CDFT geometry optimization convergence issue

[Akanksha]

Dear CP2K community,

I am trying to run constrained DFT calculations in CP2K 2024.2 to calculate the electron transfer parameters of the following reaction in an implicit DMF environment:

[trans-stilbene]- + 2bromophenol  ------------> trans-stilbene + [2-bromophenol]-

1) I have observed that the SCF wavefunction optimization keeps on fluctuating, even though I am restarting my calculation from an already optimized wavefunction under similar conditions (but without any charge constraint).

2) In other similar reactions, I have observed that even if the wavefunction reaches optimization somehow - the charges do not reach the appropriate constraint target value. The iterations are either too slow to reach convergence or start to give a drastically wrong number of valence electrons.

Here is the CDFT section i am using:

                &CDFT
                    TYPE_OF_CONSTRAINT          HIRSHFELD
                    ATOMIC_CHARGES              TRUE
                    COUNTER                     0
                    STRENGTH                    ${CDFT_STRENGTH}
                    TARGET                      ${CDFT_TARGET}
                    &ATOM_GROUP
                        ATOMS                   ${CDFT_ATOMS}
                        COEFF                   ${CDFT_COEFF}
                        CONSTRAINT_TYPE CHARGE
                    &END ATOM_GROUP
                    &DUMMY_ATOMS
                        ATOMS                   ${CDFT_DUMMY}
                    &END DUMMY_ATOMS
                    &OUTER_SCF ON
                        TYPE                    CDFT_CONSTRAINT
                        EXTRAPOLATION_ORDER     2
                        MAX_SCF                 50
                        EPS_SCF                 1.0E-3
                        OPTIMIZER               BROYDEN
                        STEP_SIZE               -2.0
                        &CDFT_OPT ON
                            BROYDEN_TYPE        BT1_LS
                            MAX_LS              5
                            CONTINUE_LS
                            FACTOR_LS           0.5
                            JACOBIAN_STEP       1.0E-2
                            JACOBIAN_FREQ       1 1
                            JACOBIAN_TYPE       FD1
                            JACOBIAN_RESTART    .FALSE.
                        &END CDFT_OPT
                    &END OUTER_SCF
                    &HIRSHFELD_CONSTRAINT
                        SHAPE_FUNCTION      GAUSSIAN
                        GAUSSIAN_SHAPE      VDW
                    &END HIRSHFELD_CONSTRAINT
                    &PROGRAM_RUN_INFO ON
                        &EACH
                            QS_SCF 1
                        &END EACH
                        COMMON_ITERATION_LEVELS 2
                        ADD_LAST NUMERIC
                        FILENAME ./${NAME}
                    &END PROGRAM_RUN_INFO
                &END CDFT

Could you please help me figure out what is causing these issues? I have attached my input and output files for your reference.

I really appreciate all your time and help. Thank you,

Akanksha

P.S. This question is a repost as I was not very clear in my previous post and didn't get a response. Sorry about that!

[uca...@ucl.ac.uk]

Hi Akanksha,

It looks like you have changed a number of the CDFT keywords from the recommended values. Have you confirmed that your chosen settings work for other systems?

Otherwise I suggest you follow the recommended keywords at

https://manual.cp2k.org/trunk/methods/dft/constrained.html
https://www.cp2k.org/_media/events:2018_user_meeting:cp2k-uk-2018-holmberg.pdf

https://pubs.acs.org/doi/full/10.1021/acs.jctc.2c00284

https://github.com/cp2k/cp2k/blob/master/tests/QS/regtest-cdft-hirshfeld-3/hirshfeld_qs.inc

The most important changes are probably:

OPTIMIZER               BROYDEN  ->  NEWTON_LS

SHAPE_FUNCTION      GAUSSIAN   -> DENSITY

MAX_SCF 20  -> 60

I expect it would be best to remove SCCS and check that CDFT converges properly without it. I don't have any experience with SCCS, but I assume that it will make convergence more challenging.

Chris