Total energy, geometry optimization, DFT+U, spin....
Moderator: bguster
-
albayrouni
- Posts: 9
- Joined: Mon Aug 02, 2010 7:23 pm
Post
by albayrouni » Mon Nov 08, 2010 11:59 am
Hello every body
I'm working on Fe and FeO using the LDA+U method in a paw framework. My problem is that the Fermi energy that I have obtained in my output is wrong , althought I have read the test t20 of v5 in abinit-5.8.4 that treats this problem in the case of NiO. I have attached my input so that you could help me.
-
mverstra
- Posts: 655
- Joined: Wed Aug 19, 2009 12:01 pm
Post
by mverstra » Sat Nov 20, 2010 12:10 am
Hi,
we'd need some more information on the output (+OS platform etc). Your input looks ok, but how can you tell the Fermi level is off? You should be able to count bands for the most part and get an answer = number of electrons. Similarly you can look at the integrated dos in both DS to see if it is correct.
Note that in LDA+U you may end up in a local (wrong) minimum if you start with the wrong dmatpu (though yours looks ok).
tolwfr 1.0d-8 is not very tight tolerance at all. This might be the issue if you end up not converging properly.
Matthieu
Matthieu Verstraete
University of Liege, Belgium
wFe3_1.in