QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: ams_nanolab on February 6, 2013, 11:53

Title: Wrong bandstructure of MoS2 monolayer with new databse
Post by: ams_nanolab on February 6, 2013, 11:53
Dear Sir,

We recently upgraded our ATK to 12.8 version. While running Ext Huckel simulations for monolayer MoS2, I noticed that the band diagram is not accurate. It is showing monolayer MoS2 as an indirect gap material, whereas it should be direct. I checked with the bulk-configuration from an old nc file from a previous version 11.8.2, it is showing perfectly accurate bandstructure.

Is this some bug? or some error in the new database?

Attached are the bandstructures.
Title: Re: Wrong bandstructure of MoS2 monolayer with new databse
Post by: Anders Blom on February 6, 2013, 12:39
The thing which immediately comes to mind is that we changed the Huckel model to be non-selfconsistent by default. You could look into that. The parameters are unchanged.
Title: Re: Wrong bandstructure of MoS2 monolayer with new databse
Post by: ams_nanolab on February 6, 2013, 14:11
I am using SCF iterations in all my calculations, with Cerda Mo and Hoffman S basis.
Title: Re: Wrong bandstructure of MoS2 monolayer with new databse
Post by: Anders Blom on February 6, 2013, 14:14
Need to see input files to spot any differences...
Title: Re: Wrong bandstructure of MoS2 monolayer with new databse
Post by: baizq on February 18, 2013, 08:53
I am using SCF iterations in all my calculations, with Cerda Mo and Hoffman S basis.

Hi,

I think you need to relax the MoS2 monolayer before band structure calculation. I have met the same problem as yours before: I calculated (by vasp) the band structure with the model extracted from the ATK 12.8 database without relaxation. It showed an indirect gap(GAMMA-K). But after relaxation (fixing cell size as that in the database & optimizing the atomic positions), the gap became direct at the high symmetry point K.

baizq
Title: Re: Wrong bandstructure of MoS2 monolayer with new databse
Post by: ams_nanolab on February 18, 2013, 11:51
Yup. I think that's the case.