Inconsistency between pymatgen output and website GUI for the electronic structure of material mp-18868

I used pymatgen to extract the data for the valence and conduction bands for material RbFe(MoO4)2 (mp-18868) and noticed a discrepancy between the pymatgen output and that on the website interactive plot.

The pymatgen output says the material is a direct band gap material, but the interactive electronic structure plot lists it as a indirect band gap semiconductor. Both also list different band gap energies.

The valence band is fairly flat, so I don’t know if it’s due to a rounding error or something along those lines. Is there any way to figure out which one is correct?

Hi @RachelOwen, welcome to the forum!

I see this inconsistency also, so we’ll have to investigate further and report back.

This is a spin-polarized material, and the band structure shows a set of bands for the spin-up and spin-down electrons. The interactive plot appears to be showing the band gap between the spin-up valence band maximum and spin-down conduction band minimum. Meanwhile, pymatgen is showing the band-gap which is from the spin-down VBM and spin-down CBM. While the spin-up and spin-down VBMs are very close in energy, it’s not clear why the order is flipped between the website and pymatgen.

As you say, the valence band is fairly flat regardless with only 5 meV between band edge energies at the Gamma and A k-points, so regardless of this bug whether the material is truly a direct-gap or indirect-gap material we probably can’t determine with confidence at this level of theory.