first off, please always reply to the mailing list as well, not just
individual people.
Dear Axel,
Many thanks for your response. You mean that I should modify the
coefficients in the potential file and then use simple command :
pair_style:tersoff and Pair_coeff: .....Tersoff C ?
no.
Do you know any other way that we can simulate N-doped Graphene sheet in
more accurate way?
your approach is backwards! what tells you that a tersoff potential is
a good approximation for this system in the first place? where would
you get the C-N mixed tersoff parameters from? the tersoff description
does not consider charges (at least not suitable for a
non-stoichiometric system) so how would you include that effect in
your description?
thus to find out what *is* an accurate (enough) model, you first have
to study the published literature in order find out, what other people
have done for these purposes and try to assess whether the accuracy
they achieve is sufficient for your needs. as a rule of the thumb, you
can assume that the level of accuracy is correlated with the level of
detail of the model and thus the computational cost. thus you can
choose on a range between a simple pairwise additive description
(which is assumes a fixed bond topology) to a DFT level description
(which is also beyond the scope of LAMMPS).
axel.
Many thanks for your response. I read a paper in which the tersoff potential was implemented. In this paper, the coefficients of tersoff potential were introduced so that I got advantage of them to examine my C-N model.
I didn’t find any related paper that tells me about Nitrogen doping in Graphene sheet. It is my pleasure if you can help me in this way and which potential should I use or which command will be helpful in LAMMPS simulation for this goal.
Regards,
Many thanks for your response. I read a paper in which the tersoff potential
was implemented. In this paper, the coefficients of tersoff potential were
introduced so that I got advantage of them to examine my C-N model.
I didn't find any related paper that tells me about Nitrogen doping in
Graphene sheet. It is my pleasure if you can help me in this way and which
potential should I use or which command will be helpful in LAMMPS simulation
for this goal.
sorry, but this is your problem and your area of interest and not mine.
what you have to do, i already outlined in my previous e-mail: you
have to determine, what is a good model for your system and the (and
only then) we can discuss how to realize this with LAMMPS. if you have
seen a publication with a suitable parameterization, then you have to
translate the parameters into a format that LAMMPS can read. as i
wrote before, you cannot simply mix and match some parameters with
some others. if you don't understand this, you would be better off
studying some more how classical models are parameterized and practice
how to run test calculations to reproduce results from publications.
it is a very bad idea to simply start with something complex without
knowing what is going on. how are you ever going to tell, whether you
made a mistake from whether the parameters don't work properly? in
classical models, the transferability of per element parameters is
quite limited (in contrast to ab inito models).
axel.
oh, and the questions you are asking are something that you should
first talk about with your adviser. as the name says, the task of an
adviser is to give advice and you obviously need some and particularly
some rather general advice on how to properly prepare and execute
classical simulations. so take advantage of that. a mailing list is a
very bad replacement for this. mailing list are good for solving
specific issues, particularly technical ones that specific to the
software.
axel.
Special thanks for the useful points you have mentioned. Sure. I will try to find a proper model and then I will try to implement this model in LAMMPS. If I faced any problem, it is my pleasure to share it here and discuss with other peers.
Wish you all the best,
Shahriar