[CP2K:7861] Including inner shell electrons in GPW calculations

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

Hi

the reason for the jumps in the energy vs. volume curve is that
you are attempting to do a constant cutoff calculation. However,
the number of grid points cannot change smoothly.
The common solution to this problem is to use a constant number 
of grid points. This is implemented in CP2K using a reference cell.

&CELL
   ....
   &CELL_REF
      ....

The reference cell is constant for all calculation and is usually
chosen slightly bigger than all the real cells used.
This is also the way you should perform constant pressure simulations.

Full convergence with the cutoff was discussed already by Marcella.
Only other point is to have REL_CUTOFF high enough.

best

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: Rustam 
Sent by: cp... at googlegroups.com
Date: 06/24/2016 07:47PM
Subject: [CP2K:7861] Including inner shell electrons in GPW calculations

Hi guys,
I need to include inner-shell electrons of Mg into my GPW calculations. That is the total number of electrons on magnesium must be 10, not just 2 valence 3s electrons.
This can be readily achieved using by using the following pseudopotential








&KIND Mg
  BASIS_SET DZVP-GTH
  POTENTIAL GTH-BLYP # 2s2 2p6 3s2 configuration&END
instead of typical
&KIND Mg  BASIS_SET DZVP-GTH  POTENTIAL GTH-BLYP-q2 # 3s2 &END
I figured (by trial and error) that including inner electrons requires significantly higher CUTOFF setting. The attached plot shows energy dependence on the cell size of MgH2.
A. 2-electron calculations are fine with 1000 Ry cutoffb. 10-electron calculations produce awful results unless the cutoff is above 8000 Ry.
QUESTION: Is it really necessary to use such a high cutoff to get smooth curves? Or are there any other settings one can tweak to get accurate results with lower (i.e. cheaper) cutoff?
Thanks!Rustam




-- 

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.



[attachment "Screen Shot 2016-06-24 at 1.15.35 PM.png" removed by Jürg Hutter/at/UZH]