People with experience know what U you should use (at least the range of
reasonable values). You can also calculate U (see our web-site), but the
calculated U values are often too large.
Since you do not tell us which system you have, we cannot help you. But for sure
bigger values of U than 8 eV
Dear Sir,
you are right. Experimental band gap for my compound is 2.0eV. The maximum
band gap on increasing U value for my calculation is 1.117eV at 10 eV. So,
should I increase U value beyond 10eV ? Untill now from my literature
review, the value of U is less than 10 eV.
Secondly, if suppose a
Hi,
The Hubbard U correction in LDA/GGA+U is to correct the failure of
LDA/GGA to describe localized d- or f-states. Therefore LDA/GGA+U is
able to give an accurate band gap only if the band gap of your system
happens to be of d-d or f-f character, which, however, not the case for
most
Dear Sir H. Jiang
I implemented GGA+U to my considered compound, keeping in mind that GGA+U
is valuable for compounds having strongly correlated system (like in d and
f orbitals). But My problem is regarding the selection of U value.
Experimental band gap for my compound is 2.0eV. The maximum
You can try larger U to see what happens, but values above 10 or 12 eV
start to be unphysical in solids. Also, the more the top of the valence
band or bottom of conduction band has an important contribution not coming
from the d or f electrons (e.g., oxygen-p), the less GGA+U will be
efficient to
thanks a lot Sir for your guidance.
On Sat, Jul 20, 2013 at 4:20 PM, t...@theochem.tuwien.ac.at wrote:
You can try larger U to see what happens, but values above 10 or 12 eV
start to be unphysical in solids. Also, the more the top of the valence
band or bottom of conduction band has an
Respected Users
I am applying GGA+U technique to considered compounds. In order to select
proper value of U for specific material, I am changing the value of U from
2eV to 10eV with a step of 1eV. I was expecting that on increasing U value
in the range 2-10eV, a value of U will appear where the
Hi,
I have never heard that the proper value of U is the one when the band gap
starts to decrease. Usually the value of U is chosen such that the
calculated properties (e.g., band gap or magnetic moment) agree with
experiment.
F. Tran
On Sat, 20 Jul 2013, Zaid wrote:
Respected Users
I am
8 matches
Mail list logo
