QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: voves on December 11, 2009, 14:13

Title: SE pseudopotentials (InSb)
Post by: voves on December 11, 2009, 14:13
Hi,
I tried different pseudopotential sets for fcc InSb crystal in ATK SE. All results show huge value of the bandgap (several eV). I appreciate any recommendation.

Jan
Title: Re: SE pseudopotentials (InSb)
Post by: Nordland on December 11, 2009, 16:15
There are no pseudopotential set in ATK-SE.

Are you taking about the basis set?
Title: Re: SE pseudopotentials (InSb)
Post by: Anders Blom on December 11, 2009, 23:10
I've seen similar results for GaAs. I think we need specific basis sets for the binary alloys; the current parameters are adapted for pure Ga and As (or In and Sb), not the alloy. This is how semi-empirical methods work - the results are only as good as the parameters.

If anyone knows of extended Huckel basis sets for the III-V alloys, we'll be very interested in testing them!

We can also try to fit our own set, but it's a bit of a task...
Title: Re: SE pseudopotentials (InSb)
Post by: Anders Blom on December 11, 2009, 23:27
I should correct my own statement a bit. There are already GaAs parameters from Cerda in ATK now, and also AlP, InAs, and some other III-V binaries. However, exactly InSb is missing...
Title: Re: SE pseudopotentials (InSb)
Post by: voves on December 14, 2009, 12:11
Could be possible to include parameters for InSb from:

Yajun Wei and Manijeh Razeghi
Modeling of type-II InAs/GaSb superlattices using an empirical tight-binding method
and interface engineering
PHYSICAL REVIEW B 69, 085316 (2004)

?
Title: Re: SE pseudopotentials (InSb)
Post by: voves on December 14, 2009, 12:15
or from:

Phys. Rev. B 57, 6493–6507
Empirical spds* tight-binding calculation for cubic semiconductors: General method and material parameters
Title: Re: SE pseudopotentials (InSb)
Post by: Anders Blom on December 14, 2009, 12:17
There are definitely plans to include tight-binding models too in ATK, but it's a bigger job than just taking some parameters, as we could (hopefully) do for the extended Huckel model, which is already implemented. I found some nice papers with parameters for InSb and several other structures, I'm trying to see how we could include them.