[Wien] How to set U in LDA+U when computing the energy gap ofHgTe

[Gerhard Fecher]

If using +U, the total energy is not longer the ground state energy.
So take much care if you like to optimize the structure.

Ciao
Gerhard


-----Ursprüngliche Nachricht-----
Von: wien-bounces at zeus.theochem.tuwien.ac.at im Auftrag von Trond.Brudevoll at ffi.no
Gesendet: Di 17.10.2006 16:38
An: wien at zeus.theochem.tuwien.ac.at
Betreff: Re: [Wien] How to set U in LDA+U when computing the energy gap ofHgTe
 
If you start with U = 0, the original s-conduction band lies much lower than -0.3 eV, in fact it also lies below the split-off band at around -1 eV. So the flipping has already taken place. The bands you see combining at the top of the valence band are the flipped light hole band and the heavy hole band. These are combining because they are forced to do so by symmetry at k=0. Applying LDA+U to the Hg 5d orbitals will push the 5d levels down and force the s-conduction band up towards the level -0.3 eV. If you increase U from 0 it does not seem that much is happening at the valence band top until the s-conduction band has risen to approach the valence band top. If you set Ueff  somewhere around 7.8 eV, the s-conduction band will touch the degenerate heavy- and light hole bands at k=0, and if you rise U further, you can even see the s-conduction band rising above the valence band edge to give an ordinary band structure with a positive gap, and the light hole band will then flip down to its "normal" non-inverted position. Thus, an Ueff somewhat less than 7.8 eV on the Hg 5 d's will give you the -0.3 eV band gap. 

 

Note however, that (mis)using LDA+U in this way is fitting, and certainly not an ab-initio procedure, so avoid doing volume optimization this way.

 

Trond Brudevoll

 

________________________________

Fra: wien-bounces at zeus.theochem.tuwien.ac.at [mailto:wien-bounces at zeus.theochem.tuwien.ac.at] På vegne av ? ?
Sendt: 17. oktober 2006 13:40
Til: Wien at zeus.theochem.tuwien.ac.at
Emne: [Wien] How to set U in LDA+U when computing the energy gap of HgTe

 

Dear users:
      An article in PRB says: the energy gap of HgTe is -0.3  .I use LDA+U to plot the energy band through adjusting U.But I didn't find the overclapping of maximum valance band and the minimum conductivity band.They are just combining.I think the energy gap is zero.But not -0.3.   What could I do? The method of LDA+U is well for HgTe or HgSe?What'the value of U?
      Thank you for advance.

                                                                               Mary

  

________________________________

????????-3.5G??,20M??! <http://cn.mail.yahoo.com>  


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 4114 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20061018/0da53bf0/attachment.bin