lammps-users Digest, Vol 58, Issue 198


I am currently working on a Si-C-H version of REBO. It is parameterized and works for my version of REBO, but I have not ported it over to LAMMPS yet. I am hesitant to share the parameters just yet, as I am still ironing out a few details, and working on the paper. This has been a long tedious process, as parameters don’t just fit themselves, but I feel I am in the final stretch of testing and writing. Not many other potentials have been fit to AIREBO / second generation REBO. I believe there is a C-H-O potential out there, as well as the C-F potential you mention, but that is it.

A note of caution: You have to be pretty careful using other peoples parameters. I am not intimately familiar with the C-F parameters, but I can say from experience that you have to pay close attention to what is reported in the papers to make sure that there are no other little modifications required for the source code. For example, for my potential, I have replaced the bicubic spline for the Pij terms with tricubic splines to more accurately fit the ternary system I am modeling (among other things).

Point being, it may or may not be as simple as “plug in parameters and play”. Bond order potentials are fairly complex beasts with many parameters. One the immediate challenges you will find is that the LAMMPS version of the code takes the spline coefficients in as input, while the papers report only the nodal values for each spline knot. This means you need a code that can take knot values and calculate the spline coefficients for the bicubic and tricubic spline functions. Not an impossible task, but it does take a bit of monkeying around and writing your own spline function codes, and of course the format has to be correct.

You may wish to contact Susan Sinnott’s group at U. Florida, she may have a lammps version of the C-F code that I don’t know about. Or she may be able to point you in the direction of a different code with the C-F parameters. If you ask very nicely she may even let you have her version of the code. She is usually very helpful in that regard.

Good luck,

J. David Schall, Assistant Professor
Dept. of Mechanical Engineering, Oakland University
130 DHE, Rochester, MI 48309