QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: ramkrishna on July 25, 2013, 07:41

Title: Hybrid functional approximation
Post by: ramkrishna on July 25, 2013, 07:41

Dear Sir,
      Can you please tell me whether any Hybrid functional approximation is used in DFT calculation or not? If not, is it possible to implement it??

Regards
Ramkrishna

Title: Re: Hybrid functional approximation
Post by: zh on July 26, 2013, 03:01

No hybrid functional approximation is implemented at the current version.  The computational cost for the hybrid functional calculation of a bulk or molecular system is quite expensive. Since the number of atoms in a two-probe device system is much larger than a bulk or molecule system, it becomes quite challenging or un-feasible to perform the hybrid functional calculation for a two-probe device system. It may be more desirable to try the low-cost but moderately accurate functional, e.g., LDA+U, self-interaction correction, and so on.

Title: Re: Hybrid functional approximation
Post by: ramkrishna on July 26, 2013, 07:50

Dear Sir,
      Thank you for your reply. I am trying to understand the Hubbard U model and trying the input of LDA+U through the tutorial http://www.quantumwise.com/documents/tutorials/latest/NiO_LDA_U/index.html/ . It is written that "Set the Hubbard U for the Ni-3d orbital to 4.6 eV" however in the new version (12.8.2) I am not getting this option (I think the tutorial is based on the older version). Can you please let me know how to implement this U value in the current version?

What is the difference between the onsite and dual U? Please explain about these terms.

And if you discuss about the self-interaction correction and how to implement this in ATK calculations then it will also be helpful for me.

Regards
Ramkrishna