SCCS Convergence issue
Quadri Olakunle Adewuyi
I run geometry optimization with SCCS implicit solvent, but SCCS does not converge after 200 iterations.
Kindly help me check if I am missing something
here is my input and output details below
&GLOBAL
PROJECT POMI
RUN_TYPE GEO_OPT
PRINT_LEVEL MEDIUM
&END GLOBAL
&FORCE_EVAL
METHOD Quickstep
&DFT
BASIS_SET_FILE_NAME BASIS_MOLOPT
POTENTIAL_FILE_NAME GTH_POTENTIALS
!Charge and Multiplicity
Charge 0
MULTIPLICITY 1
&MGRID
CUTOFF 700
NGRIDS 4
REL_CUTOFF 60
&END MGRID
&QS
EPS_DEFAULT 1.0E-10
&END QS
&SCF
SCF_GUESS ATOMIC
EPS_SCF 1.0E-5
MAX_SCF 500
&OT ON
MINIMIZER DIIS
&END OT
&END SCF
&POISSON
PERIODIC NONE
POISSON_SOLVER MT
&END POISSON
&XC
&XC_FUNCTIONAL
&PBE
PARAMETRIZATION revPBE
&END PBE
&END XC_FUNCTIONAL
&VDW_POTENTIAL
POTENTIAL_TYPE PAIR_POTENTIAL
&PAIR_POTENTIAL
PARAMETER_FILE_NAME dftd3.dat
TYPE DFTD3(BJ)
REFERENCE_FUNCTIONAL PBE
R_CUTOFF [angstrom] 16
&END PAIR_POTENTIAL
&END VDW_POTENTIAL
&END XC
&SCCS ON
DERIVATIVE_METHOD FFT
DIELECTRIC_CONSTANT 78.36
EPS_SCCS 1.0E-6
EPS_SCF 1.0E-5
MAX_ITER 200
METHOD ANDREUSSI
!METHOD FATTEBERT-GYGI
MIXING 0.2
&ANDREUSSI
RHO_MAX 0.001
RHO_MIN 0.0001
&END ANDREUSSI
!&FATTEBERT-GYGI
!BETA 1.3
!RHO_ZERO 0.0004
!&END FATTEBERT-GYGI
&END SCCS
&END DFT
&SUBSYS
&TOPOLOGY
COORD_FILE_NAME POMl.xyz
COORD_FILE_FORMAT XYZ
&END TOPOLOGY
&CELL
! unit cells that are orthorhombic are more efficient with CP2K
A [angstrom] 36.00000000 0.000000000 0.000000000
B [angstrom] 0.000000000 36.00000000 0.000000000
C [angstrom] 0.000000000 0.000000000 36.00000000
PERIODIC NONE
&END CELL
&KIND W
ELEMENT W
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q14
&END KIND
&KIND P
ELEMENT P
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q5
&END KIND
&KIND O
ELEMENT O
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q6
&END KIND
&KIND H
ELEMENT H
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q1
&END KIND
&END SUBSYS
&TOTAL_NUMBERS ON
&END TOTAL_NUMBERS
&END PRINT
&END FORCE_EVAL
&MOTION
&GEO_OPT
TYPE MINIMIZATION
MAX_ITER 200
OPTIMIZER LBFGS
&LBFGS
MAX_F_PER_ITER 20
&END LBFGS
&END GEO_OPT
&END MOTION
here is the output
-----------------------------------------------------------------------------
1 OT CG 0.15E+00 295.6 0.00073221 -3858.9269373219 -3.86E+03
*** WARNING in qs_sccs.F:594 :: The SCCS iteration cycle did not converge ***
*** in 200 steps ***
2 OT LS 0.60E+00 791.6 -3860.5237639880
3 OT CG 0.60E+00 347.1 0.00088457 -3858.9883305609 -6.14E-02
4 OT LS 0.60E-02 330.8 -3858.0304951763
5 OT CG 0.60E-02 233.0 0.00088692 -3858.9868935523 1.44E-03
6 OT LS 0.11E-01 250.0 -3858.9914525675
7 OT CG 0.11E-01 270.8 0.00084581 -3858.9954528356 -8.56E-03
8 OT LS 0.20E-01 462.9 -3859.0030589695
9 OT CG 0.20E-01 471.2 0.00079042 -3859.0171304820 -2.17E-02
*** WARNING in qs_sccs.F:594 :: The SCCS iteration cycle did not converge ***
*** in 200 steps ***
10 OT LS 0.79E-01 756.4 -3859.3170317584
*** WARNING in qs_sccs.F:594 :: The SCCS iteration cycle did not converge ***
*** in 200 steps ***
11 OT CG 0.79E-01 789.4 351.12465571 -4029.5115025971 -1.70E+02
12 OT LS 0.40E-01 280.1 -3798.2663029849
13 OT CG 0.40E-01 213.4 0.01133245 -3806.9423627571 2.23E+02
14 OT LS 0.16E+00 162.5 -3812.8065913366
15 OT CG 0.16E+00 210.6 0.00805198 -3826.1089223976 -1.92E+01
16 OT LS 0.63E+00 189.0 -3838.1840259780
17 OT CG 0.63E+00 139.3 0.01014323 -3836.3256620356 -1.02E+01
18 OT LS 0.69E-01 178.3 -3783.3719988474
19 OT CG 0.69E-01 191.3 0.00797083 -3836.4177607144 -9.21E-02
However when I changed the EPS under the SCCS section as below, it did converge. Still, it did not give information about solvation-free and polarization energy which makes with and without solvent have the same energy. &SCCS ON
DERIVATIVE_METHOD FFT
DIELECTRIC_CONSTANT 78.36
EPS_SCCS 1.0E-7
EPS_SCF 1.0E-6
output
Electronic density on regular grids: -1119.9999756896 0.0000243104
Core density on regular grids: 1119.9999999412 -0.0000000588
Total charge density on r-space grids: 0.0000242516
Total charge density g-space grids: 0.0000242516
Overlap energy of the core charge distribution: 0.00003421286733
Self energy of the core charge distribution: -7742.74303176718968
Core Hamiltonian energy: 1994.07265119910176
Hartree energy: 2522.08940846651103
Exchange-correlation energy: -632.17490075971477
Dispersion energy: -0.55798902100351
SCCS| Hartree energy of solute and solvent [Hartree] 2522.08940846651103
SCCS| Hartree energy of the solute only [Hartree] 2522.08940846651103
SCCS| Polarisation energy [Hartree] 0.00000000000000
SCCS| [kcal/mol] 0.000
SCCS| Cavitation energy [Hartree] 0.00000000000000
SCCS| [kcal/mol] 0.000
SCCS| Dispersion free energy [Hartree] 0.00000000000000
SCCS| [kcal/mol] 0.000
SCCS| Repulsion free energy [Hartree] 0.00000000000000
SCCS| [kcal/mol] 0.000
SCCS| Solvation free energy [Hartree] 0.00000000000000
SCCS| [kcal/mol] 0.000
Total energy: -3859.31382766942852
I will appreciate your kind suggestion and support.
Thank you in advance
Quadri Adewuyi
PhD student
Muhammad Saleh
Try to increase the iteration let say 800, I know it is quite long but since it is the first time, sometime it takes long iteration to reach the convergence especially if your initial coordinate is far from equilibration. The other one is to set a large SCF threshold for instance to -02 and -03 (you have set the SCF and SCC to -05 and -06). Once it converge, you can use the last convergence coordinate and restart the calculation but this time, set as -06 or -07. Hope it helps
--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/CAMAC%2BZqkBtE-Nh6Q-dhG9oHmkL9JxOaSj%3D1MSxiXVuRPhSO0tg%40mail.gmail.com.
Quadri Olakunle Adewuyi
I will try to increase the iteration and update you.
Thanks so much
To view this discussion visit https://groups.google.com/d/msgid/cp2k/CAPD_w6wwbH2mN%3DQStUg-ADigHhstgGmowO6vz6BYvAG8RvTzHw%40mail.gmail.com.