Dear Lammps-users:
I encounter a problem in using potentials. I used to use Si.tersoff for my simulation, and recently i want to research on the C interactons, I try to use SiC.tersoff and succeed. However, the silicon atoms moves not exactlly as the Si.tersoff. So I want to combine the two tersoff potential together.
Here are my old command:
pair_style hybrid tersoff morse 9.025
pair_coeff * * tersoff SiC.tersoff Si Si Si C C C
pair_coeff 13 46 morse 0.423 4.6487 1.9475
And I try to use two tersoffs:
pair_style hybrid tersoff morse 9.025
pair_coeff * * tersoff SiC.tersoff Null Null Null C C C
pair_coeff * * tersoff Si.tersoff Si Si Si Null Null Null
pair_coeff 13 46 morse 0.423 4.6487 1.9475
But it reports to be wrong.
How can I use two tersoff potential together?
Thank you very much!
Dear Lammps-users:
I encounter a problem in using potentials. I used to use Si.tersoff for
my simulation, and recently i want to research on the C interactons, I try
to use SiC.tersoff and succeed. However, the silicon atoms moves not
exactlly as the Si.tersoff. So I want to combine the two tersoff potential
together.
this is a very bad idea. even if you would hack this so that it would work
(which is technically possible by modifying the source), you are now
having only a mechanical coupling between Si and C. that means, you
have no more many-body interactions between them.
if there are differences between the two potential parameterizations,
you should first check the original literature, whether this is intentional,
and how those potentials were validated. you might be able to improve
the Si behavior by adjusting the parameterization, but then the question
is how this would change the mixed Si-C interactions.
finally, the question is: are those differences actually significant,
or rather withing the systematic error margin of the model.
Dear Axel Kohlmeyer：
Thank you for answering my letter. I have check the potential value of Si.tersoff and SiC.tersoff, and the description for Si Si Si value is different, and my research in the past all based on the Si Si Si value of Si.tersoff, that will mean i have to repeat on the work i have done the get new values from the SI Si Si values of SiC.tersoff. That’s why i want to combine the use of two tersoff potentiaol together. Will it have a way to do that?
You can't use pair hybrid in such a way. The only
thing you could do is to manually create a new Tersoff
potential file that would work for the mixture system that
has the interactions you want for each atom type. There
is no reason that the SiC tersoff file could not list
interactions for 3 Si atoms, that are identical to what
the Si tersoff file computes.
Dear Axel Kohlmeyer：
Thank you for answering my letter. I have check the potential value of
Si.tersoff and SiC.tersoff, and the description for Si Si Si value is
different, and my research in the past all based on the Si Si Si value of
Si.tersoff, that will mean i have to repeat on the work i have done the get
dear susie,
the question you have to ask yourself is where does this difference
come from (i.e. read the papers) and then figure out, whether this
difference matters or not for your research.
what you have here is a _parameterization_, i.e. a model for your
system. this _cannot_ be the "one true way" to represent silicone
with this kind of potential.
if it was, there would be a lot of upset people using all kinds
of quantum chemistry methods and spending a much larger
amount of computer time (and personal time) on their work.
new values from the SI Si Si values of SiC.tersoff. That's why i want to
combine the use of two tersoff potentiaol together. Will it have a way to do
that?
technically, you can do whatever you want. practically, you have
to ask yourself, if the result would be meaningful and there the
common understanding is, that you cannot mix and match different
parameterizations. each set of parameters has to be a compromise
between the degrees of correctness in different properties of your
material. just changing some of them would throw off the balance
and can be worse than just using the SiC.tersoff file alone.