ABINIT Discussion Forums

Hi.


I'm trying to calculate the band structure of magnetic metal system, which is one-dimensional linear atomic chain of Fe.

Simple convergence test works very well, but the problem is that it does not work if I added some input parameters for band structure calculation.


Here's my original ( works very well ) input file.


#K-point parameters
kptopt 1
nshiftk 1
shiftk 0.0 0.0 0.0
ngkpt 4 4 200

#Spin
spinat 0.0 0.0 4.0
nsppol 2

#Rather standard input for a bcc metal
iscf 5
#tolvrs 1.0d-10
acell 20 20 4.32
ecut 18.0
natom 1
nband 8

nstep 50
ntypat 1
occopt 3
rprim 1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
tsmear 0.01
typat 1
xred 0.0 0.0 0.0
znucl 26
prtden 1
toldfe 1.0d-10
# add to conserve old < 6.7.2 behavior for calculating forces at each SCF step
optforces 1


Because I'm not familiar with abinit package, I refered the input file of tutorial for spin, which is on bcc Fe.

Anyway, but, when I added some lines to calculate the band structure as below


ndtset 2

#Dataset 1: Self-consistent calculation

kptopt1 1
iscf1 5
nshiftk1 1
shiftk1 0.0 0.0 0.0
nband1 8
ngkpt1 5 5 50
prtden1 1
toldfe1 1.0d-10

#Dataset 2: Band structure calculation

iscf2 -2
getden2 -1
kptopt2 -3
nband2 8
ndivk2 50
kptbounds2 0.0 0.0 0.0
0.0 0.0 0.5

tolwfr2 1.0d-12
enunit2 1

#Spin
spinat 0.0 0.0 4.0
nsppol 2

tolvrs 1.0d-10
acell 20 20 4.32
ecut 18.0
natom 1
# nband 8

nstep 100
ntypat 1
occopt 3
rprim 1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
tsmear 0.01
typat 1
xred 0.0 0.0 0.0
znucl 26

optforces 1


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

The log says that 'some' input parameters are not correct, but I can not figure them out.

Can anyone figure out what's the problem in my input file..?


Thanks.

kptopt2 should be -1 as you are doing calculation on 1 segment.

Without testing it (and without coffee), I can't see anything else wrong (well... ngkpt 4 4 200 seems excessive and ecut seems low but it's your call).

Simon