Local cluster expansion?

Alloy Theoretic Automated Toolkit (ATAT)
1

Dear Prof. Alex Van de Walle,

Is there a way to perform local cluster expansion in ATAT, e.g. for vacancies in alloy? I would greatly appreciate your help.

Thank you,
Jingyang Wang

2

Hi,

I have an idea on the calculation of correlations for local cluster expasion.

  1. fix the spin for vacancy as 1.
  2. now this is the tricky part, we can still use the clusters.out, but you need to realize the clusters for local cluster expansion is different from those for the normal CE. You need to reduce the clusters to their subsets, meaning: the 2-site ones reduce to 1-site clusters, 3-site ones reduce to 2-site clusters. The reduced site is the one occupied by the vacancy. (the tricky part of this step is that, it is possible that not all the subsets of one multi-site cluster actually belong to one same smaller cluster. A dirty way, I think, is treating all the subsets as distinct clusters, the fitting would work anyway. You can find and sort out the real distinct clusters by comparing the correlations)
  3. okay, now you can map the energies on this new cluster set.

But I have not really tested this. Hope it could be a useful idea.

Yi Wang
Nanjing University of Sci. & Tech.

3

Dear Prof. Wang,

Thank you for your detailed explanation! It seems like a viable approach; I would surely like to try it out.

Best regards,
Jingyang Wang

4

Thank you, but I am still a Ph.D student.
Wish I get luck with your grant. :stuck_out_tongue: