This might be very silly but here goes. I can’t wrap my head around how OpenKIM is useful. Here’s a GaAs potential tested:
OpenKIM · Tersoff LAMMPS AlbeNordlundNord 2002 GaAs MO_799020228312_004 MO_799020228312 · Interatomic Potentials and Force Fields
But the data is so confusing. GaAs has zincblende structure. What’s the point of doing individual tests on Ga and As as “species” seperately?
Hi Sandip,
I think there are valid justifications for evaluating single-element properties of multi-element potentials. Perhaps not GaAs specifically, but something like an FeC potential for steel should be able to accurately describe iron.
However, you are absolutely correct that this is a limitation. We are currently actively addressing this limitation with our new project “Crystal Genome” that tests arbitrary, multi-species crystals. You can access the new interface for viewing these results from the “search for predictions” button on the new front page. We are currently running tests for 32,000 different crystals, with plans to add hundreds or thousands or even millions (we can trivially add any crystal structure, so we can scrape repositories like AFLOW and Materials Project for structures). GaAs hasn’t been run yet (the infrastructure is all in place, it just takes time to automatically run the computations), but returning to the FeC example, you can see the results here:
OpenKIM · Search by Compound · Interatomic Potentials and Force Fields. From there you can filter the results by different crystal structures and go to a page with results for each crystal.