[CPMD-list] Re: Problems with Vanderbilt PPs for Fe, Cu, Y etc (fwd)

[Fernando Alvarez Ramirez]

---------- Forwarded message ----------
Date: Tue, 10 Sep 2002 10:08:34 +0200 (MEST)
From: Juerg Hutter <hutter at pci.unizh.ch>
To: cpmd-list at cpmd.org
Cc: Fernando Alvarez Ramirez <fer at ariel.iimatercu.unam.mx>
Subject: Re: Problems with Vanderbilt PPs for Fe, Cu, Y etc


On Mon, 9 Sep 2002, Fernando Alvarez Ramirez wrote:

>
> Hi Juerg
>
> I am trying to run the CPMD code using Vanderbilt pseudopotentials for Cu.
> I have down loaded the home page Vanderbilt pseudopotential generator
> uspp-734-2.tgz from his webpage
>
> http://www.physics.rutgers.edu/~dhv/uspp/
>
> Some pseudopotentials work without problems like Si, C, O, etc, but I
> have problems when I use other pseudopotential like Fe, Cu etc.
>
> I have put the typical input line
>
> *fe_ps.uspp NEWF BINARY TPSEU
> LMAX=D
>
> but the program gives me the next outlines
>
> !!!  ERROR  !!!   AP(ILP,IL,IK) = 28.03090*******  1  6  9
>
> PROGRAM STOPS IN SUBROUTINE AAINIT|  [PROC= 0]
>
> Do you know how to solve this problem?
>
>
> thaks in advance
>
>
> Fernando Alvarez Ramirez
>
Hi

This most likely related to one of the restrictions of the
implementation of USPP's in CPMD.

1) maximum L is d, no f functions possible
2) all L channels need the same number of projectors
3) the channels are given by the PP program, they can not
   be changed in the CPMD input, i.e. LMAX=D doesn't have
   any effect.

Last possibility: your USPP has only s,p,d channels and 2 channels
each, then this might be a new bug in CPMD.

Please reply to the mailing list

thank you

Juerg