Hi All,
I have one doubt. I am creating an edge dislocation in fcc, taking data from MALAB script. I am putting one dislocation, so i need to have periodicity in both Burger’s and Line direction.i.e x(My Burger’s Direction) and z(line direction) direction lengths must be integral multiples of burgers vector. My code gives me upper and lower bounds of the domain in x, y and z; but a small difference exists to have exact periodicity. For that, i am adding the required distance in the upper and lower bounds of the length’s. i.e manually in the read_data file created by MATLAB. So will i have a little sort of a vacuum gap or something where the atoms are not there, thus affecting my simulation run later when my dislocation will cross the boundary and appear on the other side?? Or is this correct that it is not necessary to have the atoms filled in the box right till the boundary…???
Regards,
Sagar
Hi All,
I have one doubt. I am creating an edge dislocation in fcc, taking data from
... and i have many "doubts", but i suspect what you really have is a
"question", right? 
MALAB script. I am putting one dislocation, so i need to have periodicity in
both Burger's and Line direction.i.e x(My Burger's Direction) and z(line
direction) direction lengths must be integral multiples of burgers vector.
My code gives me upper and lower bounds of the domain in x, y and z; but a
small difference exists to have exact periodicity. For that, i am adding the
required distance in the upper and lower bounds of the length's. i.e
manually in the read_data file created by MATLAB. So will i have a little
sort of a vacuum gap or something where the atoms are not there, thus
affecting my simulation run later when my dislocation will cross the
boundary and appear on the other side?? Or is this correct that it is not
necessary to have the atoms filled in the box right till the boundary..???
there is not "right" or "wrong" in these cases. LAMMPS will do the
simulation of the system that you feed it. whether it actually is the
system that you *intend* to simulation, is up to you to determine. the
easiest way is to read your data into a visualization program that can
visualize periodic copies of your system and you can judge by
yourself, whether your input data is creating the system that you want
to simulate. similarly, you can use the replicate command in lammps in
combination with the dump image command to create snapshots. if your
system is not the (perfect) periodic continuation that you want, then
you need to fix your input data. since you are writing your own code
to generate the input, you have to debug it by yourself.
axel.