[CP2K-user] [CP2K:15636] Re: Dipole Moment. Help wanted.

[hut... at chem.uzh.ch]

Hi

dipole moments in periodic systems are only defined modulo 2Pi/L
(for simple cubic systems). The value is calculated through a
complex Logarithm that can lead to results on different branches
with offsets. Have a look at the literature for more complete
discussions and more general formulas. 
The choice of reference point can lead to a reduced number of
"jumps" and easier post-processing. 

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter                         Phone : ++41 44 635 4491
Institut für Chemie C                FAX   : ++41 44 635 6838
Universität Zürich                   E-mail: hut... at chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp... at googlegroups.com wrote: -----
To: "cp2k" <cp... at googlegroups.com>
From: "Alexander Kazakov" 
Sent by: cp... at googlegroups.com
Date: 06/25/2021 03:58PM
Subject: [CP2K:15636] Re: Dipole Moment. Help wanted.

Yeap. Topology is here.


Best regards,
Alexander Kazakov

пятница, 25 июня 2021 г. в 15:55:24 UTC+2, Alexander Kazakov: 
Dear CP2K community,

I am simulating 64  molecules of water using DFT in NVT ensemble using PBC. I interested in  dipole moment values as a trajectory with MD step. 

Here is what I have so far:


You  can see 4 curves: 3 of them are components (x,y,z) and the last one is  the length of the vector.  All of them are functions of the MD step.

Let's  focus on curves that correspond to x and z components. You can see that  there is a kind of "jump"  at certain ranges. These "jump" actually  resemble a crossing kind of boundary condition. You can see how well it  is fit the blue curve on the top to the bottom (the same is applied for  the purple curve).
Personally, I think it is indeed a kind of boundary condition, however, I have no idea what 30 means.

Then I found these certain steps where it happens:

MD STEP
26
    X=  -19.18787840 Y=    5.85693184 Z=   29.55510907     Total=    35.72090144 
27
    X=  -19.18076162 Y=    5.52351367 Z=  -29.53176935     Total=    35.64458192

So  the z component apparently changed the sign. So I thought that  something horrible happened in the system. I plotted position atom  change as a function of atom ID:
(if the picture has very poor quality, please find enclosed pdf version)


Alright,  there is an interesting plot here. We see the different amplitude of  change through the atom IDs. So first 2/3 of the curve correspond to  hydrogen atoms displacement, whereas the last part corresponds to oxygen  atoms.

So back to questions. Could you tell me, why some  components of the dipole moment "jump" changing the sigh? Is it possible  to force dipole moment to stay on one side of the "boundary"? There are  no doubts that hydrogen atoms should move "actively" than oxygen atoms,  but I don't understand what has changed dramatically in the system.  Lastly, if you can give me a hint for number 30 in the boundary (see  figure), I would be thankful!

As because I check CP2K  6.1, 7.1, 8.1, and 8.2 and I got the same behaviour with jumps I assume  that I did something wrong or misunderstand something.
Please find enclosed FORCE_EVAL section of my script.

Additionally I am sending topology of the system.

If additionaly information is required, please let me know.

Best regards,
Alexander Kazakov


  
  -- 
 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... at googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/437a492d-a3df-4e73-bdb0-31dc9c1dd1f1n%40googlegroups.com.
 

[attachment "W64.xyz" removed by Jürg Hutter/at/UZH]