[Pw_forum] Pseudopotential troubles

Miguel Martínez Canales wmbmacam at lg.ehu.es
Mon Oct 9 11:27:30 CEST 2006


Dear Paolo,

First, thanks for your answer.

> you do not need to specify the radial grid and other optional variables

Indeed. It's just that I was puzzled to see that different radial grids 
allowed me different rc's for RRKJ pseudization. Due to the 3d orbital I am 
indeed more interested in an ultrasoft pseudopotential than in a 
Trouliers-Martins one.

> rcloc should be larger than any rc

Thanks for this one.

> you asked for a US PP but rc  for norm-conserving and ultrasoft
> pseudization are the same. 

Yes, I know. When I increase rcutus, ld1.x complains for having only one 
reference energy. I would add additional energies per channel but I don't 
see an obvious way to do this after reading INPIT_LD1. ¿Should I add an 
empty 3s orbital with the pseudization energy .neq. 0.00?

>> Another question I have is how to add additional projectors per orbital,
>> such as the ones used in Paolo Gianozzi's pbe pseudopotential.
> 
> ???

OK, it seems I didn't explain myself. I should have written projector per 
l. If you grep Projectors out of a pseudo file you get, for example:

$ grep Projector ~/pseudo/Ca.pbe-nsp-van.UPF
     4    6             Number of Wavefunctions, Number of Projectors

This is straightforward to do in David Vanderbilt's code, as you directly 
specify the number of beta functions the pseudopotential is going to have. 
Again, to do this, I would add another let's say 3s line to &inputp. But I 
fear this won't give the desired effect.

Thanks again for your time and attention,

Miguel



More information about the Pw_forum mailing list