BANDER charges in bulk gan

[quanghieu]

Dear comunity,

I am calculating bader charges from a cube file for hexagonal GaN bulk system. The Ga and N layers in this compound are in the form of : ABA'B'ABA'B'..... where the difference between A and A' layers is that they are slightly shifted laterally with respect to each other. Similarly, the B and B' layers. First, I have generated a charge density cube file by doing post-processing of the wavefunctions collected from the scf calculation of bulk 2x2x2 GaN system in Quantum Espresso. Then I ran the bader code for the cube file to get the charges. The k-point grid I used is 12x12x12 .

The total charge of the system comes out to be correct (the expected valence charge of the system). Also, the charges on atoms of the same layer are same. But the problem is that the charges on Ga atoms of layers A and A' (and also charges on N atoms of layers B and B') are not same. This is a bulk system and all the Ga atoms are in same chemical environment. They are tetrahedrally bonded to 4 N atoms (and same is true for N atoms, N atoms in layers B and B' are tetrahedrally bonded to 4 Ga atoms). So, I expected that the charges on all the Ga atoms in this bulk system should be same. Can someone please tell me what I am missing?

[graeme]

I would have to take a look at the file(s) to really debug this. But make sure that you have a file with the core charge and you can try using the "-b weight" method for higher accuracy.