[CP2K:9842] SIC in qs_ks_utils.F

[Xiaoming Wang]

Thanks Juerg.

Unfortunately, the SIC methods with PBC correction (AD, MAURI) are only 
suitable to either isolated systems or doublet cases. My system is in 
singlet 
or triplet configurations with full PBC. So I think I need to go for Scaled 
PZ
with explicit orbitals anyway. I am planning to do some modifications of the
code to include this correction. Do I need to make any further changes of 
following lines if I want to add them to sic_explicit_orbitals?
---------------------------------------------------

      CALL cp_ddapc_apply_CD(qs_env, &
                             orb_rho_g, &
                             ener, &
                             v_hartree_gspace=work_v_gspace, &
                             calculate_forces=calculate_forces, &
                             Itype_of_density="SPIN")
-----------------------------------------------------

By the way,  since for my system, the unpaired electrons tend to be 
localized. 
So when calculating the self-interaction energy as  in
------------------------------------------------------
CALL pw_poisson_solve(poisson_env, orb_rho_g%pw, ener, work_v_gspace%pw)
------------------------------------------------------
Can I change the poisson_env to use solvers like MT or wavelet?

Best,
Xiaoming Wang

On Monday, January 8, 2018 at 11:40:50 AM UTC-5, jgh wrote:
>
> Hi 
>
> yes, I think you are right. This SIC method calculates the 
> electrostatic correction using the same Poisson solver as used 
> for the full density. For PBC this means that the G=0 term is 
> neglected and the SIC correction is calculated using a uniform 
> background charge. 
>
> The methods in "calc_v_sic_rspace" use a correction, see 
> "cp_ddapc_apply_CD" to avoid this problem. A similar correction 
> would have to be added to the methods in "sic_explicit_orbitals". 
>
> The following SIC functionals have a PBC correction 
> AD 
> MAURI_SPZ 
> MAURI_US 
>
> No PBC correction is done for 
> EXPLICIT_ORBITALS (scaled) Perdew-Zunger correction explicitly on a set of 
> orbitals. 
>
> best 
>
> 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: Xiaoming Wang 
> Sent by: cp... at googlegroups.com 
> Date: 01/08/2018 06:23AM 
> Subject: [CP2K:9842] SIC in qs_ks_utils.F 
>
> Hello, 
>
> In the lines 613 and 614 of the subroutine qs_ks_utils.F, key operations 
> of the SIC correction are shown: 
>  CALL pw_poisson_solve(poisson_env, orb_rho_g%pw, ener, work_v_gspace%pw) 
>   
> energy%hartree = energy%hartree-dft_control%sic_scaling_a*ener 
> The first line calculates the electrostatic energy of the unpaired 
> electron (self-interaction) and the second line subtracts this energy from 
> the total Hartree energy. For open boundary calculations, this is correct, 
> since the electrostatic energy calculated this way is the self-interaction 
> energy. However, for periodic boundary conditions, the electrostatic energy 
> calculated by the first line is not the self-interaction energy. Instead, 
> the interaction energy of the unpaired electron and its periodic images is 
> also included!! (Please correct me if I am wrong.) 
>  So it seems that to do a correct SIC calculation for PBC systems, one 
> needs to either add back the interaction energy of the unpaired electron 
> and its periodic images using perhaps Ewald summation method (not sure) or 
> correct the self-interaction term by changing a poisson solver. Is it 
> possible to calculate the Hartree energy using non-periodic poisson solver, 
> say multipole, for a periodic system? 
>
> Best, 
> Xiaoming 
>
>
>   
>   -- 
>  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. 
>   
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180108/04f7a44d/attachment.htm>