[Pw_forum] problems with variables in US pseudopotential generation

[Lorenzo Paulatto]

In data 10 maggio 2011 alle ore 00:45:35, Tram Bui  
<trambui at u.boisestate.edu> ha scritto:
> Dear Everyone,
>        In the pseudopotential generation input file, under &inputp, I  
> need to fill in the data for rcut (matching radius for norm-conserving)  
> and
> rcutus (matching radius for Ultra soft PP). I have looked at some  
> examples, and most of the time the rcut & rcutus are the same. So i'm  
> just
> wondering how one would have different rcut& rcutus?

Dear Tram Bui,
for norm-conserving pseudopotentials they are indeed the same. For  
ultrasoft pseudopotentials the set of rcut defines a very hard  
norm-conserving pseudo on top of which the ultrasoft one is build, but  
further pseudizing at rcutus.

As a consequence, rcut determine the hardness of the augmentation charge,  
while rcutus is linked to the hardness of the pseudo wavefunctions.  
Unfortunately the presence of ghost states often gives you little choice  
for rcut and rcutus.

As of chosing them, if you do not have a similar element to take  
inspiration, start with a reasonable norm-conserving single-projector  
pseudopotential and make it harder until 4*ecutwfc is the ecutrho you wish  
to use. Then double the number of projectors and change it to ultrasoft;  
keep increasing rcutus up from rcut until you reach a reasonable hardness.  
This procedure is difficult to apply for configurations with semicore in  
valence.

best regards


-- 
Lorenzo Paulatto IdR @ IMPMC/UPMC CNRS & Université Paris 6
phone: +33 (0)1 44275 084 / skype: paulatz
www:   http://www-int.impmc.upmc.fr/~paulatto/
mail:  23-24/4é14 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05