[CP2K:10589] Setting Electronic Temperature in Fermi-Dirac smearing

Lucas Lodeiro eluni... at gmail.com
Wed Aug 1 05:37:57 CEST 2018

Hi Kendric, the smearing is a contraption to can converge a SCF calculation
when the HOMO and LUMO are very near in energy (or the band gap is zero),
that is the case for metals or semimetals (because in the SCF procedure you
can get some metalic states in the middle of SCF or in some step of your
DM). The smearing contribute with entropic-termic energy, and that is a
problem, it can fix the SCF but gives some problem in energy, for that is
ideal to use a little value for it, but it cannot be too little neither,
because a too little value is laki that does not use the smearing. the best
way to get the better temperature is search (scan) from what value the
energy is converged (by some criteria ej 10E-4 Ha/atom), and then you can
know too what is the minimum value to does not have problems in the SCF
calculation, you can use one or some configurations. I use in metalics and
semimetalics surfaces a value among 0.01 to 0.001 Ha (it is the kb*T energy
of smearing, it is around 1000 to 100 K)

i hope i help you


2018-07-31 13:14 GMT-04:00 Kendric Roberts <robertsk... at gmail.com>:

> Hi I am new to cp2k and am wondering when using smearing for cell/geometry
> optimization how should one choose a temperature, which gives very good scf
> convergence, but will give inaccurate results? Thanks for the help.
> -Kendric Roberts
> --
> 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.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180731/492c0696/attachment.html>

More information about the CP2K-user mailing list