislocation using the Conjugate Gradient algorithm.

Now the initial output file dump.cg-minimize-initial_0 is not showing the atomic structure that I have input as w_dislo.lmp.

that is not correct. both structures look equivalent. you are disregarding periodic boundaries. your data file has atoms outside the principal simulation cell, but due to PBC, LAMMPS will wrap them into the principal cell and adjust the image flags. in the dump output, you only output the coordinates from the principal simulation cell, but not the image flags (or unwrapped coordinates) and thus there appears to be a difference. if you would output unwrapped coordinates, both geometries should look equivalent.

axel.