partial site occupancy

Message

Lilia Woods · #1 Post by **Lilia Woods** » Wed Jun 05, 2013 4:23 pm

I am calculating the electronic structure of a crystal ABC4, which has 4A atoms, 4B atoms and 32C atoms. The C atoms have 1/2 site occupancy.

Does anybody know how to deal with partial site occupancy?

Some previous posts suggested to build a supercell and remove atoms to account for the 1/2 occupancy. However, in my case there are many possibilities for that.

How can one determine how big the supercell needs to be and which atoms to remove?

Thanks.

alex · #2 Post by **alex** » Thu Jun 06, 2013 9:58 am

Hi Lilia,

supercells are the only way. Typically I would double the unit cell in one direction and remove half of the C atoms.
If there are many possibilities, you need to calculate them all. However, you might omit the nonsense stuff, like full occupation in half the unit cell.

Cheers,

alex

boris · #3 Post by **boris** » Mon Jun 10, 2013 2:44 pm

Hi

You should use special quasirandom structures. It will allow you to (1) consider the 1/2 site occupancy and (2) avoid having to build several different supercells for the configuration averaging.

Cheers