Hi all-
I am running a simulation with a silica substrate that has bonds which cross through the x and y boundaries (i.e. an atom near xhi is bonded to an atom near xlo), but I get the warning for inconsistent image flags. I believe that everything is fine with my starting configuration (in that if you correct for the boundary conditions the bonds are all of an acceptable length etc.), but I am wondering if this will cause some sort of problem with the dynamics of the simulation.
I have read some other e-mails in the mail list where people have gotten similar errors with replicate and I have seen that this has been said to be due to not initially having the correct image flags for a given molecule that spans a periodic boundary, which I understand. But I don’t see how I can fix that for a solid because if I start changing image flags then in reality, I would end up having to change all of them I think.
Anyway, I am hoping some one can assist me in understanding this.
Cheers-
Chris
You're right - you can't have fully consistent
image flags for a periodic solid that spans the box.
From Section errors of the manual:
Inconsistent image flags
The image flags for a pair on bonded atoms appear to be
inconsistent. Inconsistent means that when the coordinates of the two
atoms are unwrapped using the image flags, the two atoms are far
apart. Specifically they are further apart than half a periodic box
length. Or they are more than a box length apart in a non-periodic
dimension. This is usually due to the initial data file not having
correct image flags for the 2 atoms in a bond that straddles a
periodic boundary. They should be different by 1 in that case. This is
a warning because inconsistent image flags will not cause problems for
dynamics or most LAMMPS simulations. However they can cause problems
when such atoms are used with the fix rigid or replicate commands.
So image flags do not affect dynamics. But you can't replicate
the system via the replicate command. LAMMPS isn't smart enough
to figure out how to break/add bonds with inconsistent image flags.
Steve