fremi energy, band gap

Message

yilmaz · #1 Post by **yilmaz** » Mon Nov 21, 2005 3:35 am

hi everbody
could somebody please tell me what am i doing wrong.
i did a 64-atom bulk calculation of diamond; both with USPP and PAW-PBE.
the final structure is perfectly good.
i dont understand two thinks about density of states.

1) some filled bands have positive energy, which is not supposed to be, as i know, for pure diamond. and fermi energy comes just above the higest energy sate (11 eV) like a metal. i have changed the EMAX EMIN but no change.

2) and the bang gap, i plotted from density of states using P4VASP, is around 4.5 eV much lower than known value. (diamond bang gap is 5.5 eV).

am i using some wrong parameter, i have seen some people calculated diamond bang gap close to 5.4 eV with VASP.

i really appreciate any comment.
thanks

#2 Post by **admin** » Wed Nov 23, 2005 4:34 pm

in 3D periodic systems, the E-zero is not defined up to a constant
offset. To ROUGHLY estimate the 'absolute' energy values, please
add the value of 'alpha+bet' written in OUTCAR.
for slab systems, the vacuum energy level can be detrmined from the
LOCPOT file, if the vacuum thickness is sufficient
2) DFT calculations usually give too small band gaps because DFT is a ground state theory.

atreyee · #3 Post by **atreyee** » Tue Mar 21, 2006 3:16 am

Dearv yilmaz

can you please let me know how dou you plot DOS using p4VASP? I have tried that one but i can't get it properly.

Thanks
Atretee

#4 Post by **admin** » Thu Mar 23, 2006 2:49 pm

this question should be addressed to the p4vasp-forum

wahyu · #5 Post by **wahyu** » Wed Apr 16, 2008 5:34 pm

The Efermi from the self consistent is not the same as that from the band structure calculation (non selfconsistent). Is the Efermi from the selfconsistent the correct one?

hatdau · #6 Post by **hatdau** » Wed Mar 10, 2010 10:02 pm

I have a similar question as wahyu:
I plotted band structure using Efermi from nonselfconsistent OUTCAR and it give non sense physics as explained in following:

1. My system has 24 and has indirect gap with negative band gap where top of valance band at G point and bottom of conducting band at X point.
2. So the fermi level should go through both valence band and conducting band.
3. However using Efermi from bandstructure calculation the Fermi lever only go through conducting band which gives extra electron to the system. -> not correct

if I use efermi from selfconsistent (which is smaller the the one above) it makes more sense.

Can someone explain to me and how can I do it right

I am using VASP 4.6.36

Best,
Danny
<span class='smallblacktext'>[ Edited Wed Mar 10 2010, 11:06PM ]</span>

panda · #7 Post by **panda** » Sat Mar 20, 2010 1:52 am

to get DOS you should be plotting DOSCAR file, to get bands you should be plotting EIGENVAL or PROCAR file. You have to check the Ef from the DOSCAR and normalize the band structure when you plot PROCAR or EIGENVAL with respect to this value.

coulomb · #8 Post by **coulomb** » Sat Mar 20, 2010 12:14 pm

I have a similar problem on fermi energy:
My DOS is not consistent with the bandstructure. Although there's diffenence in the value of E-fermi between DOS and
bandstructure. In my case,it's only 0.01eV ,far from about 0.2eV to make sense.

Any comment is appreciated.
Thanks!