Is it possible to fix cluster expansion terms, obtained from

Alloy Theoretic Automated Toolkit (ATAT)
1

Dear ATAT experts,

Assume I have the following FCC A_xB_yC_z alloy, where x+y+z=1. Let’s suppose I’ve obtained a good cluster expansion for the binaries A_{1-x}B_x, A_{1-x}C_x, B_{1-x}C_x also as FCC. I would think that since in A_xB_yC_z, the cluster expansion terms involving AB, AC, or BC should behave very similar to those of the binaries -below a certain length range-, one could fix these terms and focus to calculate those that include ABC interactions. Perhaps only allowing that AB, AC, and BC interact at distances not contemplated in the lengths used in the binaries.

I know there is an input file, nbclusters.in, that allows to select which clusters to include, but I don’t think one can assign values to these clusters and keep them fixed. Can we?

Indeed, the alloy I am targeting has more then three elements.

Thanks for reading and for your support.

2

Sorry, this is neither implemented nor very straightforward to do so.
The way to define ternary effective interactions that are symmetric in the three species do not nicely separate into binary terms and ternary terms. It is not even clear that an effective interaction in a ternary and in a nested binary system should be equal (one averages over all ternary environments of the cluster considered and the other over all binary environments of the cluster considered).

On the positive side, it should not be difficult to import your binary structural energy data and then start a ternary cluster expansion from there.