how to implement LDA+U modification in VASP for the the SnO2 system ?

Message

liulcc · #1 Post by **liulcc** » Tue Aug 14, 2012 3:02 am

The LDA+U and GGA+U have been implanted to deliberate wide-band-gap semiconductors. The literature [PRB 74,045202 (2006)] proposes a novel approach to calculate the parameter U. I want to discuss the SnO2 system in the same process, and the (Sn) d10s2p2 should be amend to d9s2p9 to survey the parameter U. So, I consult how to implement this modification in VASP?

zcapd90 · #2 Post by **zcapd90** » Wed Oct 24, 2012 1:05 am

Hi

I'm having related problems using GGA+U for SnO2 bulks. I'm trying to get a realistic band gap (which I believe is around ~3.5eV), but I can't get anything over 1eV.

I'm using Dudarev type +U, and just have tried having the onsite interaction on l=2, l=1, and l=0, for just the Sn element, for a range of U values between 1 and 6. No luck though. In fact as U increases the band gap gets smaller, which I didn't expect. The highest band gap I can get is using plain GGA, which I didn't expect either. Every time I test new values of LDAUU or LDAUL, I fully relaxed the unit cell, then do highly converged singlepoint calculation to see bandgap.

Any idea's how I can improve my SnO2 bandgap using VASP?

Thanks!