[CP2K-user] [CP2K:18015] Re: Convergence failure and incorrect solvation energy with SCCS in cp2k-2022.1

[Wiko Ann]

Dear  Matthias

Can Constrained DFT work well with SCCS in cp2k?

Wiko
在2022年11月11日星期五 UTC+8 12:45:24<Wiko Ann> 写道：

> Dear  Matthias
>
> Thank you for your comment! 
> 1. I've tried to set "POISSON_SOLVER MT" and "PERIODIC none", but the 
> calculated solvation energy is different from the experimental value. For 
> example, The solvation energy of H3O+ is -108kcal/mol in experiment. With 
> "POISSON_SOLVER MT" and "PERIODIC none", it's -138.4 kcal/mol. But if I set 
> "POISSON_SOLVER PERIODIC" and "PERIODIC XYZ", The value is -107.5kcal/mol. 
> Another settings are all the same. The parameters RHO_MIN and RHO_MAX are 
> set to 0.0002 and 0.0035 respectively, which are suitable for cations 
> according to the article  [Self-consistent continuum solvation (SCCS): The 
> case of charged systems. J. Chem. Phys. 139, 214110 (2013) ].  
> The main discrepancy is from the energy in gas phase (the energies are 
> posted below). But I still don't understand why there's such a big 
> difference.
>
> PERIODIC
>   Total energy(without SCCS):                                             
>   -17.54128600763402
>   Total energy(with SCCS):                                               
> -17.71255174117677
> MT
>   Total energy(without SCCS):                                             
>   -17.49054517094935
>   Total energy(with SCCS):                                               
> -17.71112388917346
>
>
> 2. And another question arose. When I used SCCS to study a 6-atom Ag 
> cluster with H3O+ along side, there were different distance between H3O+ 
> and Ag cluster, also between O and H. I found the solvation energy of the 
> system is positive. And the energies between adjacent structures changed 
> rapidly (below). I suppose there are something settings still not 
> appropriate, maybe RHO_MIN and RHO_MAX. I'd love to know how to decide 
> RHO_MIN and RHO_MAX in such a system?  The input file is attached below. 
>
>  -239.471083048784806
>  -238.953979533671856
>  -238.946448820809621
>  -239.402925734069186
>  -239.260639793435388
>  -239.180932339281611
>  -239.014590347465059
>  -239.163178351021770
>  -239.194699170551246
>  -239.162140349715088
>
>
> 在2022年11月10日星期四 UTC+8 23:15:18<Matthias Krack> 写道：
>
>> Hi Wiko
>>
>>  
>>
>> After running checks using SCCS with POISSON_SOLVER MT and “PERIODIC 
>> none” for charged and uncharged molecules (solutes), I have correct my 
>> previous comment. CP2K/SCCS with MT seems to work fine which makes it the 
>> suggested setup especially for charged systems with SCCS. Note, that you 
>> need relatively high cutoff values (>800 Ry) and large cells for 
>> numerically accurate forces.
>>
>>  
>>
>> Sorry for confusing
>>
>>  
>>
>> Matthias
>>
>>  
>>
>> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of 
>> Krack Matthias (PSI) <matthia... at psi.ch>
>> *Date: *Tuesday, 8 November 2022 at 10:17
>> *To: *cp... at googlegroups.com <cp... at googlegroups.com>
>> *Subject: *Re: [CP2K:17994] Re: Convergence failure and incorrect 
>> solvation energy with SCCS in cp2k-2022.1
>>
>> Hi Wiko
>>
>>  
>>
>> 1.       That’s indeed a problem.
>>
>> 2.       That’s correct, SCCS is meant to work with POISSON_SOLVER 
>> periodic and PERIODIC xyz.
>>
>>  
>>
>> Best
>>
>>  
>>
>> Matthias
>>
>>  
>>
>> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of 
>> Wiko Ann <toda... at gmail.com>
>> *Date: *Tuesday, 8 November 2022 at 03:44
>> *To: *cp2k <cp... at googlegroups.com>
>> *Subject: *Re: [CP2K:17992] Re: Convergence failure and incorrect 
>> solvation energy with SCCS in cp2k-2022.1
>>
>> Dear Matthias: 
>>
>>  
>>
>> Thanks a lot for your reply. The electrostatic energy is posted below. 
>>
>> 1. I did use k points and smearing in the calculation of slab. It should 
>> be the problem. 
>>
>> 2. Did you mean if I want to obtain the solvation energy of H3O+, I can 
>> use “PSOLVER MT” and “PERIODIC NONE” to calculate cluster energy in gas 
>> phase, and then use the default setting when applying SCCS?
>>
>>  
>>
>> Regards
>>
>>  
>>
>> Wiko
>>
>>  
>>
>> 在2022年11月8日星期二 UTC+8 00:08:09<Matthias Krack> 写道：
>>
>> PS: Do not use with SCCS “PSOLVER MT” and “PERIODIC NONE” but the 
>> defaults.
>>
>>  
>>
>> M.
>>
>>  
>>
>> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of 
>> Krack Matthias (PSI) <matthia... at psi.ch>
>> *Date: *Monday, 7 November 2022 at 16:53
>> *To: *cp... at googlegroups.com <cp... at googlegroups.com>
>> *Subject: *Re: [CP2K:17989] Re: Convergence failure and incorrect 
>> solvation energy with SCCS in cp2k-2022.1
>>
>> Hi Wiko
>>
>>  
>>
>> 1.       Slab: SCCS in CP2K is not implemented for k points and not yet 
>> tested with smearing. I guess that you are using one of these features for 
>> the Ag(111) slab which might cause problems.
>>
>> 2.       H3O+: The results are sensible to the choice of the parameters 
>> RHO_MIN and RHO_MAX. Their values differ especially for neutral and ionic 
>> systems. What do you get for the electrostatic part of the solvation 
>> energy? 
>>
>>  
>>
>> Just as minor improvements (do not help for the issue above), I suggest 
>> to use for H3O+ the default diagonalization instead of OT (i.e. &OT off) 
>> and a smaller EPS_DEFAULT value (at least 1.0E-12 instead of 1.0E-10).
>>
>>  
>>
>> Best
>>
>>  
>>
>> Matthias
>>
>>  
>>
>> *From: *cp... at googlegroups.com <cp... at googlegroups.com> on behalf of 
>> Wiko Ann <toda... at gmail.com>
>> *Date: *Monday, 7 November 2022 at 14:56
>> *To: *cp2k <cp... at googlegroups.com>
>> *Subject: *[CP2K:17989] Re: Convergence failure and incorrect solvation 
>> energy with SCCS in cp2k-2022.1
>>
>> 在2022年11月7日星期一 UTC+8 21:51:15<Wiko Ann> 写道：
>>
>> Hello, everyone:
>>
>>  
>>
>> I was trying to apply solvation effect to the Ag(111)-H3O+ system, while 
>> I met several problems.  
>>
>>  
>>
>> 1. The first problem is the convergence of Ag(111) surface. I built a 
>> 4*4*3 Ag(111) slab. The system could hardly converge when implicit 
>> solvation model was used. While in gas phase, it converged successfully.  
>> Are there some tricks to make it converge?
>>
>>  
>>
>> 2. Another problem is more confusing. The solvation energy of H3O+ 
>> differs a lot from the experimental value as well as the values calculated 
>> by other solvation models such as VASPsol. I've tried to change the 
>> parameters in SCCS sections according to the article  [Self-consistent 
>> continuum solvation (SCCS): The case of charged systems. J. Chem. Phys. 
>> 139, 214110 (2013) ], and try to change the Poisson solver, but I still 
>> couldn't obtain a reasonable result. Why is there such a huge difference? 
>> Did I misuse SCCS section? 
>>
>>  
>>
>> I'm new to CP2K, and I didn't know whether I set somethings wrong in the 
>> input file or I missed some subtle but important details. I'll appreciate 
>> it if someone can give me some advice on these problems, since solvation is 
>> vital in predicting the energy of H3O+ and the interaction of metal surface 
>> and adsorbates. 
>>
>>  
>>
>> The input file is attached below.
>>
>>  
>>
>> Thanks a lot!
>>
>>  
>>
>> Wiko
>>
>> -- 
>> 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... at googlegroups.com.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/0420c53c-f0bb-49dc-85fa-2124328053d8n%40googlegroups.com 
>> <https://groups.google.com/d/msgid/cp2k/0420c53c-f0bb-49dc-85fa-2124328053d8n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>> -- 
>> 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... at googlegroups.com.
>>
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/ZRAP278MB08274F9BF3607A244882B3C3F43C9%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM 
>> <https://groups.google.com/d/msgid/cp2k/ZRAP278MB08274F9BF3607A244882B3C3F43C9%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer>
>> .
>>
>> -- 
>> 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... at googlegroups.com.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/ea085a09-02d3-425f-ac2c-cfcf2f1564e6n%40googlegroups.com 
>> <https://groups.google.com/d/msgid/cp2k/ea085a09-02d3-425f-ac2c-cfcf2f1564e6n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>> -- 
>> 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... at googlegroups.com.
>>
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/cp2k/ZRAP278MB0827413D164197DF78B9DC76F43F9%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM 
>> <https://groups.google.com/d/msgid/cp2k/ZRAP278MB0827413D164197DF78B9DC76F43F9%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM?utm_medium=email&utm_source=footer>
>> .
>>
>

-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/1930068a-0761-4db8-8e79-e10e7c7599f9n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20221111/b3153b1d/attachment-0001.htm>