Print MO Energies with OT

[Matt W]

Hi,

this is a 'feature'. It is a result of the way OT works and is not ideal.

You can instead use the MO_CUBES 
<https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/PRINT/MO_CUBES.html> 
keyword to obtain MO energies. You can set NLUMO to calculate as many 
unoccupied states as you wish. By default it will also print cube files for 
NHOMO and NLUMO states - if you do not want these files, which can get 
large, set WRITE_CUBE to false.

Matt

On Monday, July 2, 2018 at 1:04:19 PM UTC+1, Martin Pimon wrote:
>
> Dear CP2K community,
>
> I am a relatively new CP2K user and currently face some problems 
> concerning the output of the molecular orbitals (i.e. eigenvectors) for my 
> system, an insulator.
>
> I managed to converge the SCF loop of a fairly large cell using the OT 
> method (please note that this is the *only *method where I could achieve 
> convergence).
> Now, in the &DFT section it is possible to write the molecular orbital 
> coefficients with the flag:
>
> &PRINT
>    &MO
>    ...
>    EIGENVALUES .TRUE.
>    EIGENVECTORS .TRUE.
>    ...
>    &END MO
> &END PRINT
>
> However, using the OT method, no eigenvalues are printed: They are 
> all 0.000000! This poses a problem, since I need to look at the molecular 
> orbitals for a specific band.
>
> Moreover, a band of interest to me is the first band above the band gap. 
> Is it even possible to view this band with OT, since I can't use ADDED_MOS? 
> I thought a way out would be to converge with OT, and then use some other 
> method in combination with an external restart of a .wfn file and use 
> ADDED_MOS. However, as mentioned earlier, other electronic SCF methods 
> diverge almost instantly, even with the starting condition of the OT result.
>
> I would really appreciate some help! I tried to keep this question as 
> general as possible, if you need more specific details please just ask.
> Thank you in advance for your answer!
>
> Best wishes!
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180702/507b0379/attachment.htm>