I'm trying to converge a ground state scf cycle of a ferromagnetic compounds (containing half-filled Eu 4f electrons) taking into account SO coupling in PAW.
It's not converging:
Code: Select all
ETOT 250 -932.41424563692 7.682E-01 3.847E-02 4.368E+02
ETOT 251 -936.95319587233 -4.539E+00 8.552E-02 5.962E+02
ETOT 252 -940.18025543856 -3.227E+00 3.904E+00 3.913E+02
ETOT 253 -926.51020300340 1.367E+01 2.602E-01 9.431E+01
ETOT 254 -926.75519625814 -2.450E-01 2.661E-02 8.302E+01
ETOT 255 -932.57849535648 -5.823E+00 2.628E-01 3.813E+02
ETOT 256 -926.55239654446 6.026E+00 1.684E-01 6.018E+01
ETOT 257 -931.20513213756 -4.653E+00 2.538E-01 2.740E+02Here are the relevant input variables:
Code: Select all
nline 20
kptopt 4
ngkpt 4 4 4
nshiftk 1
shiftk 0 0 1/2
tolvrs1 1.0d-16
nstep1 1000
diemac1 4.0
iscf 17
nbdbuf 22
nband *224
ntypat 5
znucl 63.0 56.0 14.0 8.0 7.0
usepawu 1
lpawu 3 -1 -1 -1 -1
upawu 7.5 0.0 0.0 0.0 0.0 eV
jpawu 0.6 0.0 0.0 0.0 0.0 eV
nsppol 1
nspden 4 # Non Colinear Magnetism
nspinor 2 # Non Colinear Magnetism
pawspnorb 1 # Spin-orbit coupling
spinat 0 0 7 66*0 # Impose magnetism
diemix 0.6
pawecutdg 60
ecut 30
ecutsm 0.5
ixc 11
paral_kgb 0
I have try the following (does not work)
- Change nline (from 4 to 20)
- use dmatpawu :
Code: Select all
dmatpuopt 2
usedmatpu -3 # IMPOSE DENSITY MATRIX IN THE FIRST 10 STEPS
dmatpawu 1.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 1.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 1.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 1.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 1.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 1.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 1.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0
- Use diemixmag -4 (I also tried -2)
Does anyone have any ideas how what could be wrong?
Thank you
Samuel.