CNT - Optical Spectrum

[nicolapelagalli]

Dear Quantum ATK staffs,

I'm trying to simulate a (16, 0) cnt at different primitive cell sizes. I noticed that the band gap value considerably changes using a primitive cell sized 8nm x 8nm instead of a 4nm x 4nm one (from 0,52 eV to 0,63 eV) and the dielectric constant changes too. Does anyone of you have an explaination to this behaviour?
I'm using GGA-type DFT, BLYP functional and FHI pseudopotential. k-point sampling = 100 along the simmetry axis of the tube.

Thanks for your attention.

[Petr Khomyakov]

Optical Spectrum module is not really meant to compute dielectric properties of quasi-1D materials such as nanotubes. What it actually computes for a single nanotube (modeled with periodic boundary conditions) is an optical spectrum for an array of this nanotube plus all its periodic images. The optical spectrum and related dielectric constant are then dependent on the vacuum padding around the nanotube.

Regarding the band gap, could you attach your 2 python scripts and related logs? Which version of QuantumATK are you using? Did you try to increase vacuum padding thickness even more to see if the gap converges to some value?

[nicolapelagalli]

I'm using the 2018.06 version.
I didn't try to increase the cell size, since I thought that a 4nmx4nm cell size was enough for a (16,0) nanotube (in order to assume that the tubes do not interact). Now I'm going to increase the vacuum padding.


You can find my scripts and logs in the annex.

[Petr Khomyakov]

Just a general advice, it is a better practice to have working directory somewhere else, but not in the ./vnl folder.

Could you also attach the scripts used for the actual calculations?

Based on the scripts posted, I would suggest increasing the number of k-points from 20 to 100 in the band structure calculation, as well as explicitly specifying the route path along the nanotube axis, e.g., G, Z if the nanotube's axis is along the z-direction.

 

[nicolapelagalli]

I repeated all the simulations making sure all settings were the correct ones, now everything works. It was an error due to myself.

Thanks for your kindness and patience.