[gpaw-users] question regarding delta-scf in gpaw

[abhishek khetan]

Dear gpaw-users,

I am trying to calculate the homo->lumo  or homo->lumo+1 excitation energy
for a few big organic molecules (with about 120 atoms). I want to ask a few
questions concerning the tutorials. In the tutorial on the webpage its
written:
-----------------------------------------------------------------

# Obtain the pseudowavefunctions and projector overlaps of the# state
which is to be occupied. n=5,6 is the 2pix and 2piy orbitalsn =
5molecule = [0, 1]wf_u = [kpt.psit_nG[n] for kpt in
calc.wfs.kpt_u]p_uai = [dict([(molecule[a], P_ni[n]) for a, P_ni in
kpt.P_ani.items()])
         for kpt in calc.wfs.kpt_u]

-----------------------------------------------------------------

1) I want to understand how can I get the correct "n" for my system? My
molecule is very very large and its not easy to just guess into which
orbitals the excitations will take place. Is there a way to determine this
automatically? Can I just directly choose it to be the 'n' which is the
band index of the first unoccupied state ?
2) Whats the meaning of molecule  = [0,1]. Does this correspond to 'C' and
'O' in the tutorial? Does it mean I must use molecule = [0,1,2,3,4....120]
in my simulations ?

3) Also further, on the page it seems that 'bands:-1' is always used in
convergence. I wanted to understand is this has got anything to do with
leaving the last band of the occupies bands out of convergence for lumo? I
want to know this to understand, for e.g. I want the excitation energy to
the level above lumo, then should I use 'bands:-2'

thanks and best,
abhishek



-- 
MfG,
abhishek
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20160627/2151661d/attachment.html>