Calculation of atom charges using the REPEAT method

[emerson p l]

Dear friends,

I'm a new CP2K user.

I optimized the structure of a MOF using PBE exchange-correlation functional with D3 dispersion correction, 2x2x2 k-point grid with 80 Ry of kinetic energy cutoff using the projector-augmented wave (PAW) method.

I did this procedure in another software because I didn't know the CP2K software until then. I saw that there is a way to calculate the charge of atoms using the REPEAT method in CP2K.

I have prepared a base script (testmof.inp) attached. Could someone with more experience kindly help me?

1 - I used EPS_SCF 1.0E-8 value. But SCF did not converge. One test would be to increase the value of the number of iterations in MAX_SCF, correct?
2 - Is there a specific BASIS_SET and POTENTIAL to calculate charges using the REPEAT method? I'm using GTH_POTENTIALS and BASIS_SET DZVP-MOLOPT-SR-GTH
3 - My system has ~28 atoms, I know it's difficult to specify, but what would be the average time range to perform an SCF and REPEAT chargers calculation? I'm using 48 colors.
4 - Is this part below ok to calculate the charge of atoms using the REPEAT method? I based it on this link (https://www.cp2k.org/_media/events:2016_user_meeting:cp2k-uk-2016-tong.pdf)

&PROPERTIES
  &RESP
    USE_REPEAT_METHOD T
    STRIDE 2
    &SPHERE_SAMPLING
      AUTO_RMAX_SCALE 100.0
      AUTO_RMIN_SCALE 1.0
      AUTO_VDW_RADII_TABLE UFF
    &END SPHERE_SAMPLING
    &PRINT
      &PROGRAM_RUN_INFO
      &END PROGRAM_RUN_INFO
    &END PRINT
  &END RESP
&END PROPERTIES

&END FORCE_EVAL

If my questions are outside the standard, let me know.

Thanks in advance.

Best,

Emerson

[Jürg Hutter]

Hi

you should first concentrate on the SCF convergence. If the system is not converging you have to adapt
the SCF parameters. Don't increase SCF cycles (100 should be enough).
Once you have converged, you can restart easily and test the REPEAT options.

You claim to do 2x2x2 k-points, but that is not specified in your input?

You are using the default GPW method (not GAPW which is similar to PAW). This is ok but you need in any
case to increase the Cutoff (400-600 Ry). Remember that this is the density cutoff, not the wavefunction cutoff as
in plane wave codes.

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of emerson p l <emers...@gmail.com>
Sent: Monday, September 9, 2024 7:48 PM
To: cp2k
Subject: [CP2K:20676] Calculation of atom charges using the REPEAT method

--
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/57604889-d73c-406b-a03d-dd664ae2f6d6n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/57604889-d73c-406b-a03d-dd664ae2f6d6n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[emerson p l]

Thank you very much for your answer, Jürg Hutter.

Your comments clarified me a lot.

I have a question: how is the GAPW method calculated? I found the commands in a tutorial (https://www.cp2k.org/exercises:2017_uzh_cp2k-tutorial:gapw):

Would it be the part described above?

In fact, I ended up not putting the k_points in the script I sent, for that, would it be the sequence below?

Again, thank you very much for your time and patience, Jürg Hutter.

Best,

Emerson

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/bY4TPprMBXA/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/ZR0P278MB07598A603576A645356DED589F9A2%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.

[Jürg Hutter]

Hi

easy answer, replace GPW by GAPW for the METHOD keyword.

regards
JH

PS: The charge methods are not available for GAPW

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of emerson p l <emers...@gmail.com>

Sent: Wednesday, September 11, 2024 5:38 PM
To: cp...@googlegroups.com
Subject: Re: [CP2K:20692] Calculation of atom charges using the REPEAT method

Thank you very much for your answer, Jürg Hutter.

Your comments clarified me a lot.

I have a question: how is the GAPW method calculated? I found the commands in a tutorial (https://www.cp2k.org/exercises:2017_uzh_cp2k-tutorial:gapw):

[image.png]

Would it be the part described above?

In fact, I ended up not putting the k_points in the script I sent, for that, would it be the sequence below?

[image.png]

Again, thank you very much for your time and patience, Jürg Hutter.

Best,

Emerson

Em ter., 10 de set. de 2024 às 05:09, Jürg Hutter <hut...@chem.uzh.ch<mailto:hut...@chem.uzh.ch>> escreveu:
Hi

you should first concentrate on the SCF convergence. If the system is not converging you have to adapt
the SCF parameters. Don't increase SCF cycles (100 should be enough).
Once you have converged, you can restart easily and test the REPEAT options.

You claim to do 2x2x2 k-points, but that is not specified in your input?

You are using the default GPW method (not GAPW which is similar to PAW). This is ok but you need in any
case to increase the Cutoff (400-600 Ry). Remember that this is the density cutoff, not the wavefunction cutoff as
in plane wave codes.

regards
JH

________________________________________

From: cp...@googlegroups.com<mailto:cp...@googlegroups.com> <cp...@googlegroups.com<mailto:cp...@googlegroups.com>> on behalf of emerson p l <emers...@gmail.com<mailto:emers...@gmail.com>>

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com><mailto:cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com>>.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/57604889-d73c-406b-a03d-dd664ae2f6d6n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/57604889-d73c-406b-a03d-dd664ae2f6d6n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
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/bY4TPprMBXA/unsubscribe.

To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k%2Bunsu...@googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB07598A603576A645356DED589F9A2%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.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<mailto:cp2k+uns...@googlegroups.com>.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAH%2BbL%3DeuuZebPV%3DzgCJ1WfHkwaqyU%3DwJnq6chrR7USAacYO9%3DQ%40mail.gmail.com<https://groups.google.com/d/msgid/cp2k/CAH%2BbL%3DeuuZebPV%3DzgCJ1WfHkwaqyU%3DwJnq6chrR7USAacYO9%3DQ%40mail.gmail.com?utm_medium=email&utm_source=footer>.