ABINIT Discussion Forums

Hello,

I've a few questions regarding DMFT Hub1 solver:

1). The spectral function can be obtained by setting prtdos=1.
However, the output energy range of the spectral functions is
always limited to -0.3~+0.3 Ha. Is there any possibility to
increase it? When U is large, the correlated orbital may goes
out of the box.

2). The file "SpFunc-cor_orb-realw" corresponds to the spectral
of the correlated orbitals and the file "SpFunc-realw" corresponds
to the total spectral (with correlated orbitals included), right?

3). I suppose the output files "SpFunc" already shift Ef=0, right?

4). When I assign dmftbandi and dmftbandf, in non-SOC case,
the calculated band are separated as spin-up and spin-down bands.
If so, the band index should be assigned by up or down or their
combined order?

Thanks,
Shu-Ting Pi
UC Davis

Hello,

1) Indeed you can change the two internal variables dmft_nwr and step in
file 65_paw/m_paw_dmft.F90 in order to increase the number of points and the range of energy
(you need to recompile the code afterwards).

2) Yes indeed.

3) Yes.

4) dmftbandi and dmftbandf are not dependent on the spin. In other words, the bands used are the bands numbered {dmftbandi,....,dmftbandf}
for bands corresponding to spin up and the bands numbered {dmftbandi,....,dmftbandf} for bands corresponding to spin dn.

best regards
Bernard