Does transition metal carbide need DFT+U?

[Seunghwan_Kwon]

Hello,

I have researched about Hubbard U parameter of transition metal carbide or nitride

such as TiC, TiN, ZrC and so on.

However, there weren't any reference that calculated Hubbard U parameter of those materials.

I have read about transition metal carbide DFT papers and none of them indicated the Hubbard U parameter.

Meanwhile, there were plenty of reference about transition metal oxide U parameter.

So my question is, does transition metal carbide or nitride not need to use DFT+U method?

Thank you.

[alexey.tal]

Dear Seunghwan_Kwon ,

Did you see this paper Phys. Rev. B 79, 235126 (2009)?
There is an extensive discussion on the U value in TiN.

In principle, DFT+U was proposed to improve the description of systems with strongly correlated d and f electrons. However, whether these correlation effects play an important role depends on the compound and the properties of interest.

You might find the discussion in this thread useful.

[Seunghwan_Kwon]

I am doing DFT calculations for zirconium carbide and

I have read about ZrC DFT papers, but there weren't any

Hubbard U parameters indicated.

So, I am still confused that no one wrote zrc hubbard u parameter.

Also, is it fine not to use U parameter for lattice parameter calculations and use it for just electronic?

I calculated zrc u parameter using linear response before but the band gap is 0 so verification of it is difficult and the lattice parameter is overestimated than pure dft.

Thank you

[alexey.tal]

In transition metal oxides, the Hubbard correction is usually needed to correct the band gap or magnetic moments.
If you are interested in ZrC specifically, I would point out that ZrC is metallic and in metals the d-states are less localized and less correlated.
I would suggest that you resort to DFT+U if you get unsatisfactory results with DFT. Is that the case for you?