[CP2K:10640] electrostatic decoupling

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

Hi

I am not sure, but think DDAPC suffers from the same problem.

regards

Juerg 
--------------------------------------------------------------
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: "Xiaoming Wang" 
Sent by: cp... at googlegroups.com
Date: 08/16/2018 04:48PM
Subject: Re: [CP2K:10640] electrostatic decoupling

Dear Prof. Hutter,

Thanks for your explanations and suggestion. I will read the 
corresponding papers. Btw, does the ddapc method solve
this problem? There is a 'e_decpl' decoupling energy in the 
cp_ddapc_apply_CD. Or it still suffers the same
problem?

Best,
Xiaoming


On Thursday, August 16, 2018 at 8:12:32 AM UTC-4, jgh wrote:Hi 
 
your problem is related to the ill-defined energy of a charge 
in a periodic system. 
Your energy 1 is calculated with a background charge to nuetralize 
the charge of your orbital. In energy 2, the isolated system, 
no such background charge is needed. 
 
If you want to get some idea how to attack this problem, I would 
suggest to read the vast literature on the calculation of 
charged defects in solids. Start with the recent work of Pasquarello. 
 
regards 
 
Juerg 
-------------------------------------------------------------- 
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: "Xiaoming Wang"  
Sent by: cp... at googlegroups.com 
Date: 08/15/2018 10:05PM 
Subject: [CP2K:10631] electrostatic decoupling 
 
Hello, 
 
I'd like to decouple the the Coulomb interaction between the electron of one specific state, say HOMO,  
and its periodic images, for a fully periodic DFT calculation. The interested charge density is localized. 
I have tried to use different poisson solvers, say MT or WAVELET, to achieve my goal. So first I extracted  
the the charge density from mo_coeff. Then called the poisson solver. 
 
pw_poisson_solve(poisson_env, orb_rho_g%pw, ener1, v_gspace1%pw)  
 
with poisson environment PERIODIC3D. Next I changed the poisson_env to MT0D, then called  poisson 
solver once more. 
 
pw_poisson_solve(poisson_env, orb_rho_g%pw, ener2, v_gspace2%pw) 
 
Finally, the decoupling energy is deltaE = ener1 - ener2. I thought deltaE should be a very small 
number, because the charge density of that state is quite localized and my unit cell is big enough for 
the MT solver. However, I got a very large deltaE 0.05 Ha. Also the value is negative, which means the  
Hartree energy is higher for the decoupled case. I cannot understand this, because I think the image  
interaction would increase the energy. So can anyone give some advice? 
 
Best, 
Xiaoming Wang   
  --  
 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+... at googlegroups.com. 
 To post to this group, send email to cp... at googlegroups.com. 
 Visit this group at https://groups.google.com/group/cp2k. 
 For more options, visit https://groups.google.com/d/optout. 
  
  
  -- 
 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 post to this group, send email to cp... at googlegroups.com.
 Visit this group at https://groups.google.com/group/cp2k.
 For more options, visit https://groups.google.com/d/optout.