QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: yasheng on December 2, 2016, 16:39

Title: Huckel basis set
Post by: yasheng on December 2, 2016, 16:39

Hi All,

I have a question about Huckel basis set. is it ok to mix different Huckel basis sets in one calculation? for example, If I have copper and oxygen in my system, can I use Cerda.Copper and Muller.Oxygen ?

Thanks,
Yasheng

Title: Re: Huckel basis set
Post by: zh on December 5, 2016, 11:05

It isn't a good idea to mix them because their parameters values were optimized by means of different strategies.

Title: Re: Huckel basis set
Post by: Petr Khomyakov on December 5, 2016, 13:01

Technically, you can do it, but you have to clearly understand the reason why you want to do it, and verify if this basis set mixing gives any meaningful results for physical quantities calculated in your particular study.

Title: Re: Huckel basis set
Post by: yasheng on December 5, 2016, 13:09

Thanks for your reply.

 For oxygen, only Hoffmann.Oxygen Huckel Basis set available. If we want to do calculations for oxides, we are restrained to use Hoffmann basis sets only, if we cannot mix different basis sets.

In my experience, Hoffman basis sets are giving wrong band gap. (I mean totally unreasonable)

Title: Re: Huckel basis set
Post by: zh on December 7, 2016, 15:09

In principles, we can optimize the parameter values in the Huckel basis set by using the DFT calculated band structure and DOS of bulk materials as references.

Please see :
http://pubs.rsc.org/en/content/articlelanding/2012/dt/c2dt30298e#!divAbstract