Partial density of states in random solid solutions

Question

This question comes as a follow-up to https://mattermodeling.stackexchange.com/a/2214/116; nonetheless, it is relevant on its own as well.

To designate a particular crystal structure to a random solid solution (rss, in this question) means assigning an ensemble of all possible random states of this solution with that specific crystal structure. For instance, an FCC rss has complete FCC symmetry only in its idealized theoretical random state. In contrast, in reality, any one of the randomly possible configurations of this solution will have a symmetry that is lesser than FCC.

For an rss, every atom has a different local chemical environment that means they each would have a different electronic density of states. Given that, does a partial density of states (PDoS) matter?

Let's say for an 8 atom structure of an AB type rss, PDoS for each of the 8 atoms will be different and make sense. Does a collective PDoS for the 4 A atoms or the 4 B atoms make sense, though?

Answer 1

An 8 atom structure can have a PDoS of each atom assigned to orbitals, but your question is more about does the collective PDoS of those 4 atoms which are different combinations to form anything interesting.

A PDoS is useful for assigning which states come from what orbitals or atoms, so I would say that this probably depends on your particular system. Try to plot it and see what you get, the full DoS can be seen as the collective PDoS of all atoms, so it makes sense to split these by elements, orbitals, spin, etc. Your system will determine if anything useful can be determined.