Thanks Alex,
I tried your pseudos with hex structure, got 4.25 A eq. lattice constant,
E_KS = -5187.5720 eV instead of -2970.1303 eV/4.3 A with Siesta gga psfs,
I attach the cut-offs of these here:
Cd pb nrl nc
ATM 3.2.2 9-APR-07 Troullier-Martins
5s 2.00 r= 2.47/5p 0.00 r= 2.60/4d10.00 r= 2.41/4f 0.00 r= 2.41/
Se pb nrl nc
ATM 3.2.2 9-APR-07 Troullier-Martins
4s 2.00 r= 1.94/4p 4.00 r= 2.14/4d 0.00 r= 1.94/4f 0.00 r= 2.49/
Maybe they are not relativistic, not an expert in reading these psfs.
In any case, the above psf gives 0.85 eV gap for hexa-CdSe, yours a
little bit over 0.9 eV.
I guess discrepancy is related to poor 'convergence' for the numerical
orbitals
chosen automatically by siesta, or some bug in the installation.
However, tests passed all ok.
A puzzling thing: in the out file Siesta reports a divergent Charge
parameter
Z= 48 Mass= 112.41 Charge= 0.17977+309
which I don' t know if has any meaning.
Thanks
On 10/22/2012 10:20 PM, Oleksandr Voznyy wrote:
Hi Francesco,
have you tried to plot the bands? Do they behave OK?
You use 4x4x4 mesh with 0 shift, which is wrong, basically you reduce
the valuable amount of points in this way (should be 0.5 shift for even
number of points).
Anyhow, this won;t matter, if you plot the bands.
Here are my GGA pseudos with which I obtain 0.65eV in wurtzite.
pe Cadmium with 4d in valence
tm2 2.1
Cd pbr
0.0 0.0 0.0 0.0 0.0 0.0
8 4
5 0 2.00 0.00
5 1 0.00 0.00
4 2 10.00 0.00
4 3 0.00 0.00
2.43 2.55 1.50 3.00 1.10 0.70
#23456789012345678901234567890123456789012345678901234567890
pe Selenium
tm2 2.0
Se pbr
0.0 0.0 0.0 0.0 0.0 0.0
6 4
4 0 2.00 0.00
4 1 4.00 0.00
4 2 0.00 0.00
4 3 0.00 0.00
1.90 2.00 2.90 2.90 1.10 1.10
#2345678901234567890123456789012345678901234567890
Alex.
--
Francesco Tassone, Ph.D.
CNST IIT@POLIMI
via Pascoli 70/3, 20133 Milano
tel +39 02 2399 9880
