Getting a larger band gap in cp2k than vasp results

[Jacob Schulze]

Hey all,

I am trying to replicate the results in cp2k from a paper (https://www.nature.com/articles/s42005-018-0009-4) where they calculate band structures and DOS of the cuprate LCO in Vasp. The results for my band structure don't quite look like theirs however. One of the big differences is that I am getting a band gap that is about twice as big as theirs. 

I am using the same functional as them, and have used the same kpoint mesh. Had hoped increasing the size of the basis set would help, but didn't help at all. Any advice would be greatly appreciated. I attached my input file below.

Thanks,

Jacob

[Franklin T Cerasoli]

Hello Jacob,

Do you implement a Hubbard U parameter?

I notice that the paper mentions that there are “beyond DFT” methods used, such as DFT+U. Hubbard functional is known to correct bandgap overestimation. 

-Frank Cerasoli

On Apr 27, 2022, at 8:06 PM, 'Jacob Schulze' via cp2k <cp...@googlegroups.com> wrote:

Hey all,

--
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/8986bd88-8076-42ee-84af-4a768c11cd86n%40googlegroups.com.

[Jacob Schulze]

Hey Frank,

Sorry about the late reply. I didn't implement a Hubbard U. Seems that cp2k doesn't support DFT+U when using a kpoint mesh though.

Jacob

[Franklin T Cerasoli]

In hindsight, DFT+U widens the gap. So, my suggestion may have been irrelevant.

It may be worth seeing what other “beyond DFT” methods are used. 

-Frank

On May 12, 2022, at 6:11 PM, 'Jacob Schulze' via cp2k <cp...@googlegroups.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0d425bc1-e7f2-4699-bb21-4d5a921c4661n%40googlegroups.com.

[Matt Watkins]

Hello everyone,

I am inclined to think that this is an actual issue - I could reproduce the reported large gap with SCAN - it persisted when running gamma point only in larger cell.

I haven't used SCAN much via the libxc interface - I know it has been tested quite extensively for water and organic systems but maybe not for heavy atoms like La.

Matt