M06-2X DFT functional

[SUBHASISH MALLICK]

I want to use M06-2X DFT functional in conjunction with 6-31G* basis sets in cp2k version 6.1.0. In this purpose I used libxc library, where, only XC_MGGA_C_M06_2X functional is supported in cp2k version 6.1.0 and the  XC_MGGA_X_M06_2X and

XC_HYB_MGGA_XC_M06_2X are not supported in the current version of cp2k. May I ask you among the above three M06-2X functionals mentioned in libxc which is equivallent to the gaussian M06-2X and why other two functionals are not supported in this version? 

Is there any way to implement the HYB_MGGA functionals in cp2k version 6.1.0.

[Frederick Stein]

Dear Subhasish,

The DFT part of M06-2X should be XC_HYB_MGGA_XC_M06_2X. It is basically the combination of XC_MGGA_X_M06_2X and XC_MGGA_C_M06_2X. When you want to use M06-2X within CP2K, you have the following options:

1) Within 6.1.0 use XC_MGGA_X_M06_2X and XC_MGGA_C_M06_2X.

2) Within 6.1.0 adjust the code by either using the patches in https://groups.google.com/forum/#!searchin/cp2k/wb97x%7Csort:date/cp2k/ARya5uM-pA4/2wt5PgtkAQAJ by Stanislav or applying the patches from PRs 304 and 422 (https://github.com/cp2k/cp2k/pull/304, https://github.com/cp2k/cp2k/pull/422) and use XC_HYB_MGGA_XC_M06_2X.

3) Use the latest trunk from CP2K with XC_HYB_MGGA_XC_M06_2X. (The PRs from 2 have already been included.)

Please keep in mind that you need an additional HF section in all of these three cases.

Best regards,

Frederick

[Tianhua Wang]

Dear Frederick,

I also want to use the M06-2X functional in CP2K through the first option you provided above, because of the version (5.1) I'm using, and the &XC_FUNCTIONAL section in the input file is written as

&XC_FUNCTIONAL

  &LIBXC

    FUNCTIONAL XC_MGGA_X_M06_2X XC_MGGA_C_M06_2X

  &END LIBXC

&END XC_FUNCTIONAL

However, I'm hesitating at the keyword FRACTION in the &HF section. The M06-2X functional has 54% HF, and I wonder if I should write "FRACTION 0.54" in this section as

&HF

  FRACTION 0.54

  &SCREENING

    EPS_SCHWARZ 1.0E-6

    SCREEN_ON_INITIAL_P TRUE

  &END SCREENING

  &INTERACTION_POTENTIAL

    POTENTIAL_TYPE TRUNCATED

    CUTOFF_RADIUS 6.0

    T_C_G_DATA t_c_g.dat

  &END INTERACTION_POTENTIAL

  &MEMORY

    MAX_MEMORY 5000

    EPS_STORAGE_SCALING 0.1

  &END MEMORY

&END HF

Thank you for your attention to my consultation and looking forward to your reply.

Sincerely,

Tianhua Wang

在 2019年9月3日星期二 UTC+8下午8:15:53，Frederick Stein写道：

[Frederick Stein]

Dear Tianhua,

Either use FRACTION 0.54 with SCALE_COULOMB 1.0 (default value for SCALE_COULOMB) or FRACTION 1.0 with SCALE_COULOMB 0.54. Both should work. The second way is more convenient when you mix interaction potentials. Basically you can use every combination of SCALE_COULOMB and FRACTION as long as the product of their input values is 0.54.

Best,

Frederick

[Tianhua Wang]

Dear Frederick,

Thank you for your early reply and detailed explanation. My calculations run smoothly under this setting.

Sincerely,

Tianhua Wang

在 2019年10月16日星期三 UTC+8下午4:18:51，Frederick Stein写道：

[alejandro martinez]

Hi, Frederick Stein

What happens if instead of using the functional m06-2x you want to use m06-l or another functional?

El miércoles, 16 de octubre de 2019, 3:18:51 (UTC-5), Frederick Stein escribió:

Querido Tianhua

Utilice FRACTION 0.54 con SCALE_COULOMB 1.0 (valor predeterminado para SCALE_COULOMB) o FRACTION 1.0 con SCALE_COULOMB 0.54. Ambos deberían funcionar. La segunda forma es más conveniente cuando mezclas potenciales de interacción. Básicamente, puede usar todas las combinaciones de SCALE_COULOMB y FRACTION siempre que el producto de sus valores de entrada sea 0.54.

Mejor,

Frederick

El miércoles 16 de octubre de 2019 03:44:37 UTC + 2, Tianhua Wang escribió:

Querido Frederick

También quiero usar el M06-2X funcional en CP2K a través de la primera opción que proporcionó anteriormente, debido a la versión (5.1) que estoy usando, y la sección & XC_FUNCTIONAL en el archivo de entrada está escrita como

& XC_FUNCTIONAL

  & LIBXC

    FUNCIONAL XC_MGGA_X_M06_2X XC_MGGA_C_M06_2X

  & FIN LIBXC

& END XC_FUNCTIONAL

Sin embargo, estoy dudando de la palabra clave FRACTION en la sección & HF. El M06-2X funcional tiene 54% de HF, y me pregunto si debería escribir "FRACCIÓN 0.54" en esta sección como

Y HF

  FRACCIÓN 0.54

  &CRIBADO

    EPS_SCHWARZ 1.0E-6

    SCREEN_ON_INITIAL_P TRUE

  & FIN DE CRIBADO

  & POTENCIAL DE INTERACCIÓN

    POTENTIAL_TYPE TRUNCATED

    CUTOFF_RADIUS 6.0

    T_C_G_DATA t_c_g.dat

  & END INTERACTION_POTENTIAL

  &MEMORIA

    MAX_MEMORY 5000

    EPS_STORAGE_SCALING 0.1

  & FIN DE MEMORIA

& END HF

Gracias por su atención a mi consulta y esperamos su respuesta.

Sinceramente,

Tianhua Wang

El martes 3 de septiembre de 2019 UTC + 8 8:15:53 ​​PM, Frederick Stein escribió:

Querido Subhasish,

La parte DFT de M06-2X debe ser XC_HYB_MGGA_XC_M06_2X. Básicamente es la combinación de XC_MGGA_X_M06_2X y XC_MGGA_C_M06_2X. Cuando desee utilizar M06-2X dentro de CP2K, tiene las siguientes opciones:

1) En 6.1.0 use XC_MGGA_X_M06_2X y XC_MGGA_C_M06_2X.

2) En 6.1.0, ajuste el código utilizando los parches en  https://groups.google.com/ forum / #! Searchin / cp2k / wb97x% 7Csort: date / cp2k / ARya5uM-pA4 / 2wt5PgtkAQAJ  by Stanislav o aplicando los parches de PRs 304 y 422 ( https://github.com/cp2k/cp2k/ pull / 304 ,  https://github.com/ cp2k / cp2k / pull / 422 ) y use XC_HYB_MGGA_XC_M06_2X.

3) Use la última troncal de CP2K con XC_HYB_MGGA_XC_M06_2X. (Los RP de 2 ya se han incluido).

Tenga en cuenta que necesita una sección de HF adicional en estos tres casos.

Atentamente,

Frederick

El martes 3 de septiembre de 2019 12:59:46 UTC + 2 SUBHASISH MALLICK escribió:

Quiero usar M06-2X DFT funcional junto con conjuntos básicos 6-31G * en cp2k versión 6.1.0. En este propósito, utilicé la biblioteca libxc, donde solo se admite XC_MGGA_C_M06_2X funcional en cp2k versión 6.1.0 y XC_MGGA_X_M06_2X y

XC_HYB_MGGA_XC_M06_2X no son compatibles con la versión actual de cp2k. ¿Puedo preguntarle entre las tres funciones M06-2X mencionadas anteriormente en libxc, que es equivalente a la M06-2X gaussiana y por qué otras dos funciones no son compatibles con esta versión? 

¿Hay alguna forma de implementar los funcionales HYB_MGGA en cp2k versión 6.1.0?

[Frederick Stein]

Hi,

Then you have to use the appropriate Libxc functional, look for the original paper (or in the Libxc source code) for the Hartree-Fock fraction and adjust the Hartree-Fock section accordingly. Unfortunately, CP2K does not ask Libxc explicitly for the parameters of the Hartree-Fock section. Thus, you have to provide the Hartree-Fock section yourself.

Best,

Frederick

[alejandro martinez]

Thank you very much for the information, do you have an example where the HF section is added to the input file?

--
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/499c8742-bff8-4edd-9a12-880f178a6643%40googlegroups.com.

[Mohammad Alaraby]

Hi all,

Thanks Frederick for the useful tips.

I am failing to use M06-2X with any of the keywords you specified. I am using CP2K version 7.0 (Development Version).

With XC_HYB_MGGA_XC_M06_2X or the X or C alone, it doesn't recognize the functional. [wrong functional name]

any hints?

Regards,

Mohammad

On Tue, Nov 26, 2019 at 5:36 PM Travis <polla...@gmail.com> wrote:

Hi,

M06L is a local functional, there is no HF contribution. Just set the LIBXC section.

-T

On Tuesday, November 26, 2019 at 12:27:20 PM UTC-4, alejandro martinez wrote:

Thank you very much for the information, do you have an example where the HF section is added to the input file?

El mar., 26 de nov. de 2019 a la(s) 04:35, 'Frederick Stein' via cp2k (cp...@googlegroups.com) escribió:

Hi,

Then you have to use the appropriate Libxc functional, look for the original paper (or in the Libxc source code) for the Hartree-Fock fraction and adjust the Hartree-Fock section accordingly. Unfortunately, CP2K does not ask Libxc explicitly for the parameters of the Hartree-Fock section. Thus, you have to provide the Hartree-Fock section yourself.

Best,

Frederick

--
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 cp...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/499c8742-bff8-4edd-9a12-880f178a6643%40googlegroups.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 view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0ec7e090-47b5-4c31-b24f-d9af4d18e7a8%40googlegroups.com.

--

Mohammad Salem,
+1(587)-708-4501

[Frederick Stein]

Hi,

It seems, there is no XC_HYB_MGGA_XC_M06_2X keyword in 4.3.4. Instead, you need XC_HYB_MGGA_X_M06_2X and XC_MGGA_XC_M06_2X (+HF section). This combination runs on my machine with the development version. I will update the link with the list of available functionals in our manual.

Best,

Frederick

[Mark Forster]

Frederick

Is there an example input file for (say) water or a small organic molecule?

As an inexperienced user I need something that runs 'out of the box' as a starting point.

Mark

--

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/f3030d6f-e6c1-467e-b9ee-401cf8746043%40googlegroups.com.

[Mohammad Alaraby]

Hi,

XC_HYB_MGGA_X_M06_2X works fine for me.

XC_MGGA_XC_M06_2X is not recognized. (I also tried: XC_HYB_MGGA_C_M06_2X).

A second note: I receive a stability flag:  *** WARNING in xc/xc_libxc.F:185 :: XC_HYB_MGGA_X_M06_2X is not ***
 *** considered stable. Use with caution...

--

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/f3030d6f-e6c1-467e-b9ee-401cf8746043%40googlegroups.com.

--

Mohammad Salem,
+1(587)-708-4501

[Mohammad Alaraby]

Hi again,

I found this in the 4.3.4 libxc "#define XC_MGGA_C_M06_2X        236 /* M06-2X correlation functional from Minnesota         */.

Now my calculation is running fine with the following:

&XC
      &XC_FUNCTIONAL
        &LIBXC
          FUNCTIONAL XC_HYB_MGGA_X_M06_2X
        &END LIBXC
        &LIBXC
          FUNCTIONAL XC_MGGA_C_M06_2X
        &END LIBXC
      &END XC_FUNCTIONAL
      &HF
        &SCREENING
          EPS_SCHWARZ 1.0E-10
        &END
        &MEMORY
          MAX_MEMORY  5
        &END
        FRACTION 0.54
      &END

Is there anything that looks suspicious?

To Mark, I am attaching my input file for a water dimer. It's an ALMO calculation using GAPW and wavelet solver. You could start from there!

Cheers,

Mohammad

--

Mohammad Salem,
+1(587)-708-4501

[Frederick Stein]

Dear all,

The input section looks fine to me. I was using the same.

Basically, if you need some template input files, just have a look at our regtestsuite. It provides examples for the different functonalitites, too. For Libxc, you should have a look at the directory /your/cp2k/directory/tests/QS/regtest-libxc.

Sorry about the names. I meant XC_MGGA_C_M06_2X not XC_MGGA_XC_M06_2X! The functionals XC_MGGA_XC_M06_2X or XC_HYB_MGGA_XC_M06_2X do not exist in Libxc.

@Mohammad: What trunk version are you using? This warning should be absent in the latest trunk. We are now using a different Libxc flag for warnings.

Best,

Frederick

[alejandro martinez]

Thank you very much for the clarification

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAOsnjQ2CyOcG2xLURjhyX__0HA4pVT2-0qPgxXF7OC%3DE5WMCAg%40mail.gmail.com.