The ABINIT Users Mailing List ( CLOSED )

[mmikami@rc.m-kagaku.co.jp]

Hello, Boa tarde,  Horacio,

I am also curious about ab initio pseudopotential(PP) methods for such
rare-earth (RE) elements. I have not noticed any pseudopoential studies
for Pr (and my experience for RE PPs is too limited to disclose...) ,
but  I would like to share some of papers that appears interesting:

- N. Richard, S. Bernard, "Phase transition in cerium by pseudopotential
 methods", J. Alloys Compd. 323-324, 628 (2001)
 It seems that the shallow core electrons (5s, 5p) may be treated as
 valence electrons. (They are so localized that high ecut may be needed)
 The authors says, "For the 5s, 5p, and 5d orbitals, the cut-off radius 
is
 taken just before the outermost maximum of the corresponding wave
 functions ..." so that, r_c5s =1.29, r_c5p=1.49, r_c5d=1.99, and
 r_c4f=1.99 a.u.
 The choice of the local potential part may also be relevant...
 (in this paper, l=0 channel ..)

- M.Needels, M. Schluter, and M. Lannoo, "Erbium point defects in 
silicon"
  Phys. Rev. B47, 15533 (1993)
  If your system would be Pr-compounds (not Pr-metals), the electron
  configuration for Pr-PP making would be relevant. (Pr3+ or Pr4+ ?)
  Moreover, the treatment of f-electrons would be also important:
  the f-electrons should be explicitly treated as valence electrons,
  or they could be treated as core electrons ? (no-inclusion in calc.)
  The following paper may be interesting from this point, although it is
  based on Full-potential LMTO method:
  N.V. Skorodumova et al, "Electronic, bonding, and optical properties
  of CeO2 and Ce2O3 from first-principles", Phys. Rev. B64, 115108
  (2001)

- J. Laegsgaard and K. Stokbro, "Pseudopotential description of rare
  earths in oxides: The case of Er2Si2O7", Phys. Rev. B63, 075108 (2001)
  Although it is based on Vanderbilt-type PP, it indicates many tests on
  parameters for PP-making were needed ...

(Suggestions of other relevant papers would be highly appreciated !)

Obviously, many hurdles are to be overcome to reach physically
meaningful results. (Other hurdles might be; with/without non-linear
correction, LDAs/GGAs, ... )

In future, studies based on PAW might be expected to avoid ghost-states
etc (in principle), in my humble opinion.
Further follow-ups from experts on this forum ML may be expected.
Let us wait and see... Any further suggestions would be highly 
appreciated !

Regards,
Masayoshi

On 2003.8.12, at 00:14 Asia/Tokyo, hwlalves@ufsj.edu.br wrote:

> Is there anyone who can give me any hints about how to construct Rare 
> Earth
> (open f and d shells) Troullier-Martins pseudopotentials by using the 
> fhi98PP
> code? I have tried to construct the Pr ones including the semicore 4f 
> electrons
> but, I am in trouble in the definition of the core radii .... When I 
> use s or p channells, a lot of warnings concerning such definition 
> appears. As a consequence, a lot of ghost states are detected in the 
> pswatch program and I am not be able to correct them. Thanks in 
> advance.
>
> Horacio W. Leite Alves