How to shift density of states

[kcbhamu]

Dear Alex,
I am plotting bands and dos in one plot. I see, my DOS is slightly shifted towards the +ve y-axis. I mean, VBM and VBM for bands and DOSs are not falling on the same point. I want to shift the default DOSs with the difference in Fermi_energy of band and doss. How can I do it?

[ajjackson]

This is the correct behaviour if the DOS includes broadening. Do you see it with a tetrahedron-method DOS?

[kcbhamu]

I used the below INCARs

1. INCAR.scf
   ISMEAR = 0
   SIGMA = .05
   IBRION = -1

2. INCAR.HSE (used WAVECAR from step 1)
   IBRION = -1
   ISMEAR=-5
   SIGMA=0.01

[ajjackson]

That does look like a tetrahedron-method DOS, so this isn't due to smearing.

The Fermi energy of a non-self-consistent band structure calculation isn't really meaningful, but the energy levels should be consistent with a DOS calculated on the same density.

I wonder if it was shifted by self-consistency with more k-points. I see HSE mentioned so presumably these are self-consistent hybrid-DFT calculations with zero-weighted k-points for the band structure. Did you also zero-weight the DOS k-points? (Actually I'm not sure if Sumo really supports that, any idea @utf?) How well-converged is the self-consistent k-point mesh?

Another thing to check: which k-point has the VBM in the DOS? (You should be able to check with sumo-bandstats.) Is it definitely a point that appears in the band structure? (Usually it is Gamma and the answer is "yes" but there are exceptions...)

[kcbhamu]

I used SCF k-mesh with 0.022piAng^(-1) for SCF in both cases. Although, "zero-weighted k-points" were used for SCF and NSCF for band structure calculations while for DOS, a regular k-mesh was used.

Hole effective masses:
m_h: -4.394 | band 100 | [0.50, 0.00, 0.00] (M) -> [0.00, 0.00, 0.00] (\Gamma)
m_h: -6.520 | band 100 | [0.50, 0.00, 0.00] (M) -> [0.00, 0.00, 0.00] (\Gamma)

Electron effective masses:
m_e: 30.467 | band 101 | [0.00, 0.00, 0.00] (\Gamma) -> [0.50, 0.00, 0.00] (M)
m_e: 2.760 | band 101 | [0.00, 0.00, 0.00] (\Gamma) -> [0.50, 0.00, 0.00] (M)

For DOS, I used gamma-centered k-mesh
K-Spacing Value to Generate K-Mesh: 0.020
0
Gamma
4 4 1
0.0 0.0 0.0

for band structure, I used additional zero-weighted k-points.

I am attaching the plot.
band.pdf

[ajjackson]

Hmm, that certainly does look like a shift in the energy reference. Can you have a look at the reported Fermi energy of your band structure and DOS calculations and see if they differ by a similar amount?

I think bandplot relies on the energy levels being consistent between calculations. It's strange that they would be so different for such a flat band structure.

[kcbhamu]

I have checked the Fermi energy in both the cases, band &dos, and they differ by a small amount. I wanted to shift the doss by this difference from terminal.

…

On Mon, Dec 5, 2022, 8:13 PM Adam J. Jackson ***@***.***> wrote: Hmm, that certainly does look like a shift in the energy reference. Can you have a look at the reported Fermi energy of your band structure and DOS calculations and see if they differ by a similar amount? I think bandplot relies on the energy levels being consistent between calculations. It's strange that they would be so different for such a flat band structure. — Reply to this email directly, view it on GitHub <#190 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADZPLTYVP4NL53KGTT6QILLWLXE6BANCNFSM6AAAAAASOAJPLQ> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[ajjackson]

We do not currently provide a command-line argument to apply this shift, it is expected that the DOS and band-structure calculations will be consistent.

It is quite strange that the Fermi energies would be different from calculations with the same "active" k-points, are you sure they are both using the same Gamma-centered 4x4x1 mesh and the SCF is converged?

[mersadkhan]

This seems to be a bug that the command sumo-bandplot can not set the fermi energy to the same point that the band is set.

[ajjackson]

No, it is a feature request