Dielectric matrix and LDA+U
Posted: Sat Jun 05, 2010 1:08 pm
Dear VASP Master and all users,
I did calculation of the frequency dependent dielectric
matrix (LOPTICS = .TRUE.) for the Mott insulator using the LDA+U.
In the electronic density of states, there is a wide energy gap ~2 eV,
so the elctronic ground state is well describe by LDA+U.
But, in the optical absorption calculated directly from the dielectric
function, there is no indication of the gap,
In fact, the absorption spectra are similar as in a metal (U=0).
I have a question how the interaction U (if any) is included
in the electronic ground state used to obtain the dielectric matrix.
I have used VASP 5.2.2
Thanks for any anwsers,
Przemek
I did calculation of the frequency dependent dielectric
matrix (LOPTICS = .TRUE.) for the Mott insulator using the LDA+U.
In the electronic density of states, there is a wide energy gap ~2 eV,
so the elctronic ground state is well describe by LDA+U.
But, in the optical absorption calculated directly from the dielectric
function, there is no indication of the gap,
In fact, the absorption spectra are similar as in a metal (U=0).
I have a question how the interaction U (if any) is included
in the electronic ground state used to obtain the dielectric matrix.
I have used VASP 5.2.2
Thanks for any anwsers,
Przemek