A question about the occupation number for each Wannier function within the GAPW method

[fyya...@gmail.com]

Dear CP2K developers, 

I have a question regarding the Wannier function within the GAPW method:

when you perform a unitary transformation of the ground state orbitals to obtain their maximally localized Wannier functions (MLWFs), which is required for all-electron NMR calculation (page 18 of your paper, CP2K: An electronic structure and molecular dynamics software package - Quickstep: Efficient and accurate electronic structure calculations, Thomas D. Kühne, et al.). 

The occupation number for each Wannier function within the GAPW method is 2 (for close-shell), and the same for the original delocalized Kohn-Sham orbitals. 

Am I correct? 

Thank you!

Fangyong

[fyya...@gmail.com]

My question is not about NMR, I just want to know whether each Wannier function in GAPW method has 2 electrons (for close-shell). 

Thanks!

Fangyong

[Fangyong Yan]

Dear CP2K developers, 

The partition of electron density into three parts in GAPW method, makes me to ask this question, whether the Wannier function corresponds to the correct molecular orbital, not the wavefunctions related to the planewave part, soft part or hard part during the electron density partition. If the Wannier function corresponds to the correct molecular orbital, then the electron number is 2 for close-shell. (here I talk about insulators, so no smearing is used here).

So Wannier function in GAPW has 2 electron each? Thanks!

Fangyong

--
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/c7cb97b9-bc36-468e-b31f-19cb5d4fae4cn%40googlegroups.com.

[Fangyong Yan]

I dont think cp2k uses the smoothed Gaussian function, which is used to construct the smooth electron density, to construct the Wannier function, because I doubt the smoothed Gaussian function is the correct wavefunction, thus, it should not be used. 

However, I still want to know whether the Wannier function is orbital transformed from the correct delocalized Kohn-Sham molecular orbital. 

Thanks!

Fangyong

[Marcella Iannuzzi]

Dear Fangyong

The correct unitary transformation of the wave functions is used and each orbital is occupied with two electrons.

Regards

Marcella

[Fangyong Yan]

Dear Marcella, 

Thank you very much for your clarification!!

I am using GAPW to calculate infrared spectra. For IR spectrum, some systems need a very large planewave cutoff, for such systems, GAPW is a better choice because you dont know how large your cutoff will be, because the energy may converge for smaller planewave cutoff, but the vibrational modes may be still not accurate. For such systems, GAPW is a better choice than GPW. 

So thanks a lot!! 

Best regards,

Fangyong

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/5f9f9232-595a-4a47-8d8f-ceaffdbd013en%40googlegroups.com.