LSORBIT = .TRUE.: PARCHG problem

Message

blazing · #1 Post by **blazing** » Fri Jan 04, 2013 2:23 pm

Dear all,

I try to calculate the total charge density by means of Band decomposed chargedensity calculation (i.e. PARCHG file) with SOC included (LSORBIT = .TRUE.). I follow the instructions for the PARCHG calculation in manual. The outcome is very puzzling - when I integrate my PARCHG I see that the number of electrons is smaller than NBANDS.

Does anybody has idea why it is so?

When I do similar calculation without SOC everything is OK.

=========================

Some input details.

1st run - static SCF calculation (only relevant tags)
ISTART = 0
ICHARG = 1
ISYM = 0
#LWAVE = .FALSE.
#LCHARG = .FALSE.
NSW = 0
IBRION = -1

2nd run - PARCHG calculation - EVERYTHING IS THE SAME EXCEPT FOR
ISTART = 1
ICHARG = 11
LPARD = .TRUE.
NBMOD = -2
EINT = -17 # only occupied bands
LSEPB = .FALSE.
LSEPK = .FALSE.
===========================

Any suggestions are greatly appreciated.

Cheers,

Misha

yuuukuma · #2 Post by **yuuukuma** » Thu Mar 07, 2013 8:42 am

I also found smaller partial charge density with non collinear calculations with vasp.5.3.3. Is the compensation charge included in the PARCHG?

Best regards,
Yu