Bader volumes too large?

[moe]

Hello,

I'm running some sanity checks with Bader 1.04 on bulk NaCl and KI in Quantum Espresso. I'm following the recommended method for pseudopotential codes which uses the all-electron charge as a reference.

I am getting sensible results for Bader charge for cations and anions (around 0.8 ionic charge for Na and K, -0.8 for Cl and I).

The Bader volumes in the ACF.dat file, however, seem to be too large. The volumes of the NaCl and KI unit cells are 130 A^3 and 341 A^3, respectively. However, I am getting a Bader volume of 168 A^3 and 414 A^3 for the Cl and I, respectively.

Are these number reasonable? How can the Bader volume associated with a single atom within the simulation cell be larger than the entire cell?

I've tested this on some isolated molecule-in-a-box systems with vacuum and it's a similar trend. Atoms with a net negative charge have very large volumes.

I've attached my ACF output files as well as the Cube files used to generate them.

Thank you very much!

[graeme]

For cube files, the atomic volumes are in bohr^3.