Optimisation of Si(100) slab does not converge
142 views
Skip to first unread message
DMITRII Drugov
Nov 29, 2020, 8:24:02 PM11/29/20
to cp2k
Good day dear CP2K users, could you please have a look at my .in file to see did I specify ABC cell parameters and ALPHA_BETA_GAMMA for Si(100) with 16.29210*16.29210*5.76 A^3. I am using TZVP-MOLOPT-GTH and I allowed two top layer to relax while other bulk part is frozen. In total it only 42 atoms our of 85 is under relaxation. My system didn't reach any convergence yet, despite I run it for 30 hour on supercomputer with 120 GB memory and 36 CPU. However, I used same .in file for gold and Na slabs and with around 70-60 atoms with DZVP, and it converged within 1-2 hours. Do you think it's normal for TZVP, should I wait longer? Or smth wrong with my .in setting?
Regards,
Dmitrii
-------- Informations at step = 3 ------------
Optimization Method = LBFGS
Total Energy = 2.9038335460
Real energy change = -14.0555995686
Decrease in energy = YES
Used time = 15200.012
Convergence check :
Max. step size = 0.3482436711
Conv. limit for step size = 0.0030000000
Convergence in step size = NO
RMS step size = 0.0816981844
Conv. limit for RMS step = 0.0015000000
Convergence in RMS step = NO
Max. gradient = 1.6813091629
Conv. limit for gradients = 0.0004500000
Conv. for gradients = NO
RMS gradient = 0.4396446909
Conv. limit for RMS grad. = 0.0003000000
Conv. for gradients = NO
my setting are:
&GLOBAL
PROJECT Si100_optimisation
RUN_TYPE GEO_OPT
PRINT_LEVEL MEDIUM
&END GLOBAL
&FORCE_EVAL
METHOD QS
&DFT
BASIS_SET_FILE_NAME BASIS_MOLOPT
POTENTIAL_FILE_NAME GTH_POTENTIALS
CHARGE 0
MULTIPLICITY 1
&MGRID
CUTOFF 800
NGRIDS 5
REL_CUTOFF 70
&END MGRID
&QS
EPS_DEFAULT 1.0E-12
WF_INTERPOLATION ASPC
&END QS
&SCF
SCF_GUESS ATOMIC
EPS_SCF 1.0E-7
MAX_SCF 1000
CHOLESKY INVERSE
ADDED_MOS 20
&SMEAR ON
METHOD FERMI_DIRAC
ELECTRONIC_TEMPERATURE [K] 300
&END SMEAR
&DIAGONALIZATION
ALGORITHM STANDARD
&END DIAGONALIZATION
&MIXING
METHOD BROYDEN_MIXING
ALPHA 0.4
NBROYDEN 8
&END MIXING
&END SCF
&XC
&XC_FUNCTIONAL
&PBE
&END PBE
&END XC_FUNCTIONAL
&vdW_POTENTIAL
DISPERSION_FUNCTIONAL PAIR_POTENTIAL
&PAIR_POTENTIAL
PARAMETER_FILE_NAME dftd3.dat
TYPE DFTD3
REFERENCE_FUNCTIONAL PBE
R_CUTOFF 15.0
&END PAIR_POTENTIAL
&END vdW_POTENTIAL
&END XC
&POISSON
PERIODIC xy
POISSON_SOLVER ANALYTIC
&END POISSON
&END DFT
&SUBSYS
&CELL
ABC 16.404 16.404 50.0
ALPHA_BETA_GAMMA 90.0 90.0 120.0
PERIODIC xy
&END CELL
&COORD
Si 17.64977 -17.64977 -4.07302
Si 17.64977 -12.21907 -4.07302
Si 17.64977 -6.78837 -4.07302
Si 17.64977 -1.35767 -4.07302
Si 12.21907 -17.64977 -4.07302
Si 16.29210 -16.29210 -5.43070
Si 13.57675 -13.57675 -5.43070
Si 14.93442 -14.93442 -4.07302
Si 12.21907 -12.21907 -4.07302
Si 16.29210 -10.86140 -5.43070
Si 13.57675 -8.14605 -5.43070
Si 14.93442 -9.50372 -4.07302
Si 12.21907 -6.78837 -4.07302
Si 16.29210 -5.43070 -5.43070
Si 13.57675 -2.71535 -5.43070
Si 14.93442 -4.07302 -4.07302
Si 12.21907 -1.35767 -4.07302
Si 6.78837 -17.64977 -4.07302
Si 10.86140 -16.29210 -5.43070
Si 8.14605 -13.57675 -5.43070
Si 9.50372 -14.93442 -4.07302
Si 6.78837 -12.21907 -4.07302
Si 10.86140 -10.86140 -5.43070
Si 8.14605 -8.14605 -5.43070
Si 9.50372 -9.50372 -4.07302
Si 6.78837 -6.78837 -4.07302
Si 10.86140 -5.43070 -5.43070
Si 8.14605 -2.71535 -5.43070
Si 9.50372 -4.07302 -4.07302
Si 6.78837 -1.35767 -4.07302
Si 1.35767 -17.64977 -4.07302
Si 5.43070 -16.29210 -5.43070
Si 2.71535 -13.57675 -5.43070
Si 4.07302 -14.93442 -4.07302
Si 1.35767 -12.21907 -4.07302
Si 5.43070 -10.86140 -5.43070
Si 2.71535 -8.14605 -5.43070
Si 4.07302 -9.50372 -4.07302
Si 1.35767 -6.78837 -4.07302
Si 5.43070 -5.43070 -5.43070
Si 2.71535 -2.71535 -5.43070
Si 4.07302 -4.07302 -4.07302
Si 1.35767 -1.35767 -4.07302
Si 17.64977 -14.93442 -1.35767
Si 17.64977 -9.50372 -1.35767
Si 17.64977 -4.07302 -1.35767
Si 14.93442 -17.64977 -1.35767
Si 13.57675 -16.29210 -2.71535
Si 16.29210 -13.57675 -2.71535
Si 14.93442 -12.21907 -1.35767
Si 12.21907 -14.93442 -1.35767
Si 13.57675 -10.86140 -2.71535
Si 16.29210 -8.14605 -2.71535
Si 14.93442 -6.78837 -1.35767
Si 12.21907 -9.50372 -1.35767
Si 13.57675 -5.43070 -2.71535
Si 16.29210 -2.71535 -2.71535
Si 14.93442 -1.35767 -1.35767
Si 12.21907 -4.07302 -1.35767
Si 9.50372 -17.64977 -1.35767
Si 8.14605 -16.29210 -2.71535
Si 10.86140 -13.57675 -2.71535
Si 9.50372 -12.21907 -1.35767
Si 6.78837 -14.93442 -1.35767
Si 8.14605 -10.86140 -2.71535
Si 10.86140 -8.14605 -2.71535
Si 9.50372 -6.78837 -1.35767
Si 6.78837 -9.50372 -1.35767
Si 8.14605 -5.43070 -2.71535
Si 10.86140 -2.71535 -2.71535
Si 9.50372 -1.35767 -1.35767
Si 6.78837 -4.07302 -1.35767
Si 4.07302 -17.64977 -1.35767
Si 2.71535 -16.29210 -2.71535
Si 5.43070 -13.57675 -2.71535
Si 4.07302 -12.21907 -1.35767
Si 1.35767 -14.93442 -1.35767
Si 2.71535 -10.86140 -2.71535
Si 5.43070 -8.14605 -2.71535
Si 4.07302 -6.78837 -1.35767
Si 1.35767 -9.50372 -1.35767
Si 2.71535 -5.43070 -2.71535
Si 5.43070 -2.71535 -2.71535
Si 4.07302 -1.35767 -1.35767
Si 1.35767 -4.07302 -1.35767
&END COORD
&KIND Au
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q11
&END KIND
&KIND F
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q7
&END KIND
&KIND O
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q6
&END KIND
&KIND C
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q4
&END KIND
&KIND Si
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q4
&END KIND
&KIND S
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q6
&END KIND
&KIND N
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q5
&END KIND
&KIND Na
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q9
&END KIND
&END SUBSYS
&END FORCE_EVAL
&MOTION
&CONSTRAINT
&FIXED_ATOMS
COMPONENTS_TO_FIX XYZ
LIST 1..43
&END FIXED_ATOMS
&END CONSTRAINT
&GEO_OPT
OPTIMIZER LBFGS
MAX_ITER 300
&END GEO_OPT
&END MOTION
Lucas Lodeiro
Nov 29, 2020, 10:21:43 PM11/29/20
to cp...@googlegroups.com
Hi DMITRII,
I am not an expert on silicon systems, but I read plenty times that (001) silicon surfaces presents reconstruction when the surface is expanded (the reconstruction is not possible for the unitary surface cell), as your case. This could explain why do you have this problem with silicon and not with Au and Na. The first time that I read about this was "density functional theory: a practical introduction" of Janice Steckel and David Scholl, and in some papers, like this: https://www.researchgate.net/figure/Pictorial-view-of-the-100-surface-of-silicon-a-unrelaxed-surface-b-relaxed_fig5_282648906
Another thing is, Why do you use diagonalization instead of OT method? I expect that silicon system (even surface systems) were not metals... then the OT method will be faster.
Regards
Lucas Lodeiro
--
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 on the web visit https://groups.google.com/d/msgid/cp2k/162639f5-a816-408c-b798-55a572c13fffn%40googlegroups.com.
DMITRII Drugov
Nov 29, 2020, 10:52:41 PM11/29/20
to cp2k
Thank you for your reply!
I need to compare adsorption energy for organic molecules on different surfaces (Na, Au, Si, and C). I successfully used diagonalization for Na and Au as they are metals. However, I am not sure if I use OT for Si and C, it would be accurate to compare Eads with those of Na and Au. What do you think?
Do you have any suggestion how to expand Si surface to allow reconstructing?
Do you think fo graphite I also cannot use unitary cell?
Regards,
Dmitrii
fabia...@gmail.com
Nov 30, 2020, 4:36:28 AM11/30/20
to cp2k
Dear Dmitrii,
Your cell is not correct, all angles should be 90 degree.
The total energy you reported is positive. This indicates that your structure is far from stable. Where did you get the coordinates from? Did you optimize bulk Si with cp2k before creating the slab?
Cheers,
Fabian
oh...@chem.uni-sofia.bg
Nov 30, 2020, 4:57:02 AM11/30/20
to cp...@googlegroups.com
Dear all, may someone advice me on the following.
I want to get the proper antiferromagnetic coupling of two Cu2+ ions in
a bimetallic complex. For this purpose I use the &BS subsection. Mu
question concerns the MULTIPLICITY keyword in this case. What should be
the MULTIPLICITY in this case? If I specify only MULTIPLICITY 1, which
is the actual spin state, cp2k will do non-spinpolarized calculation,
which is not the case. That is why I use UKS T and MULTIPLICITY 1.
Is that correct setting?
Best regards,
Petko
I want to get the proper antiferromagnetic coupling of two Cu2+ ions in
a bimetallic complex. For this purpose I use the &BS subsection. Mu
question concerns the MULTIPLICITY keyword in this case. What should be
the MULTIPLICITY in this case? If I specify only MULTIPLICITY 1, which
is the actual spin state, cp2k will do non-spinpolarized calculation,
which is not the case. That is why I use UKS T and MULTIPLICITY 1.
Is that correct setting?
Best regards,
Petko
Krack Matthias (PSI)
Nov 30, 2020, 5:03:12 AM11/30/20
to cp...@googlegroups.com
Dear Petko
Yes, I think so.
Matthias
-----Ursprüngliche Nachricht-----
Von: cp...@googlegroups.com <cp...@googlegroups.com> Im Auftrag von oh...@chem.uni-sofia.bg
Gesendet: Montag, 30. November 2020 10:57
An: cp...@googlegroups.com
Betreff: [CP2K:14316] Broken Symmetry and MULTIPLICITY
Yes, I think so.
Matthias
-----Ursprüngliche Nachricht-----
Von: cp...@googlegroups.com <cp...@googlegroups.com> Im Auftrag von oh...@chem.uni-sofia.bg
Gesendet: Montag, 30. November 2020 10:57
An: cp...@googlegroups.com
Betreff: [CP2K:14316] Broken Symmetry and MULTIPLICITY
--
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 on the web visit https://groups.google.com/d/msgid/cp2k/1ebe18da-1f28-53dc-e4e8-2f70afbf90ee%40chem.uni-sofia.bg.
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.
oh...@chem.uni-sofia.bg
Nov 30, 2020, 5:05:38 AM11/30/20
to cp...@googlegroups.com, Krack Matthias (PSI)
Dear Matthias, thank you! I wanted to be sure I do not make mistake.
Petko
Petko
DMITRII Drugov
Nov 30, 2020, 6:57:39 PM11/30/20
to cp2k
Thank you for reply! I created two Si (100) systems with alpha beta gamma 90 90 90 and 90 90 120. The first system did't show any sign of convergence within 10 hours of simulation, whereas other fully converged (with 120 gamma degrees) at DZVP level within 19 hours. The system with 90 90 120 cell parameters performed about 66 steps within 19 hours, but the other 90 90 90 system performed only 2 steps for 10 hours. Before doing simulation, I created slab in avogadro and put in CP2K later. I wasn't sure should I choose 90 90 90 or 90 90 120 degree, as I saw smb published a work for Si (100) with 90 90 120 degrees cell parameters.
Here is my and ouput file for system with 90 90 120 cell angels.
Total Charge 0.002
!-----------------------------------------------------------------------------!
ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.): -6.901440296115577
-------- Informations at step = 66 ------------
Optimization Method = LBFGS
Total Energy = -6.9014402961
Real energy change = -0.0000072318
Decrease in energy = YES
Used time = 883.680
Convergence check :
Max. step size = 0.0025436647
Conv. limit for step size = 0.0030000000
Convergence in step size = YES
RMS step size = 0.0006204331
Conv. limit for RMS step = 0.0015000000
Convergence in RMS step = YES
Max. gradient = 0.0001874079
Conv. limit for gradients = 0.0004500000
Conv. in gradients = YES
RMS gradient = 0.0000472944
Conv. limit for RMS grad. = 0.0003000000
Conv. in RMS gradients = YES
---------------------------------------------------
*******************************************************************************
*** GEOMETRY OPTIMIZATION COMPLETED ***
*******************************************************************************
Reply all
Reply to author
Forward
0 new messages