Hi,
I am trying to converge a SCF calculation of a bilayer structure with SOC and
magnetic V: MoSe2/VSe2. See the input file below:
&CONTROL
calculation = 'scf' ,
/
&SYSTEM
ibrav = 0
celldm(1) = 1.0 ,
nat = 6 ,
ntyp = 3 ,
ecutwfc = 60.0,
occupations = 'smearing' ,
degauss = 0.001 ,
lspinorb = .true.,
noncolin = .true.,
starting_magnetization(1) = 0.001D0,
/
&ELECTRONS
mixing_mode = 'local-TF' ,
/
&IONS
/
&CELL
/
CELL_PARAMETERS alat
5.437914144 3.152910164 -0.003742247
5.437914144 -3.152910164 -0.003742247
-0.033355109 0.000000000 43.886125527
ATOMIC_SPECIES
V 1.000000 V.rel-pbe-spn-kjpaw_psl.0.2.3.UPF
Mo 1.000000 Mo.rel-pbe-spn-kjpaw_psl.0.3.0.UPF
Se 1.000000 Se.rel-pbe-n-kjpaw_psl.0.2.UPF
ATOMIC_POSITIONS angstrom
V -0.014505031 0.000000000 0.289175199
Se 1.913587627 0.000000000 1.854706939
Se 1.910003017 0.000000000 -1.304783106
Mo 1.916933515 0.000000000 6.724381364
Se 3.838132355 0.000000000 8.379171699
Se 3.835939653 0.000000000 5.070743199
K_POINTS automatic
1 1 1 0 0 0
The iteration steps look like:
total magnetization = 0.00 0.98 0.00 Bohr mag/cell
absolute magnetization = 847.77 Bohr mag/cell
The calculations are done in a single k-point just for testing. Anything
changes even increasing the k-mesh to 8 8 1. The pseudopotentials seem not to
be affecting the result, since the same calculation with US ones gives me
similar giant "absolute magnetization". Two important observations are:
(1) The same calculation for the monolayers (namely, only MoSe2 or VSe2)
converge perfectly.
(2) Increasing ecutwfc and ecutrho the absolute magnetization decreases but
still gives values of the order of 100 Bohr mag/cell
Any help on this will be welcome!
Thanks.
Best ,
David.
Dr. David Soriano
International Iberian Nanotechnology Lab - INL
Av. Mestre José Veiga s/n
4715-330 Braga
Portugal
Phone (INL): +351 253 140 112 Ext. 2378
Phone (mobile): +34 645 766 938
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum
