Hi everybody
I am a new user of Abinit and I am trying to get DOS for NiO. I added following lines
prtdos 3 natsph 4 iatsph 1 2 3 4 ratsph 3.1 3.1 2.8 2.8
to the input file "tldau_3.in". But I got weird graph as shown.
I will be grateful if some body help me in this regard.
Thank you,
lamsalc
DOS calculation of NiO using LDA (without PAW)
Moderator: bguster
DOS calculation of NiO using LDA (without PAW)
- Attachments
-
- Projected_DOS_3d_Ni_ in_NiO.pdf
- Projected DOS for NiO
- (7.41 KiB) Downloaded 307 times
Re: DOS calculation of NiO using LDA (without PAW)
Hi,
If you modified the tldau_3.in file as you stated, I assume you obtained this .in :
This calculation is (as tests should be) quite underconverged on many aspects to allow the test to run in a few seconds on a single processors. If you want meaningful results, several points need to be corrected :
1) The cutoff energies are too low. For Nickel, the values
should be replaced by
2) The tolerance on the total energy is high.
should be replaced by
(or less, but it can get difficult for an antiferromagnetic calculation)
3) Most importantly, you should use a muh finer k-point grid. Replace :
by (at least) :
and then, continue to converge this parameter to higher values.
Good luck!
Jonathan Laflamme Janssen
If you modified the tldau_3.in file as you stated, I assume you obtained this .in :
Code: Select all
#Spin
nsppol 1
nspden 2
nspinor 1
spinat 0 0 1
0 0 -1
0 0 0
0 0 0
#Parameters
nstep 50
ecut 15
pawecutdg 30
iscf 17
toldfe 1.0d-5
nband 40
occopt 1
#tsmear 300 K
#Structural parameters
natom 4
ntypat 2
typat 1 1 2 2
znucl 28 8
xred 0 0 0
0.0 0.0 0.5
0.5 0.5 0.25
0.5 0.5 0.75
acell 3*7.92
rprim 0.0 1/2 1/2
1/2 0.0 1/2
1.0 1.0 0.0
# Kpoint Grid
ngkpt 2 2 2
chksymbreak 0 # The k point grid is not symmetric, but the calculations being for the ground-state, this is not a problem.
# LDA+U
usepawu 1
lpawu 2 -1
upawu 8.0 0.0 eV
jpawu 0.8 0.0 eV
#Density matrix
usedmatpu 5
dmatpawu
#Occupation matrix for spin 1 and atom 1
0.90036 0.00000 -0.00003 0.00000 0.00000
0.00000 0.90036 -0.00001 0.00000 0.00002
-0.00003 -0.00001 0.91309 -0.00001 0.00000
0.00000 0.00000 -0.00001 0.90036 -0.00002
0.00000 0.00002 0.00000 -0.00002 0.91309
#Occupation matrix for spin 1 and atom 2
0.89677 -0.00001 0.00011 -0.00001 0.00000
-0.00001 0.89677 0.00006 0.00001 -0.00010
0.00011 0.00006 0.11580 0.00006 0.00000
-0.00001 0.00001 0.00006 0.89677 0.00010
0.00000 -0.00010 0.00000 0.00010 0.11580
# add to conserve old < 6.7.2 behavior for calculating forces at each SCF step
optforces 1
prtdos 3
natsph 4
iatsph 1 2 3 4
ratsph 3.1 3.1 2.8 2.8
This calculation is (as tests should be) quite underconverged on many aspects to allow the test to run in a few seconds on a single processors. If you want meaningful results, several points need to be corrected :
1) The cutoff energies are too low. For Nickel, the values
Code: Select all
ecut 15
pawecutdg 30should be replaced by
Code: Select all
ecut 30
pawecutdg 602) The tolerance on the total energy is high.
Code: Select all
toldfe 1.0d-5should be replaced by
Code: Select all
toldfe 1.0d-7(or less, but it can get difficult for an antiferromagnetic calculation)
3) Most importantly, you should use a muh finer k-point grid. Replace :
Code: Select all
ngkpt 2 2 2by (at least) :
Code: Select all
ngkpt 8 8 4and then, continue to converge this parameter to higher values.
Good luck!
Jonathan Laflamme Janssen
Re: DOS calculation of NiO using LDA (without PAW)
Hi
Thank you very much for the reply.
Thank you very much for the reply.