[Pw_forum] Orthogonality of wavefunctions

[Bo Qiu]

Dear developers and users,

I'm trying to compute some matrix elements between states k and k'. To
confirm my calculation, I first try to compute the overlap between
wavefunction k and k' as  < k| k'> in quantum espresso by taking zdoc of
state k and k' (modified the elphonon.f90 code). I do find for the same k
point, the overlap between different bands are 0. However, the overlap
between two states at different points k and k' are almost always non-zero,
indicating they're not orthogonal. I thought in theory they should all be
orthonormal because they belong to the same Hamiltonian of
the entire system. So is it because of numerical reasons that they're
actually not orthogonal in quantum espresso?

Thanks a lot for you help!

Bo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20131031/6aa229c6/attachment.html>