ABINIT Discussion Forums

hi,

I am trying to study NV defects in diamond. Right now, I am just trying to
use a simple config and get qualitative results. I am trying to find the
(populated) localized vacancy states in the output and figure out how to
go about analyzing them in different configurations, how to choose cell size
and defect density etc.

Here is the simple config I am using:

ntypat 2
znucl 6 7
natom 7
typat 2 1 1 1 1 1 1

acell 3*10.26
xcart
2.56616257 2.56616257 2.56616257
0.0 5.13232514 5.13232514
2.56616257 7.69848771 7.69848771
5.13232514 0.0 5.13232514
7.69848771 2.56616257 7.69848771
5.13232514 5.13232514 0.0
7.69848771 7.69848771 2.56616257

[ I have tried different values for ecut/ngkpt et al.
I am using nsppol 1 and the values for occ in the
output look reasonable to me. I am using ixc = 1
and Troullier-Martins psp for C/N (just looking for
qualitative features to start).]

The vacancy is at the origin and the N atom is at the position
(0.25, 0.25, 0.25). In the real material, there are 6 populated states
that are localized at the defect site (i.e. 6 electrons). In my output
from this cell, there is no density at the origin/defect in the total
density. Also, when I analyze all the wavefunctions, I can find none
with significant density at the origin, even unoccupied valence bands.

I am wondering if I am misusing abinit? Is there some reason that
a simple LDA/pspnc completely misses this phenomenon? Is there
somewhere special to look for these localized states rather than in
the WFK file?


thanks,

matt

Some follow up info:

The states were there all along, but they are hard to
identify. One needs to have a fine enough k-mesh to
resolve the bands in the DoS. So far, I can only resolve
the states in the gap. I need to figure out how to run
with large nkpt to resolve the state(s) in the occupied
region in the DoS (I think).

matt