[CP2K-user] Issues with tight binding methods

[Maxime Van den Bossche]

Hello again,

I had a closer look at the 3ob parameter set 
<http://www.dftb.org/parameters/download/3ob/3ob-3-1-cc/> and it seems like 
not all the
'bells and whistles' have been implemented in the DFTB module in CP2K
to make proper use of this parametrization:

- It is supposed to be used with 'full' third-order terms, not just the 
diagonal ones
(note: there are also the LMAX_DFTB and DFTB3_PARAM keywords which
need to be set under CP2K_INPUT/FORCE_EVAL/SUBSYS/KIND).

- Becke-Johnson damping of the D3 correction doesn't seem available in the
DFTB module (though it is available for e.g. DFT calculations).

The simplest way to deal with this, I think, is to use the DFTB+ code 
instead of
CP2K. As an example, I have attached a ZIP file with DFTB+ input files and 
the
final geometry I got out of it. There is a +20% expansion in the 
c-direction.
I have no real experience with DFTB calculations of molecular crystals, so 
I don't
know if this is a typical deviation or not. Where did the initial geometry 
come from?

/Maxime
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200227/3396cb29/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dftbplus_calculation.zip
Type: application/zip
Size: 6135 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200227/3396cb29/attachment.zip>