Hello, I'm beginner in DFT and I'm working on a MOF in the following article.
https://www.nature.com/articles/s41467-022-34278-6#Sec10
I want to execute nscf, but, I don't know how I should determine the value of nbnd.
I've heard there is a way to get the value automatically, but I don't know how.
Can anyone tell me the way?
+) I just wrote my k points as below
K_POINTS (automatic)
1 1 1 0 0 0
And the authors of the article wrote in Methods as below
A 3 × 3 × 3 Monkhorst-Pack k-points grid was used for k-points sampling
Are they saying they wrote k points like this?
K_POINTS (automatic)
3 3 3 0 0 0
Dear Choi Seung Gun ,
First the second question. Yes,
K_POINTS (automatic)
3 3 3 0 0 0
is exactly what the Authors of that paper imply. Indeed,
K_POINTS (automatic)
1 1 1 0 0 0
should be used only with a-periodic systems (molecules, clusters, atoms) or with a quite large unit cell. Or for fast test calculations.
Quantum Espresso has a way to determine the number of bands: if you use
occupation = ‘smearing’
QE determines the minimum number of bands to accommodate all the valence electrons and adds a few more (minimum 4 more bands). Having a large number of bands is important for other calculations, for example, GW.
On the other hand, you should be able to determine the minimum number of bands by counting the number of valence electrons (the pseudopotential file for each atomic species can tell you how many electron that atom contributes). You need a number of bands to accommodate all the electron — remember spin.
Finally, if you want to do a nscf calculation, that means you have already done the scf calculation. In there, you should see the number of bands that have been used. A larger value than the one used in the scf should suffice to run the nscf. A larger value might add some extra states useful for the Density of States or further calculations.
Regards,
Roberto
- Badges
- Science topic
- Similar topics
- Control Systems Engineering
- Filtering