Trends in important clusters

Alloy Theoretic Automated Toolkit (ATAT)
1

Hello Axel,

Recently I’ve been studying the phase behavior of Mg-Rare Earth binaries and ternaries, and have noticed trends in the cluster expansions for these materials. For instance, for a number of Mg-RE binaries, the 2nd nearest neighbor interaction on the fcc lattice always shows a negative eci, the magnitude of which significantly dwarfs other present eci. Other trends exist for bcc and hcp lattices as well. I’m wondering if there is actual physical significance to these trends. They seem to be so uniform across different Mg-RE binaries that I feel there must be something deeper than just "it’s what fits it well," but I don’t think I have enough intuition concerning ECI to make that statement confidently. Do you have any thoughts about interrogating the physics of these trends on a deeper level? Am I just digging too far down the rabbit hole?

Thank you,
Adam

2

This is interesting. Two possible origins to look at:
-about ~20 years ago there were a lot of papers about computing the moments of the electronic DOS (which are related to correlations).
-elastic relaxations: perhaps having "big" atom at one site makes it easier to fit a similarly big atom at a 2nd nearest neighbor site?