Implicit solvent simulation

[Lucas Lodeiro]

Hello all,

I want to simulate an heterogeneous system of organic molecules adsorbed on metal surfaces. Also I need to check different solvents which are the medium of the organic molecules, and I do not want to simulate the solvent explicitly. I never compute a calculation with implicit solvent in CP2K or a PBC code, just in ORCA and Gaussian with CPCM and PCM respectively. So I need some advice to do it here in CP2K.

I have access to CP2K7.1 and at the moment there is no possibility that newer versions will be included in the cluster. I read the manual for version 7.1, and I found two sections which deals with implicit solvent:

CP2K_INPUT/FORCE_EVAL/DFT/SCRF

CP2K_INPUT/FORCE_EVAL/DFT/SCCS

It seems SCCS is better than SCRF, but I am not sure.

Also, for SCCS it is necessary to select the variables one by one for the solvent... which is so different with respect to ORCA or Gaussian where they are tabulated. In CP2K, is there a tabulated table or keyword to call a particular solvent? or conversely, How can I extract the variables from other programs to use here?

Regards - Lucas Lodeiro

[Krack Matthias (PSI)]

Hi Lucas

 

SCRF allows only for a solute within a sphere which is not well suited for organic molecules adsorbed on a metal surface. SCCS could be applicable, but it might be difficult to get it converged for such systems. SCCS is not implemented for k points. In a first step, SCCS requires only the (experimental) dielectric constant of the solvent as an input. Alpha, beta and gamma are rather (semi-)empirical parameters retrieved somehow from experiments or other models. Alternatively, you can try the Environ plugin for QE.

 

HTH

 

Matthias

--
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/CAOFT4PLqi8ysiKHDTiy2zE2iw7ZoaKHD5xWZiPRRq78%2BAuU2dQ%40mail.gmail.com.

[Anton Lytvynenko]

Dear Lucas,

in addition to the suggestions by Krack Matthias, you may also want to consider parameters derived here: https://aip.scitation.org/doi/10.1063/1.5050938

Yours,

Anton

25.07.2022 16:25, Krack Matthias (PSI) пише:

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/FBC0EA11-01CD-4CF5-AE60-B9A346191158%40psi.ch.

[Lucas Lodeiro]

Hello Krack and Anton, thanks for your reply.

It is a fact that I am going to use SCCS. I am not familiar with Environ QE, so it is not the moment to learn and put it into practice.

About SCCS implementation, Is it implemented only for full periodic periodicity or can I use it with XZ periodicity? Also, Is it implemented to use Fermi-Dirac smearing?

Andreussi and Fatterbert-Gygi flavors of SCCS are similar?

The reference paper is very good, it has the data that I need. But I have some doubts about the informed values in Table S1 and the variables in CP2K SCCS. 

The first one is the alpha+gamma informed in the table's 4° column. Is it the same if I put this value with the ALPHA or GAMMA keyword or by separating the total value arbitrarily between both keywords?

The table S1 caption sets: delta_n = 2.0 , n_c = 7.6 and beta = -0.5 GPa. It is obvious the beta = BETA keyword, but I cannot deduce which keywords are the other two values, because the unique extra keyword is DELTA_RHO = 2.0E-5 by default. Is it implemented different the switching function in CP2K with respect to FHIaims/paper?

Regards - Lucas Lodeiro

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/448439ab-0229-31a1-89b2-009966770c58%40gmail.com.

[Anton Lytvynenko]

Dear Lucas,

27.07.2022 7:36, Lucas Lodeiro пише:

> The first one is the alpha+gamma informed in the table's 4° column. Is
> it the same if I put this value with the ALPHA or GAMMA keyword or by
> separating the total value arbitrarily between both keywords?

You separate them arbitrarily between the keywords. As gamma is, by
definition, the experimentally found surface tension of the solvent, so
personally I prefer to put the value of the surface tension to gamma and
derive alpha accordingly when I happen to try SCCS, but it should not
make any influence to the calculation if separate them somehow else.

Regarding other questions, you should wait for answers from more
qualified colleagues than me.

Yours,

Anton