Dear Dave,

First of all, in dealing with lanthanides or actinides, non-aufbau
configurations occur very often because of the closeness of energies of the
relevant orbitals. In Hartree-Fock method, the unoccupied orbitals are
artificially pushed up so you don't see this phenomenon. In DFT, the
orbital energies go up upon occupation due to the Janak relationship.
Therefore, non-aufbau happens when orbitals are not well separated
energetically. So the non-aufbau problem is not because of your basis sets.
For a "politically-correct" DFT treatment, you probably need to use the FON
method. For details, see, for example, Wang and Schwarz, J. Chem. Phys.
1995, 102, 9296; 1996, 105, 4641.

Secondly, if you include spin-orbit correction in the Create run, your
fragments will not work for your molecular calculations. At least that was
the case for earlier ADF versions. That is, you cannot use a double-group
based fragment.

Thirdly, as long as your occupation is in line with the atomic
configuration (ADF is always right for this in neutral atoms), you probably
don't need to worry about the non-aufbau situation for your Create run,
which only provides atomic fragments. For your lanthanide molecules, you
most likely will have the non-aufbau occupation if you use a non-FON
method. Usually to choose a lowest-energy configuration among these
low-lying configurations (arising from 4f^x) seems oaky for your purpose.

Jun Li

At 12:09 PM 5/20/00 -0100, you wrote:
>Dear fellow ADF users,
> We have had many problems looking at Lanthanide compounds using
>ADF 2000.02. Specifically, all the basis sets under all conditions
>(ZORA full relativistic and Spin Orbit included) converge on a
>non-aufbau configuration in the Create run and if one carries out
>a full ADF run on the isolated atom one still gets a non-aufbau
>configuration (although the occupancy is OK). This occurs even with
>a huge basis set sent to us by the people at SCM. This seems to be
>causing trouble with our complexes (strange charge distributions and such
>like).
>Has anyone similar or better experiences with such lanthanide complexes
>using ADF (not BDF)?
>
>Part of a Create run for La is shown below -
>
>Basis set IV small core (4d frozen) as an example
>
> Scaled ZORA Orbital Energies, per Irrep and Spin:
> =================================================
> Occup E (au) E(eV) Diff (eV) with prev.
> ----- -------------------- ------ --------------------
> S
> 1 2.000 -0.14246309007435E+01 -38.766 -3.14E-10
> 2 2.000 -0.13970524295544E+00 -3.802 -2.29E-10
> 3 0.000 0.11962683814617E-01 0.326
> 4 0.000 0.24314806598705E+00 6.616
> 5 0.000 0.15731510908424E+01 42.808
> P
> 1 6.000 -0.82832669410432E+00 -22.540 -2.83E-10
> 2 0.000 -0.45890565460260E-01 -1.249
> 3 0.000 0.58492696041089E+00 15.917
> D
> 1 1.000 -0.11726987813749E+00 -3.191 -2.78E-10
> 2 0.000 0.64804613564763E-01 1.763
> 3 0.000 0.63031644428328E+00 17.152
> F
> 1 0.000 -0.15391143861920E+00 -4.188
> 2 0.000 0.48835583241281E+00 13.289
> 3 0.000 0.44031663337683E+01 119.816

*******************************************************
Department of Chemistry Tel: (614)292-5227 (Office)
The Ohio State University (614)529-7104 (Home)
100 West 18th Avenue Fax: (614)292-1685
Columbus, Ohio 43210-1173 E-Mail: li.208_at_osu.edu
USA
      Homepage: http://chemistry.ohio-state.edu/~junli
********************************************************
Received on 2000-12-08 12:03:10