Free energy calculation with pGFN-FF

Dear prof. Gale, sorry to bother you.

I was searching on how to calculate the free energies with GULP and I found out this tutorial: https://www.epj-conferences.org/articles/epjconf/pdf/2011/04/epjconf_ft2m2010_03005.pdf

I tried the two options (a) grad conp prop phon or b)opti free zsisa) you mentioned there, but I am using the pGFN-FF model.
Consequently, I could not run b) calculation, as the free method is not implemented for this model. For a), the calculation runs only if I don’t use the shrink keyword, which I understand it is needed so we can integrate across the Brillouin zone to obtain a converged free energy difference.

I am interested on calculating these free
energy differences for different phases of a framework. Is there any other way I can properly calculate the free energy using pGFN-FF in GULP?

Explicit k points aren’t implemented for pGFN-FF yet, but there are 2 ways you can compute phonon frequencies sampled at different k points to get the free energy:

  1. Provided you’re using version 6.2 onwards then the code will automatically create a supercell to determine the real space force constants and then phase them to get the right k points & so shrink should work.
  2. There is a direct equivalence between k points are supercells & so you can build larger supercells until you converge the free energy per formula unit & so this avoids the need to go away from the gamma point.
1 Like