OK, thanks for explanation.
But still have some questions just to make everything clear for myself.
1. ATOM generates absolutely equivalent pseudopotentials (in .psf) for
both 'up' and 'down' in case of isp (i.e. 's') calculations. Is that
correct?
2. 'r' generates different pseudopotentials (I saw in list archive that
it's an average (marked as Down in .psf) and deviation from it (marked
as Up)). But SIESTA reads only down part of it
(see atom.for in subroutine read_vps()
! Ionic pseudopotentials (Only 'down' used) )
And as you say
>'SpinPolarized' allows spin-up and spin-down components
> (of potential, charge density...) to be different.
So if SpinPolarized calculations need different potentails, how does
SIESTA treats them without reading that potentails?
3. 'r' doesn't generate Up component for l=0. What does that mean?
4. What are the units in .psf file?
I draw plots from .psf file and see that it is not pp(q) nor pp(r) as in
corresponding files.
Why it ends up with -6.0 value at high Radial grid values for Ga, As, Al
and -10.0 for N?
5.
> The spin-orbit interaction is an extra term in the electronic
> Hamiltonian that comes from the perturbative expansion of the
> relativistic energy, and thus only important for high-speed
> electrons (core electrons in heavy atoms).
How much is it important in terms of total energy and for MD
simulations, not just energy bands shape? I need to make calculations
for GaAs surface, so how reliable will they be without spin-orbit
interaction?
Sincerely,
Alexander Vozny
University of Sherbrooke, Canada
