Orthonormality of MOs in K-point calculations

[mshakiba.k...@gmail.com]

Dear CP2K developers and users,

Hi, I have a question about the orthonormality of the molecular orbitals in CP2K when we use K-points. When I use the keyword MO_ORTHONORMALITY in &SCF%PRINT section I get this message in my output:

"K-points: Maximum deviation from MO S-orthonormality not determined"

Does this mean that MOs in the K-points are not normalized? If they are not normalized will that make a problem in K-point calculations type like computing the band-structure? Another question, how can we make sure they are orthonormal?

Thank you very much in advance for your time.

Mohammad.

[hut...@chem.uzh.ch]

Hi

it exactly means what is written. The orthogonality of the
orbitals is not tested in k-point calculations. However,
they are orthogonal by construction, unless there is a major
problem in the calculation.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie E-mail: hut...@chem.uzh.ch
Universität Zürich
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: "cp2k" <cp...@googlegroups.com>
From: "mshakiba.k...@gmail.com"
Sent by: cp...@googlegroups.com
Date: 11/01/2021 12:50PM
Subject: [CP2K:16153] Orthonormality of MOs in K-point calculations

--
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...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/c10133dd-e03a-4ba3-ae22-5e5824fb6c28n%40googlegroups.com.

[mshakiba.k...@gmail.com]

Dear Professor Hutter,

Thank you very much for your response. I got one more question from your response: Can I manually check the orthonormality of the MOs for K-point calculations myself? I mean if I print out the coefficients (using DFT%PRINT%MO) and AO overlap matrix (from DFT%PRINT%AO_MATRICES%OVERLAP) and compute c^T*S^AO*c, then I should get (almost) 1.000 for the diagonal elements, right?

Sorry, I accidentally sent this message in private too.

Thank you again for your time.

Mohammad.

[hut...@chem.uzh.ch]

Hi

orthogonality for k-points has to be enforced in k space.
AO matrices will be printed in real space.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie E-mail: hut...@chem.uzh.ch
Universität Zürich
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: "cp2k" <cp...@googlegroups.com>
From: "mshakiba.k...@gmail.com"
Sent by: cp...@googlegroups.com

Date: 11/02/2021 10:07PM
Subject: Re: [CP2K:16163] Orthonormality of MOs in K-point calculations

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/2c58fbf3-b801-436c-905c-ad89924b7d7en%40googlegroups.com.

[mshakiba.k...@gmail.com]

Thank you very much for your clarification.