[CP2K:6995] Mixing Method Recommendation

S Ling lingsa... at gmail.com
Tue Sep 15 11:10:35 UTC 2015


HKUST-1 has a reasonable gap, and you may use &OT instead of direct
diagonalization to converge SCF. Please try to replace the whole &SCF
subsection of your input with the following:

         EPS_SCF     1.000E-06
         MAX_SCF 20
         &OT T
            MINIMIZER  DIIS
         &END OT
         &OUTER_SCF T
            EPS_SCF     1.000E-06
            MAX_SCF  10
         &END OUTER_SCF
      &END SCF

Please also note HKUST-1 is a magnetic system. If you don't specify
MULTIPLICITY and/or &BS subsection in your input, your calculation will
most likely converge to a closed-shell singlet state, which is not the
ground state of HKUST-1. The triplet/ferromagnetic state, which is slightly
higher in energy than the antiferromagnetic ground state, can be obtained
by setting the MULTIPLICITY keyword properly. To get the antiferromagnetic
ground state, you will need to set the &BS subsection in &KIND section for
the two types of copper atoms.


On 15 September 2015 at 10:04, Chris Campbell <chris... at gmail.com> wrote:

> Hello CP2k Group,
> Should first state that I am new the DFT field.
> Had a query on which mixing procedure to selected to be applied to the
> density matrix.  I have selected the BROYDEN_MIXING_NEW method in my
> system but without much basis for this other than I observed the first
> version of the method being selected in the cp2k tutorial.
> Could someone please advise me on the suitability of this mixing method
> for my system or suggest possible reading material on this topic.
> The system I am modelling is the geometry optimization of the MOF HKUST-1,
> my cell is 156 atoms made up of copper, carbon, oxygen and hydrogen. I will
> be using the m06-l functional.
> Currently my single energy point calculations are converging.
> Any information would be greatly appreciated,
> Chris
> Also, apologies if this amount of information is insufficient for my query
> and will happily add additional if needed.
> --
> 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 http://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/20150915/b787709e/attachment.htm>

More information about the CP2K-user mailing list