QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: zhangguangping on December 30, 2022, 15:56

Title: Is there a way to get the value of hubbard U using the linear response approach
Post by: zhangguangping on December 30, 2022, 15:56

Dear all,

In ATK, one can use DFT+U method. However, the value of U is also very important. Is there a way or a tutorial for an ATK user to fix the value of U by using the linear response approach in DOI: 10.1103/PhysRevB.71.035105?

With best regards,

/Guang-Ping Zhang

Title: Re: Is there a way to get the value of hubbard U using the linear response approach
Post by: Anders Blom on January 18, 2023, 03:25

We have not considered that approach, rather (as most people do) the U parameters are viewed as fitting parameters. If you are looking at semiconductors, I would furthermore not recommend the Hubbard method, but rather look at DFT+1/2 or maybe meta-GGA (TB09). And in fact, these days even HSE06 is computationally reasonable, at least with LCAO basis sets.