KIM model driver for Buckingham potentials

OpenKIM
1

I am interested in models for metal oxides which quite often are modelled with a Buckingham potential plus some long-range electrostatic interactions (for instance see http://dx.doi.org/10.1080/14786430412331326290). My primary simulation engine is LAMMPS. Is there a suitable KIM model driver?

2

Hello Evangelos,

Thanks for contacting us about this. There is not currently a model driver in KIM that will work for your system. However, depending on what you would like to use KIM for, we may be able to hep you.

If you are interested in archiving or distributing a model you have developed, we can work with you to create what we call a “simulator model” in KIM. This would allow anyone with lammps and KIM installed on their system to use your model as you intended it. It would also provide you (and anyone else) with a citable DOI for the model. A simulator model would also be useful even if you are using someone else’s model, but you want to clearly present the parameters you used in a reproducible way, thus allowing your work to be more easily repeatable.

If you are looking for other models to study metal oxide systems, this approach wouldn’t help you.

What did you have in mind?
Ron

3

Hello Ron

Thanks for your reply. I am interested in having a “library” of potentials for some metal & metal oxides systems which I and my collaborators can use in LAMMPS. I have found OpenKIM very helpful so far especially for the case of metals.

Since there is no suitable KIM model driver, I am definitely interested in working with you to develop a “simulator model” in KIM. I had only a very quick look at the website. Could you please share with me any document / info regarding the preparation of such a KIM model?

Evangelos

4

Hi,

To get started on this, I think if would be good if you could me a sample lammps input script that you use for one of your simulations, as well as any auxiliary files that I would need for testing (parameters for the potentials, coordinates of atoms, etc). Something small that can run quickly would be best.

5

Hi

I attach a LAMMPS input file and the required data file with the coordinates of the atoms. The system is crystal NiO with no defects and the potential has been published in https://www.tandfonline.com/doi/abs/10.1080/14786430412331326290 (DOI 10.1080/14786430412331326290)

Please let me know if you have any questions
Evangelos

in.NiO (684 Bytes)

NiO.data (768 Bytes)

6

This seems straightforward. I’ll discuss it with the team tomorrow at our weekly meeting and get back to you soon. I think what you have sent is clear and all we need for now.

Ron

7

Thanks for looking into this Ron.
If I can assist with some coding please let me know.

8

Hi Evangelos,

A couple updates: We are working on a robust implementation of the buckingham SM. I think it should be ready soon. We want to make it general enough so that it doesn’t depend on a user implementing the same numbering of species as assumed in the model parameters, so it is taking a little longer.

Also, we accepted your Cr ADP SM after a few minor tweaks to the input (I edited the description a bit, gave it an extended kim ID according to our conventions and used our preferred format for LAMMPS versions (DD Mth YYYY).

9

Hi Ron,

thanks for taking care of the Cr ADP SM as well as implementing a Buckingham SM. I have just checked the selection form for the "KIM Model Driver " and I can spot the option for “buck/coul/long” ! Your and OpenKIM’s team response has been really fast.

10

No problem. Please let us know if you have other models to add or if you have questions. We are still working on the best way to implement the SM so that it is general. It is certainly possible, but we are debating the best way so that what we do is robust for future extensions. we’ll keep you posted.

ron

11

Is it possible to add an acknowledgement (funding project) in the submitted potentials?

Thanks
Evangelos

12

Hi Evangelos,

It is not possible at the moment, but it's a good idea. I'm surprised it hasn't come up before. We'll look into adding a funding field to all content on our site (which will also be included in associated DOI records). Assuming there are no complications, it shouldn't take too long to set this up.

Best,

Ellad Tadmor