Dear all,
I am trying to simulate an alloy containing 6 elements (Fe, Au, B, Co, Ni, Zr ). Clearly, its potential file is not in literature nor in lammps library. What are the ways of generating an EAM potential file for such a system? Should I conduct some experiment or it must be done with DFT and first-principle calculations?
Any advice and suggestions will be greatly appreciated.
Kind regards,
Kim
.
Dear all,
I am trying to simulate an alloy containing 6 elements (Fe, Au, B, Co, Ni,
Zr ). Clearly, its potential file is not in literature nor in lammps
library. What are the ways of generating an EAM potential file for such a
system? Should I conduct some experiment or it must be done with DFT and
first-principle calculations?
i do not know how to parameterize for EAM, but i would look up how
others have done their parameterization in the published literature. i
would expect there are some papers describing potentials for alloys.
if not for all of the elements, at least for some, and then it becomes
just a matter of taking what is already there and adding to it in the
same way.
if my understanding of the documentation is correct, than you may
start by using the plain eam pair style and combine multiple single
element potential files in the suitable format, and LAMMPS should
create a mixed potential for you. that may not be optimal for your
needs, but might give you a start.
of course, validation for such a complex compound is going to be a major pain...
axel.
Thank you for your reply.
As you suggested if I want to combine multiple single element potential files, is it done by pair_style hybrid command and then adding other potential files or I’m wrong?
Kind regards,
Ehsan
Thank you for your reply.
As you suggested if I want to combine multiple single element potential
files, is it done by pair_style hybrid command and then adding other
potential files or I'm wrong?
you are wrong! that is *not* what i suggested. please re-read my
e-mail and do it more carefully. also study the documentation for pair
style eam (not eam/fs or eam/alloy) *very* carefully.
axel.
...and let me emphasize, that using pair style hybrid is a
particularly bad idea for multiple EAM potentials, since it will not
share the density contribution for computing the embedding term across
the multiple hybrid potential partitions. that is why you need "real"
mixing of potentials within the pair style.
axel
Hi, you should take a look at this discussion on reasearchgate:
https://www.researchgate.net/post/Alternatives_to_GULP_for_fitting_potentials/amp
You can also look at the eam_database and eam_generate
tools in the LAMMPS tools dir. If all your elements
are not in their lists, you can contact the authors.
With B, you might need to be using MEAM.
Steve