soft and augmentation electrons

[giacsport]

Dear moderator,
I am working with VASP 4.6.35 version on charged GaAs system
I want to analyze charge behavior according to Bader analysis
So, as suggested I explicit in my INCAR the script LAECHARG=.TRUE.

and then I increase NGXF,Y, and Z up to 124
Restarting as initial guess from a previous WAVECAR (slightly different in terms of basis set from the present one), and CONTCAR

I am using hard pseudo for Ga (PAW, 3d in the semicore) and "standard" (PAW, 3d in the core)for As
I get this log

running on 2 nodes
distr: one band on 1 nodes, 2 groups
vasp.4.6.35 3Apr08 complex
POSCAR found : 3 types and 64 ions
LDA part: xc-table for Ceperly-Alder, standard interpolation
found WAVECAR, reading the header
nup: number of bands has changed, file: 0 present: 346
POSCAR, INCAR and KPOINTS ok, starting setup
FFT: planning ... 11
reading WAVECAR
augmentation electrons 0.108154249887350
soft electrons 0.000000000000000E+000
total electrons 0.108154249887350
augmentation electrons 0.108154249887350
soft electrons 0.000000000000000E+000
total electrons 0.108154249887350
augmentation electrons 6.603473210636590E-003
soft electrons 0.000000000000000E+000
total electrons 6.603473210636590E-003
augmentation electrons 1490.29547338432
augmentation electrons 1490.29547338432
soft electrons 0.000000000000000E+000
total electrons 1490.29547338432
augmentation electrons 1498.00956454160
soft electrons 0.000000000000000E+000
augmentation electrons 1498.00956454160
soft electrons 0.000000000000000E+000
total electrons 1498.00956454160
total electrons 1498.00956454160
WARNING: random wavefunctions but no delay for mixing, default for NELMDL
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 0.627098531031E+04 0.62710E+04 -0.15433E+05 49960 0.198E+03
DAV: 2 0.107769143562E+04 -0.51933E+04 -0.50459E+04 49824 0.540E+02
DAV: 3 -0.166203891255E+03 -0.12439E+04 -0.11596E+04 54304 0.198E+02
DAV: 4 -0.305564199416E+03 -0.13936E+03 -0.13618E+03 57944 0.702E+01
DAV: 5 -0.310731169767E+03 -0.51670E+01 -0.51056E+01 56584 0.124E+01BRMIX: very serious problems
the old and the new charge density differ
old charge density: 577.00000 new 578.00000
0.141E+01
DAV: 6 -0.304694282858E+03 0.60369E+01 -0.88053E+00 64718 0.123E+01 0.752E+00
DAV: 7 -0.303244706282E+03 0.14496E+01 -0.55560E+00 65840 0.607E+00 0.207E+00
DAV: 8 -0.303155481126E+03 0.89225E-01 -0.80074E-01 61694 0.727E+00 0.539E-01
DAV: 9 -0.303166028254E+03 -0.10547E-01 -0.15193E-01 63376 0.170E+00 0.210E-01
DAV: 10 -0.303167260655E+03 -0.12324E-02 -0.33206E-02 66740 0.897E-01 0.120E-01
DAV: 11 -0.303167046491E+03 0.21416E-03 -0.58759E-03 59472 0.760E-01 0.506E-02
DAV: 12 -0.303167095720E+03 -0.49229E-04 -0.16333E-03 54880 0.171E-01 0.199E-02
DAV: 13 -0.303167232256E+03 -0.13654E-03 -0.32566E-04 39600 0.825E-02 0.105E-02
DAV: 14 -0.303167224825E+03 0.74316E-05 -0.43190E-05 28920 0.745E-02
1 F= -.30316722E+03 E0= -.30316722E+03 d E =-.303167E+03 mag= 0.0000
BRION: g(F)= 0.449E-03 g(S)= 0.000E+00
reached required accuracy - stopping structural energy minimisation
writing wavefunctions
augmentation electrons 242.089083660625
soft electrons 336.009304333332
total electrons 578.098387993957
augmentation electrons 242.089083660625
soft electrons 336.009304333332
total electrons 578.098387993957
augmentation electrons -1.003642071686940E-006
soft electrons 336.009304333332
total electrons 7.584046265961230E-009
augmentation electrons -1.003642071686940E-006
soft electrons 336.009304333332
total electrons 7.584046265961230E-009

My system is characterized by 578 e- of valence (and this seems to match with some numbers in the log).
But could you kindly help me in understanding the meaning of all the numbers on top and bottom....Are they correct?

And moreover, is there in principle a "safe" way for setting good initial values of NGXF,Y and Z?

Thanks in advance

Best,
Giac

[graeme]

I don't know about the charges listed in the output; these are not related to the Bader charges.

For the values of NGXF,Y,Z, you should start with something near the default and then systematically increase the grid density and check for convergence of the Bader charges. Typically, using a grid which is twice the default is acceptable.