CO adsorbed on metal surface

1,048 views

Skip to first unread message

MR.Chem

unread,

Mar 17, 2013, 5:49:20 PM3/17/13

to cp...@googlegroups.com

Dear all,

I am studying the CO adsorbed on the metal surface, like Cu, Pd, Rh... However, my results are always larger than the experimental data, around 0.6eV. Does anyone have any experience about it?

Here is my code, can some expert help me to check it? Maybe something wrong with my code.

=================================================================================

&FORCE_EVAL

METHOD Quickstep

&DFT

CHARGE 0

LSD

POTENTIAL_FILE_NAME ../Basis_Set/GTH_POTENTIALS

BASIS_SET_FILE_NAME ../Basis_Set/BASIS_MOLOPT

&MGRID

CUTOFF 500

NGRIDS 7

# REL_CUTOFF 40

&END MGRID

&QS

EPS_DEFAULT 1.0E-10

EPS_GVG 1.0E-8

EPS_PGF_ORB 1.0E-8

EXTRAPOLATION PS

EXTRAPOLATION_ORDER 4 # find the best for your system

&DISTRIBUTION

SKIP_OPTIMIZATION TRUE

&END DISTRIBUTION

&END QS

&SCF

EPS_SCF 1.0E-8

MAX_SCF 200

EPS_SCF_HIST 1.0E-8

ADDED_MOS 500

CHOLESKY INVERSE

&SMEAR ON

METHOD FERMI_DIRAC

ELECTRONIC_TEMPERATURE [K] 3000

&END SMEAR

&DIAGONALIZATION

ALGORITHM STANDARD

&END DIAGONALIZATION

&MIXING

METHOD BROYDEN_MIXING

ALPHA 0.1

BETA 1.5

NBROYDEN 8

&END

&OUTER_SCF

EPS_SCF 1.0E-8

MAX_SCF 100

STEP_SIZE 0.1

EXTRAPOLATION_ORDER 4

&END OUTER_SCF

&END SCF

&XC

&XC_FUNCTIONAL PBE

&PBE

PARAMETRIZATION REVPBE

&END PBE

&END XC_FUNCTIONAL

&XC_GRID

XC_SMOOTH_RHO NN10

XC_DERIV SPLINE2_SMOOTH

&END XC_GRID

&vdW_POTENTIAL

DISPERSION_FUNCTIONAL PAIR_POTENTIAL

&PAIR_POTENTIAL

TYPE DFTD2

SCALING 1.25

REFERENCE_FUNCTIONAL PBE

R_CUTOFF 15.

&END PAIR_POTENTIAL

&END vdW_POTENTIAL

&END XC

&END DFT

&SUBSYS

&CELL

A 11.50460000 0.00000000 0.00000000

B -5.50230000 9.93026316 0.00000000

C 0.00000000 0.00000000 18.73890000

PERIODIC XYZ

&END CELL

&COORD

Pd 0.48065840 7.03842411 2.78730528

Pd 0.47028763 0.44875232 7.19243429

Pd 4.53327500 2.99486842 0.48000000

Pd -0.92756268 6.26067797 4.99255153

Pd 7.28442500 2.99486842 0.48000000

Pd 1.95832508 6.23995572 4.98929589

Pd 5.90885000 5.37743421 0.48000000

Pd 0.59872433 8.73424560 4.94485704

Pd -2.34460000 5.37743421 0.48000000

Pd 3.48202628 8.75574050 4.90377153

Pd 7.60058003 4.66725820 2.69914002

Pd 2.12951115 7.91373487 7.19590876

Pd -0.98640158 4.60804088 2.78207535

Pd 4.99805700 7.90489627 7.18006023

Pd -2.37121774 7.06798024 2.77804336

Pd 9.11351163 0.44812200 7.18150630

Pd 3.15770000 5.37743421 0.48000000

Pd -2.28848592 8.75288986 4.94848842

Pd 1.86796693 4.57784447 2.79012924

Pd -3.64617921 7.90460435 7.19218894

Pd 4.71963320 4.58012708 2.78282947

Pd -0.75815898 7.91750314 7.22011766

Pd 3.32828364 7.03761292 2.77891925

Pd 3.35784019 0.45975453 7.21713253

Pd 6.20941143 7.12736713 2.68973994

Pd 6.24673072 0.45867465 7.19321178

Pd -0.96902500 2.99486842 0.48000000

Pd 4.84897874 6.26179049 4.94562009

Pd 1.78212500 2.99486842 0.48000000

Pd 7.67975198 6.28516301 4.93895128

Pd 0.40655000 5.37743421 0.48000000

Pd 6.31889141 8.77579565 4.89678117

Pd 5.90885000 0.61230263 0.48000000

Pd 0.45589201 3.77441599 4.98640728

Pd -0.88579308 9.52461840 2.71360432

Pd -0.91478538 2.92497807 7.20923011

Pd 1.96179550 9.52218444 2.70458723

Pd 1.97006436 2.93242210 7.23312886

Pd 6.11012640 2.11464388 2.77204082

Pd 0.59930677 5.42011629 7.25046349

Pd 8.99428358 2.20331457 2.69500697

Pd 3.49212409 5.42106046 7.22291798

Pd -3.72017500 7.76000000 0.48000000

Pd 7.61529378 1.28270688 4.91196377

Pd -0.96902500 7.76000000 0.48000000

Pd 10.44730149 1.30931768 4.91246466

Pd 3.15770000 0.61230263 0.48000000

Pd 9.06475984 3.80294084 4.93990249

Pd 3.25324319 2.10925105 2.78044950

Pd -2.28679605 5.41093586 7.22290884

Pd 1.78212500 7.76000000 0.48000000

Pd 1.83892428 1.28295401 4.95856743

Pd 4.53327500 7.76000000 0.48000000

Pd 4.72711559 1.26487439 4.95221711

Pd 8.66000000 0.61230263 0.48000000

Pd 3.34270002 3.75714799 4.98323499

Pd 0.40655000 0.61230263 0.48000000

Pd 6.23251524 3.77764121 4.94405153

Pd 4.84358673 9.61900376 2.62576793

Pd 4.86285578 2.93105916 7.20861165

Pd -3.73917975 9.55256951 2.70716721

Pd 7.72991049 2.91981062 7.19559113

Pd 0.39747135 2.13690410 2.77102245

Pd 6.35957458 5.40832411 7.20891859

C 2.04578180 6.20000000 9.00000000

O 2.04578180 6.66763600 10.20000000

&END COORD

&KIND Pd

BASIS_SET DZVP-MOLOPT-SR-GTH

POTENTIAL GTH-PBE-q18

&END KIND

&KIND O

BASIS_SET DZVP-MOLOPT-SR-GTH

POTENTIAL GTH-PBE-q6

&END KIND

&KIND C

BASIS_SET DZVP-MOLOPT-SR-GTH

POTENTIAL GTH-PBE-q4

&END KIND

&END SUBSYS

&END FORCE_EVAL

&GLOBAL

PROJECT Pd-CO-111surface-hcp

RUN_TYPE GEO_OPT

PRINT_LEVEL low

&END GLOBAL

&MOTION

&GEO_OPT

MAX_ITER 1000

OPTIMIZER BFGS

TYPE MINIMIZATION # TRANSITION_STATE

&END GEO_OPT

&CONSTRAINT

&FIXED_ATOMS

LIST 53 51 45 43 55 33

LIST 47 57 5 3 29 27

List 7 17 31 9

&END

&END CONSTRAINT

&END MOTION

==========================================================================================

Thank you very much!!!!!!!!!!!

Ari Paavo Seitsonen

unread,

Mar 17, 2013, 6:34:42 PM3/17/13

to cp...@googlegroups.com

Dear MR.Chem,

A few issues catching my attention:

- When you say "your results are always larger than the experimental data", I presume that you mean the binding energy of the CO molecule, right? You should not look at the experimental number but compare to best theoretical DFT-GGA (+ dispersion correction, if available)

- Gamma point might not be enough in your case (One could study the magnitude of this by performing calculations with other DFT code and increase the k point sampling)

- There is the problem with the basis set superposition error with Gaussian basis sets. Did you try to do the counterpoise correction to your energies?

- LSD is not really needed in these systems (at least back then when I studied exactly these systems for my PhD thesis), but shouldn't harm either - only a factor of two in CPU time and memory

- FORCE_EVAL / DFT / QS / EPS_DEFAULT 1e-10 is quite large, I would go for 1e-14 or so for the production runs

- I don't have any idea of the EPS_GVG and EPS_PGF_ORB - do you have a reason to expect that the default values were insufficient or did you optimise these values?

- The broadening of 3000 K is very large: A typical value in the literature is rather 600 or 1200 K (0.05 or 0.1 eV); well, one should actually perform the extrapolation to zero temperature in the total energies, but let us first assume that this would be similar both in the calculation of the slab with and without the molecule and thus would cancel quite well when calculating the binding energy

- You use revPBE as the exchange-correlation functional but specify 'REFERENCE_FUNCTIONAL PBE' - why this? In addition you specify the 'SCALING' as well: First of all we usually use a value of 1.0 (where did you get 1.25?? Pure PBE should have a value smaller than one if I remember correctly), and 'REFERENCE_FUNCTIONAL' and 'SCALING' are used for the same thing, to specify the s6 coefficient in the DFT-D2 scheme. Please only insert one of these keywords

- Which lattice constant do you use, did you optimise it yourself or did you use the experimental value?

- Last but maybe not least, DFT-D2 is known to overestimate binding energies at transition metal surfaces

Here were a couple of suggestions to start with - good luck! :)

Greetings,

apsi

2013/3/17 MR.Chem <chemc...@gmail.com>

--
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 post to this group, send email to cp...@googlegroups.com.
Visit this group at http://groups.google.com/group/cp2k?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

--
-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
Ari P Seitsonen / Ari.P.S...@iki.fi / http://www.iki.fi/~apsi/
Physikalisch-Chemisches Institut der Universität Zürich
Tel: +41 44 63 55 44 97 / Mobile: +41 79 71 90 935

MR.Chem

unread,

Mar 17, 2013, 10:02:34 PM3/17/13

to cp...@googlegroups.com

Dear Seitsonen,

I am so glad you replied me. And those suggestions are really helpful. I am still learning how to use CP2K, so some part I am not very sure. I just took example from other people. Also, in my group, no one use CP2K, then, when I have some problem, the only way is searching on Google.So, anyway, Thank you very much!

Here is some question or answer for you.

1. Yes, I do mean the binding energy of the CO molecule. And I found some paper and references, there are both experimental data and theoretical results. My number are far far away, comparing no matter which one. SO, I believe there must be something wrong with my code.

2 When you mentioned the counterpoise correction, I know how to do it in Gauss, but no idea how to apply it on CP2K. Can you show me some examples? That maybe very helpful. Meanwhile, I will search it by myself.

3 About EPS_GVG and EPS_PGF_ORB , well, I just copy from other example. I will delete them.

4, For the 3000K, I am just found under that temp. the calculation will be done kinda fast, since right now, I am just testing the code. so, I only need some results ASAP. Also, I tested the calculations with using 0.05eV . The results are only small difference between F-D method and Energy_window method. For the future work, I will use the Energy_window method.

5. For the RPBE part, a friend just told me last week, He said the RPBE+D2 (scaling 1.25) works well for metal surface (based on one paper that he just found recently). and also gave me that code. So, I only need to keep one of 'REFERENCE_FUNCTIONAL' and 'SCALING', Is that right?

6 I generate the metal surface from the Material Studio, then optimize it.

Thank you again for your great suggestions. Bless you~~~~~~~~~

-Yao

Reply all

Reply to author

Forward

0 new messages