[QE-users] cannot project on zero atomic wavefunctions for part of elements - PDOS calculation using r2SCAN META-GGA and HGH pseudopotential

陈科帆 victor-chenkefan at 163.com
Wed Mar 20 05:43:37 CET 2024


Hi there,


I'm trying to obtain DOS and PDOS(both projected and partial) of VO2 using a hybrid functional calculation.


Here are my calculation setup:
input_dft='r2scan'
Pseudopotential: V.pbe-sp-hgh.UPF, O.pbe-hgh.UPF
exxdiv_treatment='gygi-baldereschi',
ecutvcut=0.8,
nqx1=6,nqx2=6,nqx3=6
occupations='smearing', (tetrahedra was used for 'nscf' calculation)
smearing='gaussian',


Yet, I found that the partial DOS only contains the information of oxygen atoms. If I sum up all the partial DOS of oxygen manually, I get the same numercial value as I find in the output Project DOS file of projwfc.x. I then tried DOS calculation of pure Vanadium. I got error message "Cannot project on zero atomic wavefunctions". Also, nothing is listed in the output file of dos.x for pure vanadium calculation. Maybe the reason is that this type of pseudopotential does not contain atomic orbitals so it's not possible to run any projection. 


My questions are:
1. For META-GGA calculations, is there any possible calculation setup that can be used to generate DOS and PDOS? Do I have to generate pseudopotential for myself? If META-GGA is not possible, which type of hybrid functional will be an efficient choice? I'm quite limited on the computational resources so I'm searching for any possible method to speed up my calculation. Also, I have to stick to hybrid functional since I want to compare the energies among different states of VO2 crystals, so DFT+U may not be appropriate for my calculation.
2. If QE cannot do such a calculation, is there any other software packages being able to run such a calculation?


Thank you very much!


Kefan
Department of Materials Science and Engineering
Sichuan University, Sichuan, China
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20240320/49e4890b/attachment.html>


More information about the users mailing list