[Pw_forum] ultra-soft psedudo potential
Lorenzo Paulatto
paulatto at sissa.it
Sat May 3 11:45:58 CEST 2008
On Sab, Maggio 3, 2008 00:10, alan chen wrote:
> What does it really mean? Is there a
> local potential if we set "lloc = -1"?
Dear Hanghui,
when generating the psedopotential you want to reproduce the scattering
properties of the all-electron core with a local potential and some
projectors functions. The projectors works at short radius, and compensate
the inability of a simple local potential to produce the desired solutions
for all energies and angular momenta.
Usually you can choose one special angular momentum and a reference energy
to be used as local channel; this has several implications: the local
channel has to be norm-conserving, cannot have a second reference energy,
cannot have ghost states.
So usually you want to choose the higher possible L for the local channel
to reduce the number of projectors (you have one for each M=-L,..,L).
Instead you often wind up taking a lower L in order to remove a ghost
state.
The point is that it is not a must to use a local channel, actually as
long as it matches the all-electron potential outside the pseudization
region, you can take *any* function as a local potential. As a drawback
you will have to use (at least) one projector for each value of L; this
can make you pseudopotential slower or less transferable. Sometimes it is
the only way to avoid the presence of ghost states.
--
Lorenzo Paulatto
SISSA & DEMOCRITOS (Trieste)
+39 040 3787 511
http://people.sissa.it/~paulatto/
----------------------------------------------------------------
SISSA Webmail https://webmail.sissa.it/
Powered by SquirrelMail http://www.squirrelmail.org/
More information about the Pw_forum
mailing list