How do I calculate lda+U with negative U?
Posted: Mon Feb 04, 2008 10:59 pm
I am doing Te related problem with negative U for Te atom.
There is something strange with LDA+U.
Here is LDA+U related lines in INCAR file.
ISPIN=2
LDAUPRINT = 1
LDAU = T
LDAUTYPE = 2
LDAUL = 1
LDAUU = -4.0
LDAUJ = 0.0
I tested an isolated single Te atom in the supercell.
After the calculaton, the amount of Te-p orbital charge is 1.943.
This value is larger than when I use positive U=4.0, which is 1.855.
For the reference the charge of Te-p for U=0.0, is 1.890.
This test contrasts with what I was guessing.
The imposing of U gives the charge localization on atomic center.
So the larger U, the larger orbital charge related with LDAUL.
I expected that the negative U would give the delocalization of charge.
Am I doing something wrong?
Please help me..
Thanks.
Geunsik Lee
There is something strange with LDA+U.
Here is LDA+U related lines in INCAR file.
ISPIN=2
LDAUPRINT = 1
LDAU = T
LDAUTYPE = 2
LDAUL = 1
LDAUU = -4.0
LDAUJ = 0.0
I tested an isolated single Te atom in the supercell.
After the calculaton, the amount of Te-p orbital charge is 1.943.
This value is larger than when I use positive U=4.0, which is 1.855.
For the reference the charge of Te-p for U=0.0, is 1.890.
This test contrasts with what I was guessing.
The imposing of U gives the charge localization on atomic center.
So the larger U, the larger orbital charge related with LDAUL.
I expected that the negative U would give the delocalization of charge.
Am I doing something wrong?
Please help me..
Thanks.
Geunsik Lee