How to Define a generic LJ 12/10 potential?

LAMMPS LAMMPS Mailing List Mirror
1

Hi all,

I’m somewhat new to LAMMPS, but I want to use it to perform some NPT simulations of the ionic solid ammonium nitrate. I’m trying to reproduce some of the results in the attached paper. My first issue is that the authors have defined the (H3N-H….O) H-bond interactions using a simple standard LJ 12-10 potential(see page 9153, third page), but I cannot seem to locate that particular pair style in the LAMMPS manual. The LAMMPS has the hbond/dreiding/lj specifically for H-bonds, but it has a cosine theta term involved. I would prefer to set up my H-bonding interactions in the same manner that was done in the paper, especially since there are parameters for the O…H interactions defined, and I would prefer using those for consistency. So is there a way to define a generic LJ 12/10 potential(if I have not over-looked it already)?

Best,

Jeffrey

Pub 218_AN Melting.pdf (696 KB)

2

Hi, Jeffrey:

LAMMPS doesn’t have a 12/10 directly available. However, you can easily modify any of the existing LJ potentials to change it from a 12/6 to a 12/10. You’ll just need to change the appropriate B and D coefficients and multiplicative factors in a few of the routines, and make sure the exponentiations of the distance are done correctly.

But this should be something you could easily do without a lot of trouble.

—AEI

3

Thanks Ahmed,

Is this something that I would have to alter in the source code and then recompile?

Jeffrey

4

Hi all,

I'm somewhat new to LAMMPS, but I want to use it to perform some NPT
simulations of the ionic solid ammonium nitrate. I'm trying to reproduce
some of the results in the attached paper. My first issue is that the
authors have defined the (H3N-H….O) H-bond interactions using a simple
standard LJ 12-10 potential(see page 9153, third page), but I cannot seem to
locate that particular pair style in the LAMMPS manual. The LAMMPS has the
hbond/dreiding/lj specifically for H-bonds, but it has a cosine theta term
involved. I would prefer to set up my H-bonding interactions in the same
manner that was done in the paper, especially since there are parameters for
the O..H interactions defined, and I would prefer using those for
consistency. So is there a way to define a generic LJ 12/10 potential(if I
have not over-looked it already)?

a) you can always use pair style table and create a tabulated
potential with an external program/script (python/matlab/awk/fortran).
b) there is the mie pair style which is a generalized LJ pair style
(it is rather slow, though)
.http://lammps.sandia.gov/doc/pair_mie.html

axel.

5

Ok great,
I have options now. I may try these out as well.

Jeffrey