El 11/07/2005, a las 20:38, Oleksandr Voznyy escribió:
There seem to be an error in OUT file of ATOM -
in most places concerning the list of 'nl' values for
pseudopotential generation, the 'n' value is shown starting from 1,
instead of showing just valence electrons:
I would not say "in most places". It is just in one place, and for
information which is
quite technical. It can be considered a bug, but, as the correct
bookeeping of the
n values would take quite a lot of extra code, the solution is going
to be simply to
remove the fake "n" values.
As pseudopotential generation: Troullier-Martins
nl s eigenvalue rc cdrc delta
1s 0.5 -1.062777 1.496969 0.283853 -1.503458
2p -0.5 -0.395428 1.898289 0.308215 -1.059701
2p 0.5 -0.375457 1.898289 0.291133 -1.119899
3d -0.5 0.000000 2.499179 0.000000 -11.713099
3d 0.5 0.000000 2.499179 0.000000 -11.714532
4f -0.5 0.000000 3.968839 0.000000 -18.116540
4f 0.5 0.000000 3.968839 0.000000 -18.117588
And another question:
How can one be sure that the pseudopotential is norm-conserving
(i.e. that rc is not too small - too close to the node)?
or this error indicates exactly to problems with norm-conservation?:
error in gamfind (aka v0pp) - delta not found
stop parameter =863
The norm conservation is one of the constraints in the search of the
pseudopotential parameters (that
"delta" is one of them). If the program cannot find a set of
parameters, then you do not have a proper
pseudpotential, and either you are doing something wrong with the
configuration or have chosen a value of rc
which is too big or too small.
Alberto
Prof. Alberto Garcia (http://fisica.ehu.es/ag)
Dep. Fisica de la Materia Condensada, Facultad de Ciencia y Tecnologia
Universidad del Pais Vasco, Apartado 644, 48080 Bilbao, SPAIN
Mail: [EMAIL PROTECTED]; Tel: +34-946.012.490; Fax: +34-946.013.500
