Dear all,
There is a command( pairmodify tail yes) which can calculate long-range correction for LJ potential but it does not work EAM potential. Does anyone know how to calculate tail correction for EAM potentials?
Thanks,
Ramin
There is no tail correction for EAM. It goes smoothly
to 0.0 at the cutoff.
Steve
Thanks Steve, I wrote a code which coupled with lammps to perform Gibbs ensemble hybrid Monte Carlo and it works for lj system and for lj system I had to include energy beyond the cut off( tail correction) now I want to simulate liquid alloys and I’m wondering how I can calculate energy beyond the cutoff for eam potential. Is that possible?
Thanks,
Thanks Steve, I wrote a code which coupled with lammps to perform Gibbs
ensemble hybrid Monte Carlo and it works for lj system and for lj system I
had to include energy beyond the cut off( tail correction) now I want to
simulate liquid alloys and I'm wondering how I can calculate energy beyond
the cutoff for eam potential. Is that possible?
let me rephrase what steve wrote: if a potential is going smoothly to zero
at the cutoff, there cannot be a tail correction like it is computed for
lj/cut and alike. the tail correction is computing the part of a potential
that is truncated through the cutoff.
now, if you want to compute a long(er)-range interaction contribution for
EAM, you have a completely new model.
axel.
Thank you axel, so if I want to include longer range in my calculation I need to redesign EAM potential or there is another way to calculate that part of interaction without changing the current potential?
And my second question is that my EAM potential is working for regular MD and I see a phase separation in a box, so can I conclude that the interaction beyond the cutoff has no physical meaning in my system When it’s working in MD? Or they can be a game changer when it comes to Gibbs ensemble method?
Thanks,
Ramin
Thank you axel, so if I want to include longer range in my calculation I
need to redesign EAM potential or there is another way to calculate that
part of interaction without changing the current potential?
this is not my area of research. i have explained what i know.
And my second question is that my EAM potential is working for regular MD
and I see a phase separation in a box, so can I conclude that the
interaction beyond the cutoff has no physical meaning in my system When
it's working in MD? Or they can be a game changer when it comes to Gibbs
ensemble method?
when a model is not considering interactions beyond a cutoff, that doesn't
mean, there are none. it simply means that they are considered negligible
or that they cancel out. for homogeneous(!) liquid metals, for example,
that is a good approximation to assume that interactions cancel beyond a
certain distance.
since gibbs ensemble is not my area of research i cannot tell you how this
translates. in fact, your line of questioning suggests, that you need to
validate whether your use of tail corrections for lj/cut interactions is
actually giving you an improvement for a correct physical reason or just by
accident. once you have explored and fully(!) understood that, it should be
more clear how to address issues that you are seeing.
another question that comes to my mind is: does your implementation
consider that EAM interactions are not pairwise additive?
in conclusion, i think you need to start a conversation with your
adviser/supervisor and dig deeper into the literature. this mailing list is
focused on technical issues around LAMMPS. we usually don't have the time
(or interest?) to provide advice on planning research projects.
axel.
Actually tail correction for lj systems is in literature and without that I was getting false density and pressure in the system for different potential parameter so it has a physical reason.
No, in my implementation I just get the total potential energy from Lammps whenever I need it in my Monte Carlo calculation and assume it is a configurational energy and treat it like the potential energy plus tail correction energy in lj system. Is that wrong?
Thank you axel.
Actually tail correction for lj systems is in literature and without that
I was getting false density and pressure in the system for different
potential parameter so it has a physical reason.
you are not paying attention (and there are few things that irritate me
more than people that are not paying attention). the fact that you get the
desired result is - by itself - in no way conclusive proof, that it is
physically meaningful. ever heard the term "cancellation of errors"? so it
is quite possible, that you are getting the "right" answer for the "wrong"
reason. now, which is which, i cannot tell, since - as i already mentioned
- your area of research is not mine, so i don't claim any expertise in it.
all i've explained was the physical reasoning for what is computed with the
tail correction.
No, in my implementation I just get the total potential energy from Lammps
whenever I need it in my Monte Carlo calculation and assume it is a
configurational energy and treat it like the potential energy plus tail
correction energy in lj system. Is that wrong?
same as before. this is not my area of expertise. LAMMPS will give you
what you ask it to do. the rest is not my concern.
axel.
There is no energy for EAM beyond the cutoff. It is not
like LJ in that respect. The energy goes to 0.0 at the
cutoff.
Steve