[Pw_forum] fhi2upf.f90 and 'reasonable assumption' of nwfs=lmax_+1

[Brad Malone]

Hi, I'm trying to convert a Ga pseudopotential from FHI to UPF format. The
pseudopotential was generated from the following FHI input file:
--------------------
31.00  6  2  8  1.00 : z  nc  nv iexc rnlc
    1  0   2.00      : n  l   f
    2  0   2.00
    2  1   6.00
    3  0   2.00
    3  1   6.00
    3  2  10.00
    4  0   2.00
    4  1   1.00
2  t                 : lmax  s_pp_def
2  0.00  5.00  t     : lt  rct  et  s_pp_type
--------------------

I'm keeping the 3d states in the core, but treating it with a NLCC. Anyway,
when I try to convert this file it asks me to specify the wavefunctions and
occupancies. So I typed in

Wavefunction # 1: label, occupancy > 4S,2
Wavefunction # 2: label, occupancy > 4P,1

However, the code asks for one more wavefunction, under the 'reasonable
assumption' that nwfs = lmax_+1 (see below from fhi2upf.f90)

>  ! reasonable assumption
>   rel = 1
>   rcloc = 0.0d0
>   nwfs  = lmax_+1  <----
>

I imagine I'm missing something here, but why is this a reasonable
assumption? Is not reasonable for me to pseudize only the valence 4s and 4p
states in Gallium? Being unafraid of being called unreasonable, I changed
this line to be "nwfs=lmax_" and entered in only my 4s and 4p wavefunctions
(treating l=0 as the local component) and my pseudopotential was then
generated "successfully". I then tried out this pseudopotential along with
an As one that I also generated with FHI (and similarly hacked fhi2upf.f90
to convert it) on a GaAs system and received the following error:

>    from read_pseudo_nl : error #         1
>      Reading pseudo file (BETA)
>      from read_pseudo_nl : error #         1
>      Reading pseudo file (BETA)
>

So it looks like something isn't working out here, and I wonder if it's this
'reasonable assumption' that I don't understand. Thanks in advance for any
feedback or advice on this.

Best,
Brad Malone
UC Berkeley
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20091102/d6c591e8/attachment.html>