I'm try to calculate density of state by using HSE hybrid functional, but in Vasp example unclear for me (http://cms.mpi.univie.ac.at/wiki/index.php/FccNi_DOS). Why they didn't use ICHARG=11 or How calculatinon density of state by using HSE dffer from normal LDA or GGA?
In VASP, it is not necessary to do a separate calculation to obtain the density of states. After a self-consistent run, the DOSCAR file will be there and the density of states can be extracted. Often time, the use of a non-self consistent run (i.e. ICHARG=11) is because a more detailed DOSCAR is desired. In this case, one can set ICHARG = 11 to keep the charge density constant from the scf run, but the number of k-points can be changed as can NEDOS to provide a more detailed DOS. Doing it this way would be faster than redoing the scf calculation on the denser grid. However, it isn't always necessary depending upon the complexity of the DOS or what exactly you want to use the DOS for. As far as I can tell, it should work the same way in HSE as it does in LDA or GGA.
A little more info here:
http://cms.mpi.univie.ac.at/vasp/vasp/Accurate_DOS_Band_structure_calculations.html
I would like to add more points from Dr. Andew:
For calculating DOS of a big system (say more than 100 atoms), it is practical to perform a self-consistent calculation to obtain charge density. Afterwards, this charge density is used to proceed another non-self consistent run by setting ICHARG = 11. The actual benefit is that we can afford DOS of a large system at less computational cost and memory even though dense k-point mesh of BZ is sampled.
However, this approach works only for LDA and GGA, not hybrid-type functionals, e. g., HSE06. The reason is that the charge density obtained by these particular cases is not only from Kohn-Sham orbitals but also from Hartree-Fock wave functions. So, one cannot evaluate HSE06 DOS by reading the converged charge density, i. e., using ICHARG = 11.
Thanayut
Dear, Dr. Thanayut Kaewmaraya, I want to ask here one doubt as you told one cannot evaluate HSE06 DOS by reading the converged charge density (i.e ICHARG=11) so in case of hybrid functional how does calculate dos spectra by only SCF calculation.
Thank You
Shilendra
Dear Shilendra
My answer is similar to Dr. Andrew's. You can simply get HSE06 DOS by performing a SCF run and the DOS is given in DOSCAR file. In fact, HSE06 is computationally demanding due to the evaluation of the HF exchange term. Unlike LDA and GGA, it is quite difficult to sample many k-points in order to have nicely smooth DOS.
However, there is a warning associated with this limitation. It is important to ensure that you sample sufficient k-points especially for indirect-gapped semiconductors. The listed k-points in BZ (i. e., IBZKPT file) must explicitly include the point at which the conduction band minimum is located. Otherwise, the calculated DOS and band gap is less reliable.
I hope that my answer can help you.
Best wishes,
Thanayut
Thank You Dear, Dr. Thanayut Kaewmaraya for your reply, If I will have any doubt I will ask you in future !
Shilendra
How can we sample KPOINTS for the calculation of Band Structure using HSE06. If we have hexagonal and body centered tetragonal system how, can we add desired KPOINTS with zero weight in addition to the KPOINTS from IBZKPT file.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Semiconductor