Convergence problem for ionic relaxation in a charged system
Posted: Tue Jun 26, 2012 5:30 pm
I am optimizing a structure in a charged system for a particular magnetic state; However, Even if I try difference relaxation strategy(quasi newton and CG), different subspace diagonalization method(Davison and Residual minimum), symmetry tag(ISYM= -1,0, default), and different initial geometry, I am always getting a bouncing dE(the energy fails to decrease continuously) as follows: Do you have some idea to overcome this ionic convergence problem?
48 F= -.26219840E+03 E0= -.26219840E+03 d E =-.327147E-03 mag= 6.0000
49 F= -.26219893E+03 E0= -.26219893E+03 d E =-.527741E-03 mag= 6.0000
50 F= -.26219919E+03 E0= -.26219919E+03 d E =-.259314E-03 mag= 6.0000
51 F= -.26219809E+03 E0= -.26219809E+03 d E =0.109917E-02 mag= 6.0000
52 F= -.26219562E+03 E0= -.26219562E+03 d E =0.247148E-02 mag= 6.0000
53 F= -.26219434E+03 E0= -.26219434E+03 d E =0.127219E-02 mag= 6.0000
54 F= -.26219575E+03 E0= -.26219575E+03 d E =-.140522E-02 mag= 6.0000
55 F= -.26219585E+03 E0= -.26219585E+03 d E =-.105190E-03 mag= 6.0000
56 F= -.26219623E+03 E0= -.26219623E+03 d E =-.373362E-03 mag= 6.0000
57 F= -.26219663E+03 E0= -.26219663E+03 d E =-.398108E-03 mag= 6.0000
58 F= -.26219668E+03 E0= -.26219668E+03 d E =-.557682E-04 mag= 6.0000
59 F= -.26219637E+03 E0= -.26219637E+03 d E =0.307803E-03 mag= 6.0000
60 F= -.26219559E+03 E0= -.26219559E+03 d E =0.781366E-03 mag= 6.0000
Below is my input file:
PREC = accurate
ISTART = 0
ICHARG = 2
ISPIN = 2
NELECT = 176 #177 for nuetral
IALGO =48
LREAL = Auto
NBAND =140
EDIFF = 1E-5
ISYM =-1
NELMDL = -5
NELMIN = 8
NELM = 40
LMAXMIX = 4
IBRION = 1 # I also tried 2
ENCUT = 530
NSW = 60
EDIFFG = -0.01
ISMEAR = 0
SIGMA = 0.05
ISIF = 0
VOSKOWN =1
MAGMOM =5 -5 -5 -5 5 -5 5 -5 5 -5 5 -5 18*0
LDAU = .TRUE.
LDAUTYPE = 1
LDAUL = 2 0 0
LDAUU = 4.0 0 0
LDAUJ = 0.58 0 0
48 F= -.26219840E+03 E0= -.26219840E+03 d E =-.327147E-03 mag= 6.0000
49 F= -.26219893E+03 E0= -.26219893E+03 d E =-.527741E-03 mag= 6.0000
50 F= -.26219919E+03 E0= -.26219919E+03 d E =-.259314E-03 mag= 6.0000
51 F= -.26219809E+03 E0= -.26219809E+03 d E =0.109917E-02 mag= 6.0000
52 F= -.26219562E+03 E0= -.26219562E+03 d E =0.247148E-02 mag= 6.0000
53 F= -.26219434E+03 E0= -.26219434E+03 d E =0.127219E-02 mag= 6.0000
54 F= -.26219575E+03 E0= -.26219575E+03 d E =-.140522E-02 mag= 6.0000
55 F= -.26219585E+03 E0= -.26219585E+03 d E =-.105190E-03 mag= 6.0000
56 F= -.26219623E+03 E0= -.26219623E+03 d E =-.373362E-03 mag= 6.0000
57 F= -.26219663E+03 E0= -.26219663E+03 d E =-.398108E-03 mag= 6.0000
58 F= -.26219668E+03 E0= -.26219668E+03 d E =-.557682E-04 mag= 6.0000
59 F= -.26219637E+03 E0= -.26219637E+03 d E =0.307803E-03 mag= 6.0000
60 F= -.26219559E+03 E0= -.26219559E+03 d E =0.781366E-03 mag= 6.0000
Below is my input file:
PREC = accurate
ISTART = 0
ICHARG = 2
ISPIN = 2
NELECT = 176 #177 for nuetral
IALGO =48
LREAL = Auto
NBAND =140
EDIFF = 1E-5
ISYM =-1
NELMDL = -5
NELMIN = 8
NELM = 40
LMAXMIX = 4
IBRION = 1 # I also tried 2
ENCUT = 530
NSW = 60
EDIFFG = -0.01
ISMEAR = 0
SIGMA = 0.05
ISIF = 0
VOSKOWN =1
MAGMOM =5 -5 -5 -5 5 -5 5 -5 5 -5 5 -5 18*0
LDAU = .TRUE.
LDAUTYPE = 1
LDAUL = 2 0 0
LDAUU = 4.0 0 0
LDAUJ = 0.58 0 0