why are you using such boundary conditions? You seem firing an ion in the Y direction, and then choose a “s” boundary condition. Is the condition “f” not better for your case? If by top/bottom you meant “in the Z direction”, so there is no physical top/bottom. @+ Julien.
Dear Julien,
Thanks for your answer!
]I am sorry I do not quite understand why ‘f’ is better. The velocity of PKA is along the radial direction, which in my case is along the Y direction. The cylinder represents a nanowire. And I think the boundary condition of nanowire should be free at radial directions and periodic at axis direction.
The z direction is periodic, so I think the defects should not appear along this direction. I cannot figure out why there are so many defects at the ends of the z direction of my simulation box.
With My Best Regars
Liu
If I understand correctly, commenting the following lines seems to eliminate the defects:
#region PKA_Au sphere 0 (-v_r) (v_l/2) 0.1 side in
#delete_atoms region PKA_Au compress no
#create_atoms 1 region PKA_Au
It may be that atoms are being created that overlap strongly with existing atoms. I am not sure why you need to delete and then create atoms. Why not just assign the existing atoms to the group Cascade_Au.
Also, I am suspicious of the preceding command:
#change_box all x final (-v_r) (v_r) y final (-v_r) (v_r) z final 0 $l boundary p p p
Why is it necessary to change the box dimensions? In fact, if you strain the simulation box, the original lattice may be unchanged, and so the create_atoms command will generate a new crystal region that is not matched to the original crystal. That may be the problem.
Aidan
Dear Aidan,
Thanks very much for your reply!
The induced atom by deleting one atom and then creating one atom, sometimes do overlap with the existing atoms in my simulations. But I introduced the pka at the middle of the axis, defects appeared at this place are expected. What confused me a lot is the defects at the ends of my cylinder, which can be seen from the attachments in my first email. There are far away from the middle and I think they should not be influenced.
I want to create pka at the free surface of my cylinder, but it seems after relaxation the position of free boundaries will slightly change. And it make it is difficult to define a region which contains ONLY one atom. But I will have a try.
I use change_box because one problem of creating the pka at the relaxed free surface. I posted my questions about this before on mailing list. https://sourceforge.net/p/lammps/mailman/lammps-users/thread/544E006D.3030407%40gmail.com/#msg32975879 and got my solution from the help of Steve.
I am running some extra similar simulations that pka induced in the center of cylinder, in which I do not need change_box. I am tring to see whether change_box makes my problem.
Okay, let us know what you find out.
Dear Aidan,
I think change_box in my scripts caused the problem. I have tried several runs and none of them showed the same kind of defects at the ends of my cylinder. In all these simulations, previous change_box command is commented. And the pka is generated by assigning velocity to an existing atom, which is done according to your advice.
I may need to do more similar runs to validate my conclusion and trying to find out other possible problem of my scipts.
I still have one question about generating the pka. In order to separate it from this question since it is a totally different one. I’d like to post a new question. I am looking forward to your and others’ help.
Thank you very much for your help!