SW Potential for GaNO

I want to make a sw potential file for GaNO, I am very puzzled on how to even start, any suggestions ? Thank you !

Are you asking about how to translate parameters from a publication into a potential file or about how to derive parameters from scratch?

Sorry, the first one, thank you !

Then I suggest a strategy with the following three steps:

  • carefully study the pair style sw documentation. It has the functional form of the potentials and there you can see how LAMMPS names the different parameters. It also has a section describing in detail the format of the potential file. Make sure that the functional form in LAMMPS is exactly the same as used in the publication (some people use modified Stillinger-Weber potential functions, then you have a problem and may need to do some C++ programming first).
  • look up and study the publication for a Stillinger-Weber potential for which LAMMPS already has a potential file included. Try to understand how the author of the file translated the parameters from the publication to the ones used in LAMMPS. Best do this first for a parameter set with just one element and then for a parameter set for multiple elements (which is significantly more complex).
  • first create a potential file for just one element for which a potential file for LAMMPS already exists (doesn’t have to be Stillinger-Weber and doesn’t have to be bundled with LAMMPS) and then run some simple tests in order to see if the two potential file will produce similar results. Keep in mind that absolute energies are meaningless, only energy differences and optimized geometries. Of course, then you need to graduate to creating the final multi-element file and then reproduce the results from the publication.

Beyond that, you should probably also look for assistance by a more experienced member of your research group or ask your adviser to try and find a suitable collaborator that can assist with this. Don’t expect more help from this group beyond such general advice. That is, unless you get very lucky and somebody here reading this does have the necessary experience and is interested in translating the same publication into a usable file and you can collaborate on the process of translating and testing.

Thank you !