A few days ago, I put this question on the ccl :
 "Does anyone know special ECP (for lanthanides and actinides)
  developped for DFT methods ?
  Do the Hartree-Fock's ECP work correctly with DFT ?"
 I thank you very much all who replied.
 Here are their answers :
         You can try:
 T.V. Russo, R.L. Martin, P.J. Hay,
 "Effective Core Potentials for DFT Calculations"
 J. Phys. Chem. 1995, 99, 17085-17087.
 Marek Sierka.
 "Politics consists in the art of taking votes from the poor and money
  from the rich under the pretext of protecting each from the other."
 Marek Sierka
 e-mail: mas.,at,
 We have used  ab initio ECPs (i.e. those obtained from a fit to Hartree-Fock
 or quasirelativistisc SCF all-electron data) in our DFT studies on transition
 metal systems. While the emphasis has been mainly on NMR-Properties, which
 adds some more aspects to the problem, we also optimized structures and looked
 at energies and vibrational frequencies. For the usual 'small-core' ECPs,
 the transferability of these
 ECPs in DFT applications is excellent, and I expect similar accuracy for
 lanthanide and actinide systems. Similar experience has also been made by others
 (e.g.: C. van W\"ullen Int. J. Quant. Chem., in press.
 T. V. Russo, R. L. Martin, P. J. Hay J. Phys. Chem. 1995, 99, 17085.
 See also the work of H. Schwarz and W. Koch and their groups in Berlin. As far
 as I remember, they also looked at some lanthanide systems using hybrid DFT/HF
 Some of our own results may be found, e.g. in:
 J. Am. Chem. Soc. 1995, 117, 1851; ibid 1995, 117, 8492.
 Chem. Phys. Lett. 1995, 235, 382.
 Chem. Eur. J. 1996, 2, 24.
 Chem. Eur. J. 1996, 2, 194.
 A particularly nice example is the unusual structure of W(CH3)6, where e.g.
 DFT and MP2 calcs. with ECPs give virtually identical results
 (M. Kaupp, JACS, in press).
 We have also tested to what extent ECPs for main-group atoms may be
 transferred (M. Kaupp, H.-J. Flad, A. K\"oster, H. Stoll, D. R. Salahub,
 manuscript in preparation). The answer depends on the core size. Thus,
 e.g., I do not recommend to transfer a 1-valence-electron ECP for Na or K,
 or a 3-valence-electron ECP for Ga, but smaller cores, such as e.g.
 for the halogens look fine.
 Hope this helps,
 Martin Kaupp
 | Dr. Martin Kaupp                                               |
 | Max-Planck-Institut fuer Festkoerperforschung,                 |
 | Heisenbergstrasse 1, D-70569 Stuttgart, Germany,               |
 | Tel.: country-code+711/689-1532                                |
 | Fax.: country-code+711/689-1562                                |
 | email: kaupp.,at,                       |
 |                                                                |
 | and Institut fuer Theoretische Chemie, Universitaet Stuttgart, |
 | Pfaffenwaldring 55, D-70569 Stuttgart, Germany                 |
 | Tel.: country-code+711/685-4399                                |
 | Fax.: country-code+711/685-4442                                |
 laurent JOUBERT (joubert.,at, wrote asking about ECPs for
 lanthanides for use in DFT calculations, and whether or not the HF ECPs
 work correctly.
 I cannot address the first question, but the second one was addressed in the
 recent paper by Hay and coworkers
 Russo, T. V.; Martin, R. L.; Hay, P. J., J. Phys. Chem., 1995, 99, 17085.
 The basic consesus of the paper is that while the ECP itself may be used
 with good reliability, at least when compared to all electron calculations,
 the HF basis set is probably not reliable with DFT calculations.
                                                 -fred arnold
 Salut Laurent,
 En general, les orbitales HF peuvent marcher pour des calculs DFT, mais il
 existe des problemes de BSSE (Basis Set Superposition Error)
 Typiquement, si tu dissocies par exemple une molecule,
 H2O, il serait bon, lorsque que tu augmentes une distance O-H (chemin de
 dissociation), de retrouver pour une grande distance, la somme des energies
 OH(-) et H(+). Generalement, avec les orbitales HF, ca ne marche pas, et les
 orbitales DFT (celles de DMol ou celles de DGauss/deMon) sont faites pour la
 methodologies DFT, de facon a supprimer les pb de BSSE. Tu peux voir la
 N. Godbout, D.R. Salahub, J. Andzelm et E. Wimmer, Can. J. Chem., 70, 560,
 pour les orbitales DFT (DGauss/deMon) pour la methodologie LCGTO-MCP (Linear
 Combination of Gaussian-Type Orbitals-Model Core Potential). Pour celles de
 DMol, tu peux regarder la doc. sur DMol.
 Sinon, si tu as des infos sur des ECP (avec effet relativiste), sur des
 actinides/lanthanides, je suis preneurs.
 En esperant que cela t'aide,
 Frederic Bouyer
 Groupe Chimie Corrosion - RNE-EMA
 EDF-DER Les Renardieres              Tel:   (33) 1-60-73-69-65
 Route de Sens, Ecuelles, BP1         Fax:   (33) 1-60-73-68-89
 77250 MORET-SUR-LOING - FRANCE       Email: Frederic.Bouyer.,at,
 En generale les ECP HF (ex: Hay and Wadt '85) peuvent etre utilises aussi pour
 des calculs DFT. Je suis en train de les utiliser sur des clusters de palladium
 et les resultats semblent tres bons et coherents. C'est quand meme vrai que si
 tu peut utiliser des ECP developpes pour la DFT (et fiables) alors cela
 pourrait etre encore mieu.
 Bon courage,
 Laurent JOUBERT
                          Laurent JOUBERT
           Ecole Nationale Superieure de Chimie de Paris
       Laboratoire d'Electrochimie et de Chimie Analytique
                   11, rue Pierre et Marie Curie
                    75231 PARIS CEDEX 05- FRANCE
                         Tel : 44-27-66-94
                         Fax : 44-27-67-50
                   E-Mail : joubert.,at,
                WWW site :