Excited state geometry optimization and wrong state

Simil Thomas

unread,

Feb 14, 2023, 2:37:45 AM2/14/23

to cp2k

Dear all,

I am trying an excited state optimization with a supercell approach.

Excitation energy at the first step is 1.73eV and it increased to 2.06 after a few geometry optimization. Actually first few state are missing in later optimization steps.

Could anyone tell me what is the problem? I am attaching my input:

(My geometry is not optimized even after 200 steps).

&GLOBAL
PROJECT ex-geom
RUN_TYPE GEO_OPT
PRINT_LEVEL MEDIUM
PREFERRED_DIAG_LIBRARY SL
&END GLOBAL

&MOTION
&GEO_OPT
TYPE MINIMIZATION
OPTIMIZER LBFGS
MAX_DR 3.00E-03
MAX_ITER 200
&END GEO_OPT
&END MOTION

&FORCE_EVAL
METHOD Quickstep
&PROPERTIES
&TDDFPT
NSTATES 20
MAX_ITER 50 !default
CONVERGENCE [eV] 1.0e-5 !default
&END TDDFPT
&END PROPERTIES
&PRINT
&FORCES ON
&END
&END PRINT

&SUBSYS
&KIND Br
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE
&END KIND
&KIND Pb
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE
&END KIND
&KIND C
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE
&END KIND
&KIND N
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE
&END KIND
&KIND H
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE
&END KIND

&CELL
A 5.97560000000000 0.00000000000000 0.00000000000000
B 0.00000000000000 5.97560000000000 0.00000000000000
C 0.00000000000000 0.00000000000000 5.97560000000000

PERIODIC XYZ
MULTIPLE_UNIT_CELL 4 4 4
&END CELL

&TOPOLOGY
MULTIPLE_UNIT_CELL 4 4 4
&END TOPOLOGY

&COORD
C 3.1299567815 3.5956831663 3.1377383495
N 4.2931353279 2.9981258408 3.1326107225
N 1.9686944529 2.9953377709 3.1527119635
H 3.1283865094 4.6859526863 3.1313428606
H 5.1101384877 3.5957093479 3.1053467119
H 4.4233324318 1.9894674912 3.1442365125
H 1.8385358141 1.9863125590 3.1652668075
H 1.1513074717 3.5927213004 3.1507668836
Pb 0.0812552926 0.0761824861 0.1207708511
Br 3.0968659468 0.1463361906 0.0098695983
Br 0.1375704475 3.0569071373 0.1548687543
Br 0.1502768675 -0.0394324052 3.1039318603
&END COORD
&END SUBSYS

&DFT
CHARGE 0
MULTIPLICITY 1

BASIS_SET_FILE_NAME ./BASIS_MOLOPT
POTENTIAL_FILE_NAME ./POTENTIAL
&POISSON
PERIODIC XYZ !default
&END POISSON

&QS
METHOD GPW
EPS_DEFAULT 1.0E-16 !default
&END QS
&MGRID
NGRIDS 4 !default
CUTOFF 280 !default
REL_CUTOFF 40 !default
&END MGRID
&EXCITED_STATES T
STATE 1 ! the excited state to be optimized
&END EXCITED_STATES
&SCF
SCF_GUESS ATOMIC !default
EPS_SCF 1.0E-7 !default
MAX_SCF 300
&OT
MINIMIZER DIIS
PRECONDITIONER FULL_SINGLE_INVERSE
&END
&OUTER_SCF

EPS_SCF 1.0E-7
MAX_SCF 10
&END
&END SCF
&XC
&XC_FUNCTIONAL PBE
&END XC_FUNCTIONAL
&VDW_POTENTIAL
DISPERSION_FUNCTIONAL PAIR_POTENTIAL
&PAIR_POTENTIAL
REFERENCE_FUNCTIONAL PBE
TYPE DFTD3(BJ)
PARAMETER_FILE_NAME dftd3.dat
LONG_RANGE_CORRECTION .FALSE.
VERBOSE_OUTPUT TRUE
&END PAIR_POTENTIAL
&END VDW_POTENTIAL
&XC_GRID
XC_DERIV SPLINE2_SMOOTH
&END XC_GRID
&END XC
&END DFT
&END FORCE_EVAL

with regards

Simil

Jürg Hutter

unread,

Feb 21, 2023, 4:43:23 AM2/21/23

to cp...@googlegroups.com

Hi

you have replicated your system 64 times. I assume this leads to a very dense spectrum at the
absorption edge. Such a high (close) degeneracy can lead to numerical difficulties for the
gradient calculation. Did you check that all parts of the calculation did converge smoothly and
the results are reasonable, e.g. initial gradients?

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Simil Thomas <sim...@gmail.com>
Sent: Tuesday, February 14, 2023 8:37 AM
To: cp2k
Subject: [CP2K:18445] Excited state geometry optimization and wrong state

--
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<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/52f385da-d503-4797-8209-adba6008e6fbn%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/52f385da-d503-4797-8209-adba6008e6fbn%40googlegroups.com?utm_medium=email&utm_source=footer>.

SIMIL THOMAS

unread,

Feb 21, 2023, 6:18:32 AM2/21/23

to cp...@googlegroups.com

Dear Prof. Hutter,

Yes. I think all parts converged smoothly. At the first step:

Max. gradient = 0.0036330171

- TDDFPT run converged in 6 iteration(s)
-------------------------------------------------------------------------------

R-TDDFPT states of multiplicity 1

State Excitation Transition dipole (a.u.) Oscillator
number energy (eV) x y z strength (a.u.)
------------------------------------------------------------------------
TDDFPT| 1 1.73629 2.4288E-01 -7.6831E-02 -6.3067E+00 1.69469E+00
TDDFPT| 2 1.94853 -7.4425E+00 -8.6367E-01 -1.1093E-01 2.68043E+00
TDDFPT| 3 1.95222 -4.5158E-02 -5.1202E-03 4.6035E-03 9.98006E-05
TDDFPT| 4 1.95227 3.0732E-02 5.5029E-03 2.1076E-02 6.78669E-05
TDDFPT| 5 1.95911 4.1849E-04 -3.4382E-06 -5.9889E-03 1.72990E-06
TDDFPT| 6 1.95913 -2.5205E-04 -7.0178E-04 -6.6425E-03 2.14452E-06
TDDFPT| 7 2.01440 -7.5332E-01 7.6379E+00 -6.7623E-03 2.90710E+00
TDDFPT| 8 2.06607 3.6495E-05 3.8731E-04 -9.5517E-04 5.38416E-08
TDDFPT| 9 2.06609 2.7989E-04 -1.6367E-04 -2.6528E-03 3.61548E-07
TDDFPT| 10 2.14649 -5.3729E-06 1.2738E-06 8.0865E-05 3.45488E-10

With regards

Simil Thomas

You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/eEMybDTQKg8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759116E40EEDF86E28CC1D59FA59%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.

Reply all

Reply to author

Forward