Dear All,
I use EAM/fs potential function. I would like to ask whether it is possible to use different cutoff radius in different conditions. If it is, how can I do this? Should I use the " if " command ?
Regards,
Nazanin
Dear All,
I use EAM/fs potential function. I would like to ask whether it is possible
to use different cutoff radius in different conditions. If it is, how can I
cutoffs for EAM are embedded in the EAM potential files.
since those are text files, you can edit them. whether
that is a good idea or not, is a different question. you
most certainly will have to test for correctness after
*any* manipulation of a potential file.
what do you mean by "different conditions"?
do this? Should I use the " if " command ?
that by itself won't do it. please study the
documentation to see how the if command
works. please keep in mind that all lammps
commands are executed immediately. so it
may not work as you might assume.
axel.
Dear Axel,
Thanks for your reply. I have already tried using different cutoffs. They worked correctly but gave me different results which could be because of the cons of the potential function.
Regarding to my previous email, a reference proposed rc=4.2 Å as an optimal choice. With this value, first and second nearest neighbors (NN ) in NiTi are taken into account in the calculation but since it lies closely below the distance of the third nearest neighbor shell, a smooth cut-off behavior has to be ensured. So I have to use two different cutoffs in my simulation. Assuming Ɛ=0.2 Å, if (r < rc- Ɛ ) I want to use the coefficient file of rc=4.2 and if (rc- Ɛ < r < rc) use the file of rc=4.4.
I would like to know if it is possible to use two different potential coefficient files during a simulation. If so, how can I do that?
Regards,
Nazanin
Dear Axel,
Thanks for your reply. I have already tried using different cutoffs. They
worked correctly but gave me different results which could be because of the
cons of the potential function.
Regarding to my previous email, a reference proposed rc=4.2 Å as an
optimal choice. With this value, first and second nearest neighbors (NN ) in
NiTi are taken into account in the calculation but since it lies closely
below the distance of the third nearest neighbor shell, a smooth cut-off
behavior has to be ensured. So I have to use two different cutoffs in my
simulation. Assuming Ɛ=0.2 Å, if (r < rc- Ɛ ) I want to use the
coefficient file of rc=4.2 and if (rc- Ɛ < r < rc) use the file of rc=4.4.
there is nothing smooth about this. in fact it is almost guaranteed
that you have bad energy conservation with a scheme like this.
I would like to know if it is possible to use two different potential
coefficient files during a simulation. If so, how can I do that?
no, not without writing c++ code. what you describe is an entirely
new pair style (and an ill conceived one to boot in my personal opinion).
whether
that is a good idea or not, is a different question.
And I'll answer it - it's most definitely a bad idea to
just change the cutoff for EAM All the functions tabulated
in the file are fit for that cutoff. You'd need to
re-parameterize the entire potential if you change the
cutoff.
Steve
thanks for your answer, steve and axel.
I always appreciate your kind support for the lammps users.
Best,
Nazanin