Bandgap and DOS contradicting in InSb nanowire

[Shan]

Hello sir,

I got contradicting band structure and density of states, when I simulated InSb nanowire using GGA-rPBE, 1*1*40 k-pts, DZDP.
the band gap shown semi-conducting with a bandgap of 0.14 eV. whereas the DOS shows high quantity of states on fermi level.

Then what is the nature of the InSb nanowire I simulated Semi-conducting or metallic?

I exported the DOS to a text file to check the fermi level, it shows fermi level is at -4eV. but the band gap from DOS plot at -4eV is much higher than 0.14 eV.

I am a bit confused after watching these results... Can u please suggest me what is the nature of the nanowire (metal or semi-conducting or semi-metallic) and why is it like this?

I attached the band structure, DOS and exported file of DOS for your reference....

Thank you sir.

[Jess Wellendorff]

It looks like there are bands crossing the Fermi level at k-points that are not sampled in your band structure calculation. The DOS sums up the electronic eigenvalue over the full Brillouin Zone, while the band structure plot only maps out certain paths in the BZ.

[Anders Blom]

I think you have only 1 kpoint in your DOS calculation, but as always, it's impossible to say unless you post your script...

[Shan]

Yes, I used 1 k-point in DOS which was the default value.

here I attached the script as you asked.

[Anders Blom]

1x1x1 is also the default for the SCF k-point sampling but you chose 1x1x41 for the system. Same should be the case for DOS (or even larger), of course, else you just sample the DOS at the Gamma point.

[Shan]

ok...  got it.

Used 5*5*5 k-point sampling and got accurate result of DOS.

Do I need to maintain same k-point as in the new calculator even in DOS and optical spectrum?. Is it necessary?
here i used 5*5*5 in dos and got accurate result.

please clarity sir.

[Anders Blom]

Weeeeell, you didn't quite get it

Let me try to be pedagogical, and ask you, why you chose 1x1x41 k-points for the calculator (quite specifically, why 1x1)? If you can answer that, then you should be able to figure out that the proper k-point sampling for the DOS is not 5x5x5. If not, I can explain it, but it's very important you understand these points, if you want to be successful in using ATK.

You don't need to use the same k-points for the calculator and the DOS, often it is necessary to use more points for the DOS to get really accurate results. But the logic of how the k-point sampling is chosen should be the same, and that's what I'm referring to above.

[Shan]

Thank you..
Dr. Anders.