[CP2K:2177] Re: PDC

[Teodoro Laino]

Dear Alexey,

periodicity can be controlled in a very modular way in CP2K.
You have two places where you can specify the periodicity. The first one 
is in the CELL section.
&CELL
     PERIODIC ...
     ...
&END

controls the behavior of the periodicity for the atomic positions 
(essentially in the construction of the neighbor lists).

While:
&DFT
   &POISSON
        PERIODIC ...
        ...
   &END
&END

controls the periodicity of the charge distribution 
(electronic+nuclear). If you want a fully isolated system you need to
set consistently PERIODIC NONE in both places.
Why periodicity splitted in two places?
Because this  allows for more flexibility (when doing something more 
naive in non-standard QM/MM).

If you are new to calculations using plane-wave you may want to have a 
look at the literature on how to decouple
periodic images. CP2K has implemented few possibilities.. None of them 
is perfect.. But you can definitely choose the best
for your problem.

Cheers
Teo




agilemolecule wrote:
> Hi Flo,
>
> Thank you. The results really changed but still wrong.
> 28 C was 3.42   now 1.55
> Here is something totally wrong.
>
>   
>> For consitency you will have to add in the DFT section
>> &POISSON
>>     
> I do not understand this option. What is it?
>
>   
>> Another thing, sometimes FULL_ALL works better with smaller values of
>> the ENERGY_GAP see input refernce:
>>     
> I tried. Even a slight reduction ENERGY_GAP to 0.1 leads to no
> convergence SCF.
>
>   
>> Try once with the proper non periodic setting, maybe the charges get better.
>>     
> This is what confusing me. There are a few key words like PERIODIC
> NONE.
> How to setup periodic and, accordingly, non-periodic conditions
> correctly?
>
> Best wishes,
> Alexey
> >
>