Projected DOS not normalized

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
Apophis
Posts: 2
Joined: Tue Nov 06, 2012 12:51 pm

Projected DOS not normalized

Post by Apophis » Tue Aug 20, 2013 4:23 pm

Hello,

i am trying to calculate the projected DOS for Ni3C using PAW. While getting the compete DOS worked fine, the projections are wrong. For one, the integrated DOS at the Fermi energy is a multiple of the number available electrons, for all atoms. Also the DOS for the spin up and down channels have magnitudes while the form of the plot is roughly the same with no shift in energy (There is also not magnetization showing in the full DOS, so that part may be correct). The sum of the integrated spin down LDOSs roughly at the Fermi energy over all atoms alone roughly resembles the total number of electrons in the cell. The same sum for spin up is a multiple of that.

Stderr contains a lot o messages of the type
dens_in_sph : this state is not normalized : norm= 1.00699
dens_in_sph : this state is not normalized in real space : norm= 1.00699
dens_in_sph : this state is not normalized : norm= 0.84502
dens_in_sph : this state is not normalized in real space : norm= 0.84502


Here is my input file:

Code: Select all

acell 8.68823391483284E+00  8.68823391473050E+00  9.60347684124753E+00
rprim
  1.0 -4.38964554255e-29 -8.43617061338e-57
  -0.500000000002 0.866025403783 1.0184651346e-27
  0.452348355634 0.261163444886 0.852745343119
ntypat 2
znucl  28 6
natom 8
typat
 2 2 6*1
xred
 -3.94426681151606E-18 -2.32635776975131E-17  7.71373150378205E-19
 -1.31961253500590E-17  5.00000000000000E-01  5.00000000000000E-01
  1.70593337472857E-01  2.50000000000000E-01  2.50000000000000E-01
  5.00000000000000E-01  9.20593337472857E-01  2.50000000000000E-01
  1.70593337472857E-01  4.20593337472857E-01  7.50000000000000E-01
  8.29406662527143E-01  5.79406662527143E-01  2.50000000000000E-01
  5.00000000000000E-01  7.94066625271432E-02  7.50000000000000E-01
  8.29406662527143E-01  7.50000000000000E-01  7.50000000000000E-01

  chkprim 1

  toldff 1e-11

   ecut   36.0
   pawecutdg 52
   ecutsm .05

   occopt 3 #fermi
   tsmear 1e-3
   nband 44

ngkpt   3*12

nshiftk   1 # unshifted grid to make for results to be comparable with those of
            # strained cell
 shiftk   0 0 0

  nstep   100
# enforce calculation of forces at each SCF step
 optforces 1

prtwf 0
prtdos 3
natsph 8
iatsph 1 2 3 4 5 6 7 8
prtden 1

nsppol 2
spinat
   0 0 0
   0 0 0
   0 0 2
   0 0 2
   0 0 2
   0 0 2
   0 0 2
   0 0 2


I also tried pawprtdos=1/2. The on site contribution is tiny, which may not be surprising for a maetalic substance. The main contribution comes from the plane wave part, which shows the same caractersitics as described above.

I would be thankful for some suggestions.

michel.roger
Posts: 1
Joined: Fri Sep 06, 2013 5:09 pm

Re: Projected DOS not normalized

Post by michel.roger » Fri Sep 06, 2013 5:25 pm

I am a beginner in abinit.
I have the same normalisation problem of the projected DOS on Co sites in Na0.5CoO2, with
prtdos=3
pawprtdos=1 or pawprtdos=0

Coming back to the test program tpaw1_4.in in section 4 of the PAW1 tutorial
The projected DOS is correct after running the program sequentially.

When running the same program in parallel, using n processor, the results are roughly (but not exactly) multiplied by the number n of processors.

However I do obtain the correct results with
prtdos=3
pawprtdos=2

I am suspecting some bug in the parallel code?

ksmirnov
Posts: 19
Joined: Thu Jun 13, 2013 8:56 am

Re: Projected DOS not normalized

Post by ksmirnov » Tue Jul 07, 2015 3:58 pm

Indeed. The weird result is that the plane-waves contribution to DOS is multiplied by the number of processors in PAW calculations (tested on the tpaw1_4 exercise of the PAW1 tutorial).

Locked