Print MO Energies with OT

[Martin Pimon]

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!

[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 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

[Martin Pimon]

Thank you so much for your quick response. I've been running tests all day using your suggestion and the results seem promising. If I can find out more, I will post here again.