[CP2K-user] [CP2K:12514] Re: Dielectric constant calculation in CP2K

[Vladimir Rybkin]

Dear Nt, 

you write:


When I multiplied the result with the total dipole moment I calculated 
above, the answer was far from 80.  I used another version of the equation 
that is more common in published literature without the epsilon_0 term in 
the denominator and without the "e" terms.  It gives me an answer much less 
than 1.

So, if I understand correctly, you used to formulas: one from the script 
and another one, that you find more proper. The latter gives less than 1. 
What is the number from the first one? 

Yours,

Vladimir


вторник, 19 ноября 2019 г., 16:51:52 UTC+1 пользователь Nt написал:
>
> Dear Vladimir:
>
> Thank you very much indeed for your reply.  I haven't managed to find 
> anyone who has performed dielectric constant calculations in CP2K so far 
> and my results have been so strange (always ~1 for water no matter which T 
> I use in my simulations) that I have been wondering if maybe CP2K is not 
> appropriate for this purpose for some reason or whether there's a problem 
> with the way I outputted my dipole moments files.  I just can't figure out 
> what is going wrong. 
>
> Thank you for your suggestion; I did write a script myself, but keep 
> getting values of ~1 (instead of ~80) at 20 C.  I have tried several 
> versions of the fluctuation expression I provided in my initial post, I 
> have checked my units repeatedly, and nothing helps.
>
> For the past few days I have been trying to understand the tutorial script 
> you referred to in case there's a hint there as to why my own script 
> doesn't generate reasonable results.  However, there are a few points in 
> the tutorial script that I don't understand, such as what the "weights" 
> parameter is and where it comes from (it looks like the dipole moment files 
> generated by CP2K and used in the tutorial script must have been formatted 
> differently to current-day moment files so that makes it tough to figure 
> out what is going on--i.e., exactly which data the tutorial script is 
> using).  It's also not possible to tell exactly what M_vec is made up of.  
> Thank you very much for the link to the paper, but unfortunately, it 
> doesn't provide the kind of nitty gritty I need to follow the example in 
> the tutorial and my own calculations.
>
> I have been hoping that someone here who has run these calculations 
> successfully might see my post and give me some hints or maybe even let me 
> know if CP2K is not the method to use for this.  Thanks a lot anyway.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20191120/b635ca0b/attachment.htm>