Reading through some first principles papers concerning iron and nickel, something lead me to ask me this question :

"What is the definition of the solubility energy of an impurity in a crystal phase ?

Shall we call solubility energy the energy given by (for Fe case for example) :

E_sol(X)=E_0(N.Fe,X) - N*µ_ref(Fe) - µ_ref(X)

Without according a thought of the site where the impurity is ?

Because in several papers, solubility energies is given for sites which turn out to be transition sites at the end.

More i think about it and more it appears to me that the expression "solubility energy" should be used for stable sites only (which imply to verify the band structure of the system with the defect).

I'm curious to read what you think about it.