Hello to all! I am performing radiation damage study for bulk tungsten using lammps. Since we have to spline a long-range pair potential (my case, EAM FS) with a short-range repulsive ZBL potential, I am having a dilemma on how to specify the switching function and global cut off for ZBL potential. Like is there any thumb rule to specify it? Do we need to specify according to first and second nearest neighboring separation as mentioned in EAM file database.
Check out the tail end of the discussion here where this topic came up.
In general, searching the LAMMPS mailing list and forum archives for previous discussions can save you asking questions that have been asked before. A smart search on google will usually do the trick.
To follow up, the answer to how careful you have to be is likely dependent on whether you are using a hybrid or hybrid overlay and whether the zbl is for all atoms in the system or just interactions with incoming projectiles and the substrate.
For example, if zbl is just for incoming projectile and substrate then these choices are likely not as important and in my experience the rinner and router suggested in the manual will probably do you fine. There has been some good work on W radiation the past that seems to use this approach.
However, if you are overlaying zbl with another potential for bonds within the substrate this is where you need to be careful as zbl can add extra energy to a bonded pair that you don’t want. In that case see the discussion Axel mentioned for some more detail.
Thank you for your suggestion. I went through the discussion and got many points cleared. I am overlaying the zbl potential and want the interactions with all atoms. In the ReaxFF related discussion, I went through one of your queries about if there is a possibility that the zbl interactions approach to zero at long-range distances. Is it true?
If you plot out zbl similar to what was shown in the other thread you will see that it will eventually trend to zero. The problem that I found, and that you will likely run into, is that it will add energy to close approaches when you in fact don’t want there to be additional energy. In my case my substrate was properly developed and equilibrated (with no spikes), when I overlayed the zbl it called some of the bonds close approaches and added energy to them that caused an undesirable energy spike. This is why there was a suggestion to make sure it is ramped to zero before the length of the smallest bond. If this does not happen to you then I don’t see a major issue with letting the cutoffs be large. You can test it by developing your substrate and equilibrating with and without the zbl. If zbl leads to a large spike in energy you may have to do something about it.
In my case I am able to just use zbl for projectile and substrate. I am not sure what you are simulating but I would really think if you need zbl between substrate atoms.
I am trying to study radiation damage in single-crystal tungsten and the primary knock-on atom is a centrally chosen tungsten atom. Thank you for your prompt and valuable suggestion.