[Pw_forum] RRKJ pseudopotentials and raman calculations

[Paolo Giannozzi]

Merlin Meheut wrote:

> I would like to know if there is a fundamental difference between 
> rrkj pseudopotentials generated with the opium software and rrkj"us" 
> pseudopotentials, such as the carbon pseudopotential C.pbe-rrkjus.UPF 
> present on your website. Apparently, the former work for Raman 
> calculations, on the same level as trouillier-martins, whereas the 
> latter is considered as "ultrasoft" and therefore the calculation crashes

the calculation doesn't "crash", it stops with an error message.
C.pbe-rrkjus.UPF is not "considered" ultrasoft: it "is" ultrasoft,
obtained by ultrasoft pseudization of a hard norm-conserving
pseudopotential. The latter is (usually but not necessarily)
generated using the RRKJ (Rabe-Rappe-Kaxiras-Joanoupoulos)
algorithm. This is why they are called "rrkjus". Norm conserving
PP generated by ld1.x using the RRKJ pseudization should be more
or less the same as those generated by OPIUM (just a guess, I never
tried to compare them)

> For my use, these rrkj pseudopotentials are perfect

No problem, use them

P.
-- 
Paolo Giannozzi, Democritos and University of Udine, Italy