Using MTP forcefield that has more than required elements

[pshinyeong]

Hi all,

I was trying to generate GeTe amorphous structure using melt quench method.

This is a general question, is it safe to use MTP forcefield that has more elements than my required configuration?

More specifically, is it safe to use the built-in MTP for MD/Geo-opt of crystal GeTe which is QuantumATK_MTP_GST_2022_12 (Ge, Sb, Te)?

Thanks.

[AsifShah]

Hi,
One approach is to test and verify it for your application as:
1. Optimize your material using this MTP and DFT and compare bond lengths, angles etc.
2. Run a small one atom movement NEB using MTP and DFT and compare results.

After verification, if you are using it for molecular dynamics make sure you verify/optimize all states of interest with DFT to check consistency.

[Anders Blom]

Yes, that should be safe with any MTP. The question on transferability of forcefields is more about which kinds of systems they are fitted for, rather than the exact amount of elements. You could for example have a forcefield for carbon, but it's only created for diamond, so it would not work for graphene. In the MTP case for GeTe, it's fitted and tested for both crystalline and amorphous materials (mainly 3D), so for your purposes it should work fine.

[pshinyeong]

Thanks all