CP2K to study ORR mechanism for MOFs

DMITRII Drugov

unread,

Mar 25, 2022, 5:38:13 AM3/25/22

to cp2k

Dear CP2K users,

Could you please share your thoughts on correct CP2K settings to study oxygen reaction reaction on MOFs surface (186 atoms, C, H, O, N, Zn)? I used to conduct similar job on Quantum Espresso for to study HER on graphite slab (86 atoms) and it took ages to finish single energy calculation with soft potentials (ecutrho = 0.4000000000d+03

ecutwfc = 0.5000000000d+02). According to my experience at gamma point CP2K is much faster but I never added k-point mesh to CP2K.

Could you please let me know what my CP2K should be? Does CP2K allow to set up k-points for energy calculation? My system is a slab with vacuum of 2*x or y in z dimension.

Best regards,

Dmitrii

Anton Lytvynenko

unread,

Mar 25, 2022, 6:32:42 AM3/25/22

to cp...@googlegroups.com

Dear Dmitrii,

CP2k has implementation of k-points, but it could be incompatible with some methods and features.

But why do you need it? MOFs cells are typically large and have low symmetry.

Yours,

Anton

пт, 25 бер. 2022 р. о 10:38 DMITRII Drugov <dresear...@gmail.com> пише:

--
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/9a22b6dc-aca9-4476-816d-38096d65d2c1n%40googlegroups.com.

DMITRII Drugov

unread,

Mar 27, 2022, 2:47:01 AM3/27/22

to cp2k

Dear Anton,

Thank you for your reply.

I even did not think about low symmetry of MOFs. Thank you for this point.

I attach my input file below for cell_opt first, then I will do energy calculation.

Could you please have a look at my settings and express your opinion on its accuracy?

What do you think I need to change for energy calculations, expect of deeper EPS_SCF from 1.0E-6 to 1.0E-7 or -8?

Do I need to use spin polarisation for Co and Zn atoms when I calculate adsorption energy for ORR reaction coordinate, or I can reach desirably accuracy without it? My job is to compare different facets reactivity for ORR.

Best,

Dmitrii

&GLOBAL
PROJECT MOF_011_optimisation_larger_cell
RUN_TYPE CELL_OPT
PRINT_LEVEL MEDIUM
!EXTENDED_FFT_LENGTHS
&END GLOBAL
&FORCE_EVAL
STRESS_TENSOR ANALYTICAL
METHOD QS
&DFT
BASIS_SET_FILE_NAME BASIS_MOLOPT
POTENTIAL_FILE_NAME GTH_POTENTIALS
LSD
CHARGE 0
MULTIPLICITY 2
&MGRID
CUTOFF 800
NGRIDS 5
REL_CUTOFF 70
&END MGRID
&QS
EPS_DEFAULT 1.0E-14
!WF_INTERPOLATION ASPC
&END QS
&SCF
SCF_GUESS ATOMIC
EPS_SCF 1.0E-6
MAX_SCF 300
&OT
MINIMIZER CG
PRECONDITIONER FULL_KINETIC
# ENERGY_GAP 0.01
&END OT
&OUTER_SCF
EPS_SCF 1E-6
MAX_SCF 300
&END
!CHOLESKY INVERSE
!ADDED_MOS 100
!&SMEAR ON
! METHOD FERMI_DIRAC
! ELECTRONIC_TEMPERATURE [K] 1000
! &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 8.0
&END PAIR_POTENTIAL
&END vdW_POTENTIAL
&END XC
SURFACE_DIPOLE_CORRECTION T
SURF_DIP_DIR Z
&POISSON
PERIODIC xy
POISSON_SOLVER ANALYTIC
&END POISSON
&END DFT
&SUBSYS
&CELL
ABC 18.55406 16.35222 50
ALPHA_BETA_GAMMA 90.0 90.0 90.0
PERIODIC xy
SYMMETRY ORTHORHOMBIC
&END CELL
&COORD
C 2.69109 3.50908 2.31039
C 2.69109 12.34314 2.31039
C 0.51227 7.92611 3.89871
N 5.94837 3.44592 0.26929
N 1.20316 9.05092 3.62466
N 1.20316 6.8013 3.62466
N 5.94837 12.4063 0.26929
C 6.87686 10.39498 0
C 5.92646 4.78521 0.67203
C 2.43496 8.59814 3.1409
C 2.43496 7.25408 3.1409
C 6.87686 5.45724 0
C 5.92646 11.06701 0.67203
H 7.0639 9.48822 0.08937
H 5.34848 5.15102 1.30235
H 3.13933 9.1391 2.86446
H 3.13933 6.71312 2.86446
H 7.0639 6.364 0.08937
H 5.34848 10.7012 1.30235
C 2.26139 4.168 1.04379
C 2.26139 11.68422 1.04379
H 1.56208 4.79832 1.23086
H 2.83938 11.95708 0.32767
H 1.56208 11.0539 1.23086
H 2.83938 3.89514 0.32767
H 3.00789 4.62555 0.65058
H 2.30689 10.73099 1.14626
H 3.00789 11.22667 0.65058
H 2.30689 5.12123 1.14626
H 1.93449 3.50312 0.43134
H 1.35819 11.94159 0.83884
H 1.93449 12.3491 0.43134
H 1.35819 3.91063 0.83884
O 5.79166 7.92611 2.27822
Zn 4.81431 13.8838 0.89604
Zn 0.60158 10.90496 3.87488
C 6.93752 15.29816 2.4367
C 0.69089 10.88113 6.85373
N 3.68024 13.25705 2.37354
N 2.08951 12.13224 3.49836
N 0 11.15518 5.72892
N 5.94837 15.36131 1.52279
C 3.70215 13.65979 3.71283
C 2.75175 12.98777 4.38486
H 4.28013 14.29012 4.07864
H 6.48928 15.85222 5.64074
H 2.56471 13.07713 5.29162
C 7.36722 14.03155 3.09562
C 2.05244 10.27344 6.85373
H 2.83938 15.81052 1.4644
H 8.06653 14.21863 3.72594
H 2.36587 10.18765 5.95054
H 6.78923 13.31544 2.82276
H 6.62072 13.63835 3.55317
H 2.65402 10.83346 6.35805
H 7.32172 14.13403 4.04885
H 1.35819 15.82601 0.95323
H 1.93449 15.4185 1.36074
H 7.69412 13.4191 2.43074
H 2.01537 9.40362 6.44622
H 8.27042 13.82661 2.83825
O 3.83695 15.26599 6.85373
O 5.92646 12.31097 6.85373
Co 0.60158 4.94726 3.87488
Zn 4.81431 1.96842 0.89604
C 6.93752 0.55406 2.4367
C 0.69089 4.97109 6.85373
N 2.08951 3.71998 3.49836
N 3.68024 2.59517 2.37354
N 5.94837 0.49091 1.52279
N 0 4.69704 5.72892
C 2.75175 2.86445 4.38486
C 3.70215 2.19243 3.71283
H 2.56471 2.77509 5.29162
H 6.48928 0 5.64074
H 4.28013 1.5621 4.07864
C 7.36722 1.82067 3.09562
C 2.05244 5.57878 6.85373
H 2.83938 0.0417 1.4644
H 6.78923 2.53678 2.82276
H 2.36587 5.66457 5.95054
H 8.06653 1.63359 3.72594
H 7.32172 1.71819 4.04885
H 2.65402 5.01876 6.35805
H 6.62072 2.21387 3.55317
H 1.93449 0.43372 1.36074
H 1.35819 0.02621 0.95323
H 8.27042 2.02561 2.83825
H 2.01537 6.4486 6.44622
H 7.69412 2.43312 2.43074
O 3.83695 0.58623 6.85373
O 5.92646 3.54125 6.85373
C 17.36317 7.92611 3.89871
N 15.96455 4.57074 2.6476
N 14.37382 12.4063 1.52279
N 15.96455 11.28148 2.6476
N 14.37382 3.44592 1.52279
N 18.05406 9.05092 3.62466
N 18.05406 6.8013 3.62466
C 15.30231 5.45724 1.79207
C 14.35191 11.06701 1.12005
C 15.30231 10.39498 1.79207
C 14.35191 4.78521 1.12005
H 15.48935 6.364 1.70271
H 13.77393 10.7012 0.48972
H 15.48935 9.48822 1.70271
H 13.77393 5.15102 0.48972
C 10.68684 11.68422 0.74829
C 10.68684 4.168 0.74829
C 16.00162 7.92611 4.5064
H 9.98753 11.0539 0.56121
H 11.26483 3.89514 1.4644
H 15.68819 8.8293 4.59219
H 15.68819 7.02292 4.59219
H 9.98753 4.79832 0.56121
H 11.26483 11.95708 1.4644
H 11.43334 11.22667 1.14149
H 10.73234 5.12123 0.64581
H 15.40004 7.43043 3.94638
H 15.40004 8.42179 3.94638
H 11.43334 4.62555 1.14149
H 10.73234 10.73099 0.64581
H 10.35994 12.3491 1.36074
H 9.78364 3.91063 0.95323
H 16.03869 7.5186 5.37622
H 16.03869 8.33362 5.37622
H 10.35994 3.50312 1.36074
H 9.78364 11.94159 0.95323
O 12.1276 7.92611 2.46887
Zn 13.23976 13.8838 0.89604
Zn 17.45248 10.90496 3.87488
C 15.36297 12.46945 2.4367
C 11.11654 15.42446 2.31039
C 17.54179 10.88113 6.85373
N 14.37382 15.36131 0.26929
N 10.51496 15.63536 3.49836
N 16.8509 11.15518 5.72892
N 12.10569 14.51055 2.37354
C 11.1772 14.77984 4.38486
C 15.6191 11.63894 6.1817
C 12.1276 14.10781 3.71283
H 10.99016 14.69047 5.29162
H 14.91473 11.91538 5.64074
H 12.70558 13.47749 4.07864
C 15.79267 13.73605 3.09562
H 15.21468 14.45217 2.82276
H 16.49198 13.54898 3.72594
H 11.26483 15.81052 0.32767
H 15.74717 13.63358 4.04885
H 15.04617 14.12926 3.55317
H 16.69587 13.941 2.83825
H 16.11957 14.3485 2.43074
H 10.35994 15.4185 0.43134
H 9.78364 15.82601 0.83884
O 12.2624 12.50162 6.85373
O 14.35191 15.45663 6.85373
Zn 17.45248 4.94726 3.87488
Zn 13.23976 1.96842 0.89604
C 15.36297 3.38277 2.4367
C 11.11654 0.42776 2.31039
C 17.54179 4.97109 6.85373
N 14.37382 0.49091 0.26929
N 12.10569 1.34167 2.37354
N 16.8509 4.69704 5.72892
N 10.51496 0.21686 3.49836
C 12.1276 1.74441 3.71283
C 15.6191 4.21328 6.1817
C 11.1772 1.07238 4.38486
H 12.70558 2.37473 4.07864
H 14.91473 3.93684 5.64074
H 10.99016 1.16175 5.29162
C 15.79267 2.11617 3.09562
H 16.49198 2.30324 3.72594
H 15.21468 1.40005 2.82276
H 11.26483 0.0417 0.32767
H 15.04617 1.72296 3.55317
H 15.74717 2.21864 4.04885
H 16.11957 1.50372 2.43074
H 16.69587 1.91122 2.83825
H 9.78364 0.02621 0.83884
H 10.35994 0.43372 0.43134
O 12.2624 3.3506 6.85373
O 14.35191 0.39559 6.85373
&END COORD
&KIND Co
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q17
&END KIND
&KIND Zn
BASIS_SET DZVP-MOLOPT-SR-GTH
POTENTIAL GTH-PBE-q12
&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 H
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q1
&END KIND
&KIND N
BASIS_SET TZVP-MOLOPT-GTH
POTENTIAL GTH-PBE-q5
&END KIND
&END SUBSYS
&END FORCE_EVAL
&MOTION
&GEO_OPT
OPTIMIZER LBFGS
MAX_ITER 300
&END GEO_OPT
&CELL_OPT
EXTERNAL_PRESSURE [bar] 0.0
KEEP_ANGLES TRUE
KEEP_SYMMETRY TRUE
OPTIMIZER LBFGS
&END
&END MOTION
&END

MOF_011_geometry.inp

ZIF864309_011.xyz

Anton Lytvynenko

unread,

Apr 1, 2022, 8:46:11 AM4/1/22

to cp...@googlegroups.com

Dear Dmitrii,

how did you obtain the xyz? Did you just export it from CIF? It looks like it is full of disordered hydrogen atoms.

MULTIPLICITY 2

Are you sure?

What do you think I need to change for energy calculations, expect of deeper EPS_SCF from 1.0E-6 to 1.0E-7 or -8?

I don't know. AFAIR, the default is 1e-5, I used to lower it to 1e-6, while 1.0e-7 resulted in very slow convergence (I mean that the calculation reached the residues of ca 1e-6 and then started to turning around it, until "accidentally" reached 1e-7 -- I guess that some other parameters are to be tuned to avoid numeric noise). Anyway, you always could check the consistency w.r.t. to the considered parameter by yourself. The key point in your case is to get rid of numeric noise in gradients -- you can assess it via observing the course of your geometry optimization (if you see some random moves around the equilibrium rather than smooth convergence to it, this might be the case).

Do I need to use spin polarisation for Co and Zn atoms when I calculate adsorption energy for ORR reaction coordinate, or I can reach desirably accuracy without it? My job is to compare different facets reactivity for ORR.

I probably don't understand your question. If your system is paramagnetic, you have to use the spin-unrestricted formalism (LSD aka UKS) anyway.

I don't think the idea of crystallographic facets makes real sense in the case of MOFs. The active centers in the case of MOFs are either the metal ions with the coordinated atoms, or the fragments of ligands, these centers are unlikely to be affected if you change the facet.

Yours,

Anton

нд, 27 бер. 2022 р. о 08:47 DMITRII Drugov <dresear...@gmail.com> пише:

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/b4c22c8c-31db-4c3d-bac6-6fb057c6e3b8n%40googlegroups.com.

DMITRII Drugov

unread,

Apr 1, 2022, 10:09:43 AM4/1/22

to cp2k

Dear Anton,

Thank you for your reply.

I set up LSD and multiplicity 2 only because I directly create a slab 011 from cif of zif-8 by Avogadro software.

That gave me odd number of electrons. I also realised that 10-7 even 10-6 will result in a strong drift for convergence.

I will try 10-5 instead for correct slab structure.

I am a bit puzzled with slab termination as to create a facet of interest by direct translation of unit cell is a bit difficult.

https://www.youtube.com/watch?v=ywR5pWqbllE

It is difficult to find a right center of symmetry without creating unbounded atoms. Chemistry of termination is also different depending how deep you slice it. I attached output file from run with previous settings in case you are interested, but that structure is unreasonable. I also got weird metal-like dos because of odd number of electrons and poor Avogadro slab preparation job.

Would you be able to look at facets which I created for non-periodic calculations and tell how reasonable to study ORR on them, or it is better to stick to periodic cif like structure to do ORR studies?

Do you think when I do reaction coordinate studies according to CHE method, should I consider only metal as an absorption site or nitrogen atom of imidazolium ligand as well? Ideally I would chose only one adsorption center.

Best,

Dmitrii

zif8.cif

zif8_001.xyz

zif8_011.xyz

out

Anton Lytvynenko

unread,

Apr 7, 2022, 5:32:50 AM4/7/22

to cp...@googlegroups.com

Dear Dmitrii,

пт, 1 квіт. 2022 р. о 16:09 DMITRII Drugov <dresear...@gmail.com> пише:

I set up LSD and multiplicity 2 only because I directly create a slab 011 from cif of zif-8 by Avogadro software.

It is not the correct way. Indeed, you need an odd number of electrons, but it does not mean that you may put any even multiplicity here. Co2+ in this coordination environment has an exact preferred multiplicity, you can deduce it from the literature -- just search for complexes of Co(II) with the same environment in literature or refer to the "Comprehensive Coordination Chemistry" book. I am out of the topic for some time, but I guess the multiplicity will be 4 (you should check it by yourself anyway). As you have only one Co ion per a pretty large cell, this should be enough (a few close paramagnetic ions would complicate the task sending you directly into the realm of magnetochemistry). Anyway, if in doubt or there is no data, please employ the relax_multiplicity feature of CP2k -- at least for a preliminary run to deduce the correct multiplicity.

Regarding the slabs, unfortunately I can't help here at the moment. To get CP2k working correctly, you definitely need to generate all atoms needed to form the slab (without gaps), all of them must be unique with respect to translations of the slab supercell and all bonds must be correctly terminated (either with explicitly declared atom or with a replica obtained by the translations). Maybe there is a program that does it correctly, but I am afraid some homebrew code will be needed, and I am unable to produce it immediately (despite I also need it for some other tasks).

Do you think when I do reaction coordinate studies according to CHE method, should I consider only metal as an absorption site or nitrogen atom of imidazolium ligand as well? Ideally I would chose only one adsorption center.

What do you mean under CHE method?

Adsorption of what?

Yours,

Anton

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0fb91eb0-d9fd-4b69-b258-536e5c9a1d9fn%40googlegroups.com.

DMITRII Drugov

unread,

Apr 7, 2022, 10:22:59 AM4/7/22

to cp...@googlegroups.com

Dear Anton,

Thank you so much for your deep scientific reply. I will try to figure out the multiplicity issue which I really found challenging in DFT calculations. My background is more about electrochemistry and materials engineering than condensed matter physics. Regarding CHE, I meant computational hydrogen electrode. I need to calculate reaction coordinate for ORR according to CHE method proposed by Norskov and his colleagues, e.g., NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-28516-0. Therefore, I thought about which active site of the terminated surface I need to consider for adsorption of intermediates, metals or its ligands?

Do you think to estimate the right multiplicity I need to consider the correct number of unpaired electrons for the whole Co-doped ZIF-8 unit cell? Relax multiplicity can give me a rough idea for this number?

Best,

Dmitrii

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAGdLAUEs3a6LLX7rBf6BGHCvDvfRcVHWV%3Dkc9K4CSgPDqu3s9w%40mail.gmail.com.

Anton Lytvynenko

unread,

Apr 8, 2022, 5:00:14 AM4/8/22

to cp...@googlegroups.com

Dear Dmitrii,

чт, 7 квіт. 2022 р. о 16:22 DMITRII Drugov <dresear...@gmail.com> пише:

Therefore, I thought about which active site of the terminated surface I need to consider for adsorption of intermediates, metals or its ligands?

I don't know which list of intermediates you are about to consider. I think you should write the list according to the literature and the chemical intuition, then think about which kinds of ZIF atoms and/or ions could exhibit specific interactions with these species. It is possible that separate preliminary calculations are necessary, I am not so familiar with the field.

Do you think to estimate the right multiplicity I need to consider the correct number of unpaired electrons for the whole Co-doped ZIF-8 unit cell?

How many Co ions per unit cell do you model? I've seen 1 in the file, is it true?

Each of them can have either 3 or 1 unpaired electron, depending on the ligand field strength (as I've told, you have to consult the literature to figure out if Co(II) is low-spin or high-spin in this coordination environment). If they are not involved in the strong antiferromagnetic interaction [10.1098/rspa.1952.0181] (as you "dilute" your Co ions with diamagnetic Zn2+, they are definitely are not), the final number of the unpaired electrons is just a sum of their numbers at each Co2+ ion in the cell.

Relax multiplicity can give me a rough idea for this number?

Yes.

Anton

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAG2SAD-CYTcY_mh0JZ_%3DOET1fxEFU4PkcWq6CnDemttPZDmXyQ%40mail.gmail.com.

DMITRII Drugov

unread,

Apr 15, 2022, 10:13:06 PM4/15/22

to cp2k

Dear Anton,

Thank you for your reply and valuable prospective on electronic state of Co-doped ZIF-8 system.

I had to optimised Zn:Co ration to 9:1, so for that cell I used only 1 Co atom.

But I have to redo everything from the begging to build correct slabs.

Do you think DOS profile will be the same for 011 and 001 terminated cell built from the same cif file?

Best,

Dmitrii

Anton Lytvynenko

unread,

Apr 20, 2022, 11:29:35 AM4/20/22

to cp...@googlegroups.com

Dear Dmitrii,

сб, 16 квіт. 2022 р. о 04:13 DMITRII Drugov <dresear...@gmail.com> пише:

Do you think DOS profile will be the same for 011 and 001 terminated cell built from the same cif file?

Actually, I wonder if the idea of DOS profile is relevant for such systems -- rather than having bands, they have discrete electronic levels, which should not drastically change between the slabs.

Basically, you have two ways to terminate the MOF crystal -- either splitting of fragments bound by non-covalent interactions, or by cleavage of the coordination bonds (most probably, with addition of solvent or water molecules to the vacant positions in coordination spheres of metal ions, if any). A MOF crystal will behave much more close to a molecular crystal rather than the metal or covalent one.